JP3858418B2 - Electronic discussion support device - Google Patents

Description

【００５２】

属性名：：＝任意の文字列の集合の要素
属性値：：＝任意の文字列の集合の要素｜任意の定数の集合の要素
上限値：：＝定数｜システム定数
下限値：：＝定数｜システム定数
関連種別：：＝任意の文字列の集合の要素
キーワード：：＝任意の文字列の集合の要素
【００５３】
上記の内容は、本実施形態における討論空間のための検索式の文法を表現している。例えば、
dispace-attrib（意見総数,[0,1000]）:article-attrib（発言者,{甲,乙}）, article-attrib（発言者,{丙,丁}）
は検索式である。
部分討論検索述語および全討論検索述語の意味は次のように定義される。
【００５４】
ａｒｔｉｃｌｅ−ａｔｔｒｉｂ（意見属性名：［下限値，上限値］）＝
｛〈ｉｄdi,Ｐdi，Ｒdi，｛ａj｜ａj∈Ａdi,〈ｎ′，ｖ〉∈Ｐaj，ｎ′＝ｎ，ｌ＜＝ｖ＜＝ｈ｝〉｜ｄi∈Ｄ｝
ただし、ｎ＝意見属性名、ｌ＝下限値、ｈ＝上限値、Ｄ＝討論ビューサーバ管理下の全ての討論空間の集合である。これは、「意見属性名ｎで示された属性値の値が下限値と上限値の間にあるような意見のみを抽出した意見集合をもつ討論空間の集合」を表している。例えば、発言時刻がある時刻からある時刻にある意見のみを抽出するといった用途に使うことができる。なお、上限値と下限値は任意の実数値をとることができるほかに、システムが用意する特殊な定数を利用できる。例えば、現在の時刻を表す定数や、現在の時刻からｎ時間前の時刻を表す定数など様々なものが利用できる。なお、このように、ある性質を持った意見集合を切り出した形の意見集合をもつ討論空間を「部分討論空間」と呼ぶ。
【００５５】
ａｒｔｉｃｌｅ−ａｔｔｒｉｂ（意見属性名：｛属性値１、・・・、属性値ｍ｝）＝｛〈ｉｄdi，Ｐdi，Ｒdi，｛ａj｜ａj∈Ａdi，〈ｎ′，ｖ〉∈Ｐaj，ｎ′＝ｎ，ｖ∈Ｖ｝〉｜ｄi∈Ｄ｝
ただし、ｎ＝意見属性名、Ｖ＝｛属性値１、・・・、属性値ｍ｝、Ｄ＝討論ビューサーバ管理下の全ての討論空間の集合である。これは、「意見属性名ｎで示された属性値が属性値の集合｛属性値１、・・・、属性値ｍ｝の要素であるような意見のみを抽出した意見集合をもつ討論空間の集合」を表している。たとえば、発言者がこの人かあの人となっている意見のみを抽出するといった用途に使うことができる。
【００５６】
ａｒｔｉｃｌｅ−ｒｅｆ（｛関連種類１、・・・、関連種類ｍ｝）＝
｛〈ｉｄdi，Ｐdi，Ｒdi，｛ａj｜ａj∈Ａdi，〈ｉｄ，ｖ〉∈Ｒaj，v∈Ｖ｝〉｜ｄi∈Ｄ｝
ただし、Ｖ＝｛関連種類１、・・・、関連種類ｍ｝、Ｄ＝討論ビューサーバ管理下の全ての討論空間の集合である。これは、「関連種類が｛関連種類１、・・・、関連種類ｍ｝の要素であるような意見のみを抽出した意見集合をもつ討論空間の集合」を表している。たとえば、関連が「質問」「返答」になっている意見のみを抽出するといった用途に使うことができる。
【００５７】
ａｒｔｉｃｌｅ−ｂｏｄｙ（キーワード）＝
｛〈ｉｄdi，Ｐdi，Ｒdi，{ａj｜ａj∈Ａdi，Ｂajがｋを含む｝〉｜ｄi∈Ｄ｝
ただし、ｋ＝キーワード、Ｄ＝討論ビューサーバ管理下の全ての討論空間の集合である。これは、「キーワードを意見本文に含む意見のみを抽出した意見集合を持つ討論空間の集合」を表している。
【００５８】
ｄｓｐａｃｅ−ａｔｔｉｂ（討論空間属性名：［下限値、上限値］）＝
｛ｄi｜ｄi∈Ｄ，〈ｎ′，ｖ〉∈Ｐdi，ｎ′＝ｎ，ｌ＜＝ｖ＜＝ｈ｝
ただし、ｎ＝討論空間属性名、ｌ＝下限値、ｈ＝上限値、Ｄ＝討論ビューサーバ管理下の全ての討論空間の集合である。これは、「討論空間属性名ｎで示された属性値の値が下限値と上限値の間にあるような討論空間の集合」を表している。たとえば、討論開始時刻がある時刻からある時刻にある意見のみを抽出するといった用途に使うことができる。なお、上限値と下限値は任意の実数値をとることが出来るほかに、システムが用意する特殊な定数を利用できる。たとえば、現在の時刻を表す定数や、現在の時刻からｎ時間前の時刻を表す定数など様々なものが利用できる。
【００５９】
ｄｓｐａｃｅ−ａｔｔｒｉｂ(討論空間属性名：｛属性値１、・・・、属性値ｍ｝)＝
｛ｄi｜ｄi∈Ｄ，〈ｎ′，ｖ〉∈Ｐdi，ｎ′＝ｎ，ｖ∈Ｖ｝
ただし、ｎ＝討論空間属性名、Ｖ＝｛属性値１、・・・、属性値ｍ｝、Ｄ＝討論ビューサーバ管理下の全ての討論空間の集合である。これは、「討論空間属性名ｎで示された属性値が属性値の集合｛属性値１、・・・、属性値ｍ｝の要素であるような討論空間の集合」を表している。たとえば、主催者がこの人かあの人となっている討論空間のみを抽出するといった用途に使うことができる。
【００６０】
ｄｓｐａｃｅ−ｒｅｆ（｛関連種類１、・・・、関連種類ｍ｝）＝
｛ｄi｜ｄi∈Ｄ，〈ｉｄ，ｖ〉∈Ｒdi，ｖ∈Ｖ｝
ただし、Ｖ＝｛関連種類１、・・・、関連種類ｍ｝、Ｄ＝討論ビューサーバ管理下の全ての討論空間の集合である。これは、「関連種類が｛関連種類１、・・・、関連種類ｍ｝の要素であるような討論空間の集合」を表している。たとえば、関連が「下位議論」になっている討論空間のみを抽出するといった用途に使うことができる。
【００６１】
また、論理和（＋）および論理積（＊）および論理差（＃）の意味は次のように定義される。
Ｄ１＊Ｄ２＝
｛〈ｉｄdi，Ｐdi，Ｒdi，Ａdi∩Ａdj〉｜ｄi∈Ｄ1，ｄj∈Ｄ2，ｉｄdi＝ｉｄdj｝、
これは、図８に示すように、討論空間の集合Ｄ１、Ｄ２の双方に含まれている同一のＩＤを持つ討論空間のみについて、双方に共通して含まれる意見のみを残した形の新たな討論空間の集合を生成する操作である。
【００６２】
Ｄ１＋Ｄ２＝
｛〈ｉｄdi，Ｐdi，Ｒdi，Ａdi∪Ａdj〉｜ｄi∈Ｄ1，ｄj∈Ｄ2，ｉｄdi，＝ｉｄdj｝∪｛ｄi｜ｄi∈Ｄ1，ｄj∈Ｄ2，ｉｄdi≠ｉｄdj｝∪｛ｄj｜ｄi∈Ｄ1，ｄj∈Ｄ2，ｉｄdi≠ｉｄdj｝、
これは、図９に示すように、討論空間の集合Ｄ１、Ｄ２の双方に含まれている同一のＩＤを持つ討論空間に関して、双方に含まれている意見を合わせた形の新たな討論空間の集合と、さらに、どちらか片方にのみ含まれている討論空間の集合を合わせた討論空間の集合を生成する操作である。
【００６３】
Ｄ１＃Ｄ２＝
｛〈ｉｄdi，Ｐdi，Ｒdi，Ａdi＼Ａdj〉｜ｄi∈Ｄ1，ｄj∈Ｄ2，ｉｄdi＝ｉｄdj｝∪｛ｄi｜ｄi∈Ｄ1，ｄj∈Ｄ2，ｉｄdi≠ｉｄdj｝、
これは、図１０に示すように、討論空間の集合Ｄ１、Ｄ２の双方に含まれている同一のＩＤを持つ討論空間に関して、Ｄ１に含まれている意見からＤ２に含まれている意見を削除した形の新たな討論空間の集合と、さらに、Ｄ１のみに含まれている討論空間の集合を合わせた討論空間の集合を生成する操作である。
【００６４】
さらに、限定子（：）および接続子（，）の意味は次のように定義される。
Ｄ１：Ｄ２はＤ１＊Ｄ２であり、これにより、討論空間の集合Ｄ２をＤ１との論理積により限定することができる。
Ｄ１，Ｄ２はＤ１＋Ｄ２であり、これにより、討論空間の集合Ｄ１とＤ２を論理和により結合することができる。
【００６５】
「討論空間検索要求の構成」
上記した討論空間検索要求（ｑｕｅｒｙ）の構成を、図１１及び図１２を参照して説明する。
通常の討論空間検索要求は図１１に示すような構成となっており、検索要求ＩＤ、処理結果返信先、検索要求伝搬経路、検索要求種別、検索式を含んでいる。
検索要求ＩＤは、ネットワーク５上で一意に定められる検索要求の識別番号であり、ホスト名や時刻や乱数を用いて生成することができる。
処理結果返信先は、討論ビューサーバ２０が討論ビュー（討論展望図）を送付する対象であり、討論空間検索要求を発信した討論ビュークライアント１０を指し示している。この処理結果返信先情報はホスト名やポート番号などを用いてネットワーク５上で一意に定めることが出来る。
【００６６】
検索要求伝搬経路は、複数の討論ビューサーバ２０の間で討論空間検索要求が配送される際の配送履歴であり、討論空間検索要求が通過した討論ビューサーバ２０のホスト名からなるリストである。
検索要求種別は、処理の種別を表す情報であり、図１１に示す通常の討論空間検索要求では「検索」を表す記号が格納されている。
検索式は、上述の討論空間検索式そのものが格納されている。
【００６７】
図１２に示す討論空間検索要求は、過去に要求した討論空間検索の要求を取消すための特別な討論空間検索要求であり、検索要求ＩＤと処理結果返信先と検索要求伝搬経路は通常の討論空間検索要求と同様である。
検索要求種別は、討論空間検索の取消では「取消」を表す記号が格納されている。
また、討論空間検索の取消では、検索式の代わりに、取消対象の検索要求ＩＤが付加されており、このＩＤ情報によって取消対象の討論空間検索要求が指定される。
【００６８】
「討論ビューサーバの動作」
本実施形態における討論ビューサーバ２０は下記のように動作する。
図１３は討論空間検索制御手段２１の動作を表現するＰＡＤ図である。この討論空間検索制御手段２１は、ネットワーク５経由で討論空間検索要求を受信し、その要求の種別により検索要求の登録または取消処理を行うものである。また、討論空間検索制御手段２１は、検索連携手段２４に検索要求を渡して他の討論ビューサーバ２０との連携を図ったり、検索要求に対して定期的に検索を行うための専用スレッドを新たに生成するなどの処理も行っている。
【００６９】
すなわち、図１３に示すように、討論空間検索制御手段２１は、実行開始時において、手続き「討論空間検索制御」を実行し、この手続きはステップｓ１−００１においてすぐにイベント処理のための無限ループに入る。このループ内では、ステップｓ１−００２においてイベントｅの取得を行い、取得したイベントｅが「新規検索要求の到着」すなわち討論空間検索制御手段２１が検索要求を受信した場合には、ステップｓ１−００３からステップｓ１−００４以降の一連の処理に分岐する。
【００７０】
ステップｓ１−００４では受信された検索要求ｒを取り出し、ステップｓ１−００５でこの検索要求ｒが検索状況記憶手段２２にすでに登録されているか否かを検索要求に含まれている検索要求ＩＤを用いて判定する。もし、検索要求ｒが検索状況記憶手段２２に登録されていない場合には、処理はｓ１−００６以降に移行し、ステップｓ１−００６では検索要求ｒを検索状況記憶手段２２に登録するとともに、この検索要求ｒに対応する検索状況を初期化する。なお、検索状況については後に説明する。
【００７１】
そして、ステップｓ１−００７、ｓ１−００８において、検索要求ｒの検索要求ＩＤを取り出し、さらに、このＩＤを識別番号に用いた「検索制御スレッド」を新たに生成する。この処理において、後述する手続き「検索制御スレッド」に書かれた処理を行うスレッドが形成される（このスレッドによる処理は、他の処理から独立して並行処理される処理になる）。
さらに、ステップｓ１−００９において、検索要求ｒを検索連携手段２４に渡すことにより、複数の討論ビューサーバ２０間の連携動作を行わせる。なお、検索連携手段２４の動作は後に説明する。
【００７２】
一方、取得したイベントｅが「検索取消要求の到着」すなわち討論空間検索制御手段２１が検索取消要求を受信した場合には、ステップｓ１−０１０からステップｓ１−０１１以降の処理に分岐する。
ステップｓ１−０１１では受信された検索取消要求ｒを取り出し、ステップｓ１−０１２で検索取消要求ｒより取消対象の検索要求ＩＤを変数ｉｄに取り出す。そして、ステップｓ１−０１３では、この検索要求ＩＤに対応する検索状況が検索状況記憶手段２２に存在するかどうかを判定し、もし、存在する場合にはステップｓ１−０１３からステップｓ１−０１４以降の処理に分岐し、ステップｓ１−０１４、ｓ１−０１５において、検索状況記憶手段２２からｉｄに対応する検索状況を削除し、ｉｄに対応する検索制御スレッドを削除する。
また、ステップｓ１−０１６では、検索取消要求ｒを検索連携手段２４に渡すことにより複数の討論ビューサーバ２０間の連携動作を行わせる。なお、検索連携手段２４の動作は後に説明する。
【００７３】
図１４には、上述した「検索制御スレッド」手続きの処理内容を示してあり、この検索制御スレッドは変数ｉｄに検索要求ＩＤを伴って生成される。
まず、ステップｓ２−００１において、検索状況記憶手段２２によりｉｄに対応する検索要求から検索式ｑを抽出し、ステップｓ２−００２において検索状況記憶手段２２よりｉｄに対応する検索要求から処理結果返信先ｂを抽出して、ステップｓ２−００３で検索制御のための無限ループに入る。
ループ内では、一定時間待った後（ステップｓ２−００４）、検索式ｑを用いて討論空間検索手段２３に討論空間検索を行わせて検索結果ａを得て（ステップｓ２−００５）、さらに、検索状況記憶手段２２上で検索状況を更新し（ステップｓ２−００６）、検索結果ａを処理結果返信先ｂに返す（ｓ２−００７）、といった処理を繰り返す。
【００７４】
図１５には、検索状況記憶手段２２が格納している情報のデータ構造を示してある。
検索状況記憶手段２２には、図示のように検索要求とその検索要求に対する検索結果との対を基本としたリスト構造が格納されており、この検索要求とその検索要求に対する検索結果の対が検索状況である。
検索要求には討論空間検索要求そのものが格納され、検索結果には討論空間全体のうちの検索要求を満たす部分のみが討論空間と同様なデータ構造をとって記憶される。なお、返信先に返される検索結果は、この形のデータが適切な符号化を用いて返信先に送信される。
【００７５】
図１６は討論空間検索手段２３の動作を表現するＰＡＤ図であり、この手続き「討論空間検索」は検索式ｑをともなって検索制御スレッドより呼び出される。まず、ステップｓ３−００１では、検索式ｑが文法的に正しいかどうか判定される。これは前述の検索式の文法規則に照らし合わせて行うことができる。そして、正しくない場合には、ステップｓ３−００２において、エラーを返してこの手続きを終了する。
一方、正しい場合には、ステップｓ３−００３において検索結果を格納する変数ａを空リストに結びつけ、ステップｓ３−００４において討論空間管理テーブルの行を指すポインタｄｐを討論空間管理テーブルの先頭を指すように初期化する。
【００７６】
そして、ステップｓ３−００５で討論空間管理テーブルの行を先頭から末尾まで走査するループに入り、ステップｓ３−００６においてポインタｄｐの指す討論空間管理テーブルについて、それが検索式ｑに適合するかどうか判定される。この判定には、ポインタｄｐが指している討論空間管理テーブルの行に結びつけられている「属性リスト」の各属性値と、「関連リスト」の関連種別の値と、ｑに現れる限定子「：」の左辺に指定される全討論検索式が用いられる。
ｄｐの指す討論空間管理テーブルの行が、検索式ｑに現れる全討論検索式のうちの一つでもマッチすれば、この討論空間は適合すると判定される。ただし、限定子「：」の左辺値が空の場合には、必ずマッチする。なお、全討論空間検索式のマッチの判定は、上述した討論空間検索式の意味に基づいて判定される。すなわち、各全討論検索述語が意味付ける属性値や関連種別値の一致と、それら述語の間の論理演算の意味に基づき判定される。
【００７７】
ｄｐの指す討論空間が適合する場合には、処理はステップｓ３−００７へと分岐する。ステップｓ３−００７において、変数ｑ０にｄｐが指す討論空間管理テーブルの行に適合する討論検索項のみを取り出す形で検索式ｑを再構成したものを代入する。ステップｓ３−００８では、新規空リストを生成し、これを変数ａ０に結びつける。次いで、ステップｓ３−００９において、変数ｄｐが指す討論空間管理テーブルの行に結びつけられている意見管理テーブルの先頭を指すポインタを変数ａｐに代入する。
【００７８】
そして、ステップｓ３−０１０で意見管理テーブルの行を先頭から末尾まで走査するループに入り、ステップｓ３−０１１において、ａｐが指す意見管理テーブルかその意見管理テーブルからたどれる意見本文が検索式ｑに適合するかどうか判定される。この判定には、ポインタａｐが指している意見管理テーブルの行に結びつけられいる「属性リスト」の各属性値および「関連リスト」の関連種別の値および意見本文と、検索式ｑ０の各討論検索項に現れる限定子「：」の右辺に指定される部分討論検索式が用いられる。
ａｐの指す意見管理テーブルの行およびこれからたどれる意見本文が、検索式ｑ０に現れる部分討論検索式のうち一つでもマッチすれば、この意見は適合すると判定される。ただし、限定子「：」の右辺値が空の場合には、必ずマッチする。なお、部分討論空間検索式のマッチの判定は、上述した討論空間検索式の意味に基づいて判定される。すなわち、各部分討論検索述語が意味付ける属性値や関連種別値やキーワードの一致と、それら述語の間の論理演算の意味に基づき判定される。
【００７９】
そして、ａｐの指す意見が適合する場合には、処理はステップｓ３−０１２へと分岐し、リストａ０の末尾にａｐが指す意見テーブルの行が追加される。ステップｓ３−０１３では、意見管理テーブルの行を示すポインタａｐを先に進め、ループｓ３−０１０の先頭に戻る。このループは、意見管理テーブルの末尾まで達すると終了し、処理はステップｓ３−０１４に分岐する。
ステップｓ３−０１４においては、ａ０が空リストでないすなわち何らかの適合意見が見つかっている場合は、処理はステップｓ３−０１５に分岐し、検索結果の構成を行う。ステップｓ３−０１５では、変数ａｌにｄｐが指す討論空間管理テーブルの行を複製して結びつける。さらに、ステップｓ３−０１６では、変数ａｌの意見リストを適合意見のリストである変数ａ０につなぎ変える。これにより、適合意見のみからなる意見リストａ０を意見リストとして持った討論空間を表現する、新たな討論空間管理テーブルの行ａｌが構成される。そして、ステップｓ３−０１７において、検索結果ａにこの新たな討論空間管理テーブルの行ａｌを追加する。
【００８０】
また、ステップｓ３−０１９では、討論空間管理テーブルの行を指すポインタｄｐを先に進め、ループｓ３−００５の先頭に戻る。このループは、討論空間管理テーブルの末尾まで達すると終了し、ステップｓ３−０１９にて、結果ａを返す。
このように、本例における検索結果は、適合する討論空間毎に、適合する意見群を抽出してそれをリストに結びつける形となる。
【００８１】
図１７は検索連携手段２４の動作を表現するＰＡＤ図である。この図における検索連携制御手続きは、検索要求ｒをともなって討論空間検索制御手段より呼び出される。
まず、ステップｓ４−００１において、検索要求ｒの検索要求伝搬経路長が所定の長さ（ホップ数制限：これは、各討論ビューサーバ毎に予め定められている）以下であり、なおかつ、検索要求伝搬経路長に自ホスト名が含まれていない場合に、ステップｓ４−００２以降に分岐してその処理を行う。この条件判断は、検索要求がいつまでも消滅せずにネットワーク５上をたらい回しされるのを防ぐのが目的である。
【００８２】
ステップｓ４−００２では、検索要求ｒの検索要求伝搬経路のフィールドに自ホストを追加し、新たな検索要求ｒ’を得る。そして、ステップｓ４−００３、ｓ４−００４において、所定の転送先ホスト（これは、各討論ビューサーバ毎に予め定められている）各々に検索要求ｒ’を転送する。検索要求ｒ’を受信する相手は、相手ホストｄの討論空間検索制御手段である。
【００８３】
「討論ビュークライアントの動作」
図１に示したように、本実施形態の電子的討論支援装置では、討論ビュークライアント１０はＷＷＷブラウザ１１および討論実行手段１２と協調して動作する。この討論実行手段１２は討論空間中の意見を読んだり、討論空間中に新たな意見を置いたりするための機構であり、本発明が適応対象の一例としている電子会議システムのインフラストラクチャーの一部である。
【００８４】
図１８には本実施形態における、ＷＷＷブラウザ１１から見たユーザインターフェースの一例を示してあり、同図中のＦ１およびＦ２は討論実行手段１２の出力であり、Ｆ３は討論ビュークライアント１０が司るユーザインターフェースである。
この図１８に示す状態では、表示領域Ｆ３には「検索式」にマッチする現在の部分討論空間への入口が「討論空間Ａ」「討論空間Ｂ」「討論空間Ｃ」というアイコンで表示されて、その内の「討論空間Ａ」がユーザにより選択されており、また、表示領域Ｆ２には、表示領域Ｆ３で選択された部分討論空間を含む討論空間が表示されている。なお、表示領域Ｆ２の表示で点線表示になっている部分は、その討論空間の意見の内で検索結果の部分討論空間に含まれない意見であり、実線表示になっている部分が検索結果の部分討論空間に含まれるる意見である。この表示領域Ｆ２の表示は、討論実行手段１２が、討論ビュークライアント１０から渡される検索結果の部分討論空間に含まれる意見と、部分討論空間のもとの討論空間の全ての意見を比較することで生成している。
また、図１８に示す状態では、表示領域Ｆ２で影が付けられている意見がユーザにより選択されており、その意見の内容が表示領域Ｆ１に表示されている。
【００８５】
討論ビュークライアント１０は、ＷＷＷブラウザ１１および討論実行手段１２と協調動作し、ユーザが表示領域Ｆ３の部分討論空間アイコンをダブルクリックするなどの所定のアクションを入力すると、当該部分討論空間の入口に対応した討論空間および部分討論空間を表示領域Ｆ２に図１８に示すように表示させることを、ＷＷＷブラウザ１１および討論実行手段１２に指示することが出来る。
また、「検索式」は複数持つことができて、図１９に示すように、複数の検索式をポップアップリストＰＬによって表示させて、その内から選択することが出来る。なお、Ｆ３の領域に現在表示されていないものも、内部的には常に保持・更新されている。
【００８６】
図２０は討論展望図関連ＵＩ手段１３の動作を示すＰＡＤ図である。
まず、ステップｓ５−００１にてイベント待ちの無限ループに入り、ステップｓ５−００２において、イベントを待って、そのイベントを結果ｅに代入する。イベントはユーザの操作に伴って、ＷＷＷブラウザ１１より発生するか、ネットワーク５経由で討論ビューサーバ２０より検索結果がもたたされたときに発生する。
【００８７】
次いで、ステップｓ５−００３において、イベントｅが部分討論空間へのアクセス要求であった場合には、処理はステップｓ５−００４に分岐する。このイベントは、ユーザが図１８に示した表示領域Ｆ３上で部分討論空間の入口をダブルクリックするなどの所定のアクションを起こした際に発生する。
この場合、ステップｓ５−００４において、ユーザのアクションにより指定された部分討論空間を特定し、更に、ステップｓ５−００５で、討論実行手段１２に対して部分討論空間の表示をする。この結果、図１８に示した表示領域Ｆ２上に該当する部分討論空間が表示される。これにより、ユーザは自分の興味に沿った討論の部分に対して素早いアクセスをおこなうことできる。
【００８８】
一方、ステップｓ５−００６において、イベントｅが表示討論展望図の変更要求であった場合には、処理はステップｓ５−００７に移る。このイベントは、ユーザが図１９に示したようにポップアップリストＰＬ上で選択中の「検索式」を変更した際に発生する。
この場合、ステップｓ５−００７において、選択された「検索式」に対する検索要求ＩＤを変数ｉｄに取り出す。ここで、討論ビュークライアント１０は、ユーザによって起動時より指定されてきた全ての討論空間検索式とその検索結果（討論ビュー）を保持している。複数の検索式と検索結果を識別する鍵は、検索要求発行時に生成される検索要求ＩＤに基づいている。
そして、ステップｓ５−００８において、検索要求ＩＤよりそれに対応する検索結果ｖ｛ｉｄ｝を取り出し、ｓ５−００８においてこれを再表示する。なお、ｖ｛ｉｄ｝は検索要求ＩＤ番号をキーとして検索結果を格納した連想配列変数である。
【００８９】
一方、ステップｓ５−００９において、イベントｅが討論展望図更新要求であった場合には、処理はステップｓ５−０１０に移る。このイベントは、展望図構成手段１４が検索結果を受信した際に発生する。
この場合、ステップｓ５−０１０において、ステップｓ５−００７と同様に検索要求ＩＤを取り出し、ステップｓ５−０１１において、この検索要求ＩＤと検索結果を組み合わせて記憶する。さらに、ステップｓ５−０１２、ｓ５−０１３において、検索要求ＩＤに対応する討論展望図が表示されている最中であれば、表示の更新も行う。
【００９０】
一方、ステップｓ５−０１４において、イベントｅが討論空間検索登録要求であった場合には、処理はステップｓ５−０１５に移る。このイベントは、ユーザが「検索式」フィールドに新たな検索式を記入し、「登録」ボタンを押した際に発生する。
この場合、ステップｓ５−０１５において、登録対象の検索式ｑを取得し、ステップｓ５−０１６において、討論空間検索要求生成手段１５に対して、検索式ｑに対する検索要求の発行を指示する。
【００９１】
一方、ステップｓ５−０１７において、イベントｅが討論ビュークライアント１０の停止要求であった場合には、処理はステップｓ５−０１８に移る。このイベントは、ユーザの指示により、ＷＷＷブラウザ１１が討論ビュークライアント１０の表示をとりやめたり、ブラウザ１１自身が終了される場合に発生する。
この場合、ステップｓ５−０１８において、討論空間検索要求生成手段１５に対して討論ビュークライアント１０が保持している全ての討論空間検索要求に対する検索取消要求を発行するように指示する。
【００９２】
図２１には展望図構成手段１４の動作を示すＰＡＤ図である。
この展望図構成手段１４は、ネットワーク５を監視して、討論ビューサーバ２０から検索結果が返されたときに動作し、すぐにｓ６−００１においてイベント待ちのループに入る。そして、ステップｓ６−００２において、イベントｅを取得し、ステップｓ６−００３においてイベントｅが討論空間検索結果の返信である場合は動作を起こす。
【００９３】
そして、ステップｓ６−００４において、返信された結果から検索結果ａを取り出し、これをもとに、ステップｓ６−００５において、討論展望図の構成を行い結果をｖに代入する。本実施形態では討論展望図は構造を持たないので、部分討論空間への入口を何らかのジオメトリ管理ポリシーに則って配列するだけの処理を行うが、もちろん、他の構成を行っても構わない。
そして、ステップｓ６−００６において、討論展望図関連ＵＩ手段１３に対して、討論展望図をｖに更新するように要求する。上述したように、この要求は、討論展望図関連ＵＩ手段１３ではイベントとして取得され、更新が行われる。
【００９４】
図２２は討論空間検索要求生成手段１５の動作を示すＰＡＤ図である。
図２２（ａ）に示すように、手続き「討論空間検索要求発行」は検索式ｑを伴って、討論展望図関連ＵＩ手段１３より呼び出される。ステップｓ７−００１において、検索式ｑに対して、図１１に示したようにフィールドを追加することにより検索要求ｒを生成し、ステップｓ７−００２において、ユーザが設定した討論ビューサーバ２０に向けてこの検索要求ｒを送信する。
また、図２２（ｂ）に示すように、手続き「全討論空間検索要求取消発行」も討論展望図関連ＵＩ手段１３より呼び出される。ステップｓ７−００３で、討論ビュークライアント１０が保持する全ての検索要求に対して、図１２に示したような検索取消要求を生成し、ユーザが設定した討論ビューサーバ２０に向けて送信する。
【００９５】
「動作例」
図２３、図２４、図２５には、本実施形態における動作例を示してある。
ユーザの興味が、多くの討論でキーとなる人物「発言者Ａ」および「発言者Ｂ」にあり、さらに、これらの人物の最近３時間の動向を知りたいとする。
この場合、ユーザは自分の興味を検索式で表現し、討論ビュークライアント１０を用いて、これを討論ビューサーバ２０に送信する。この場合の検索式は“ａｒｔｉｃｌｅ−ａｔｔｒｉｂ（意見発言時刻，［３時間前，現在時刻］）＊ａｒｔｉｃｌｅ−ａｔｔｒｉｂ（発言者，｛発言者Ａ，発言者Ｂ｝）”となる。
【００９６】
図２３には最初の状態を示してあり、ユーザの興味に合致する部分が特定され、ユーザはユーザインターフェース上に表示される討論展望図を通して、その部分に簡便にアクセスすることができる。
さらに、本実施形態ではユーザの興味の対象を時間を経て追跡しており、これにより、討論の進行に伴って刻々と移り変わる討論の最新の状況を、ユーザが苦労することなく簡単に把握できる。
図２４には図２３に示した状態より若干時間が過ぎた時の状況を示してあり、討論の進行に伴い、ユーザの興味に合致する部分討論空間が変化している。また、この時間経過に伴い、今までユーザの興味に関係なかった「討論空間２」が新たにユーザの興味の対象として出現している。
図２５には図２４に示した状態より、さらに若干時間が過ぎたときの状況を示してあり、さらにユーザの興味に合致する部分討論空間が変化している。
本発明によれば、このような討論の動的に変化する状況をユーザが操作することなく的確に追跡し、必要に応じて簡単にアクセスすることが出来るようになった。
【００９７】
次に、本発明に係る第２の実施形態について説明する。
本実施形態における装置の構成および討論空間の構造は上記の第１の実施形態と同様である。だたし、討論展望図の概念は若干異なるため、説明する。
【００９８】
「第２実施形態における討論ビューの概念」
本実施形態におけるサービスおよび討論展望図の概念構成は図２６に示すようであり、第１実施形態との違いは、ユーザの意図として、検索式（ｑｕｅｒｙ）のみならず、座標軸（ａｘｉｓ）を利用可能になっている点である。この座標軸は、例えば、時間軸や発言数などの数値情報、発言者名や関連種別などの文字列情報などを自由に設定することが出来る。
座標軸の導入により、第１実施形態では討論展望図には構造がながったところ、本実施形態では指定した座標軸に沿った構造化を行うことができるようになった。これにより、討論展望図の中で各部分討論空間の位置づけを明確化でき、ユーザは討論の現状をさらに的確につかむことが出来るようになった。
本実施形態では、第１実施形態と異なり、討論展望図構成手段１４が座標軸に基づいた討論展望図を構成するようになるため、部分的に変更されている。以下にこの変更部分について説明する。
【００９９】
「討論ビュークライアントの動作」
本実施形態における、ＷＷＷブラウザ１１から見たユーザインターフェースの例を図２７に示してある。第１実施形態と異なり、図２７中の領域Ｆ３のインターフェースデザインが変更されている。第１実施形態ではここには「検索式」にマッチする現在の部分討論空間の入口が構造を持たずに表現されていたが、本実施形態では、部分討論空間の入口アイコンが２つの座標軸（ａｘｉｓ）により張られる２次元グラフ上にマップされる。
座標軸の選択は、検索結果を評価してその座標軸へ写像するための評価関数の入力を行っていることになる。もちろん、検索結果としての部分討論空間が座標軸という何らかの評価尺度により多角的に分析可能であることが本実施形態の本質であるので、表現形態としては、３次元表示やより多次元の表示を縮退させたものなど他の様々な形態をとっても構わない。
【０１００】
図２８に本実施形態における討論展望図関連ＵＩ手段１３の動作を表現するＰＡＤ図を示す。この動作における上記した第１実施形態との相違点は、討論展望図再表示の手続き（ステップｓ８−００８、ｓ８−０１３、ｓ８−０２１）が座標軸に則った構造表示を行うように変更される点である。
本実施形態では、部分討論空間の属性または部分討論空間に含まれている意見の属性またはそれらを何らかの関数で評価した値を、ユーザが指定する座標軸上に単純に写像して表示する。本実施形態では、２次元グラフなので写像される領域は２次元図形を構成する。なお、この表示の一例は後述する。
【０１０１】
図２８における第１実施形態との違いは、受理可能なイベントｅに座標軸変更要求が追加されたことである（ステップｓ８−１９）。
このイベントは、図２７に示すように座標軸のポップアップリストＬによりユーザが座標軸を変更するときに発生する。この場合、処理は、ステップｓ８−２０、ステップｓ８−２１に分岐し、討論展望図を新しい座標軸に則って再表示する。
【０１０２】
「第２実施形態の動作例」
図２９および図３０には、本実施形態における動作例を示してある。
ユーザの興味が、最近３時間の議論の動向を知りたいとし、さらに、発言がどのような関連をもってなされたかを合わせて知りたいとする。
この場合、ユーザはまず自分の興味を検索式で表現し、討論ビュークライアント１０を用いて、これを討論ビューサーバ２０に送信する。この場合の検索式は“ａｒｔｉｃｌｅ−ａｔｔｉｒｂ（意見発言時刻、［３時間前、現在時刻］）”となる。
そして、ユーザは討論展望図を表示する座標軸を決定する。図２９では、検索式として「発言時刻が最近３時間以内」が、座標軸として「意見発言時刻」と「意見関連種別」が選択されている。
【０１０３】
図２９に示す状態は最初の状態であり、時刻１１：００時点での討論状況を示している。本実施形態では、討論展望図の表示において、部分討論空間の入口を表現する矩形は、該当する討論部分空間の属性値の範囲を示している。例えば、“Ｄｓｐａｃｅ−１”で表示される「討論空間１」の部分討論空間に対する入口は、この部分討論空間で最も古い発言の時刻が９：００であり、最も新しい発言の時刻が１０：３０であることを示している。また、意見の関連種別でみると、質問と回答が発言されたということを表している。
図３０に示す状態は図２９よりも時間が経ったときの状態である。この時点で、前の時間では質問と回答ばかりだった「討論空間１」の討論が、回答、提案、まとめなどのより生産的な討論中心に移ってきたのが分かる。また、「討論空間３」の討論は、前の時間では提案もなされていたが、この時点では質問のみになり、あまりうまく討論が進行していないことが分かる。
もちろん、本実施形態でも、第１実施形態と同じように、討論展望図の表示を用いて、所望の部分討論空間へのアクセスをおこなうこともできる。
このように、本実施形態では座標軸を導入することで、ユーザは、現在の討論の状況をより正確かつ適切に質にまで踏み込んで把握できるようになる。
【０１０４】
次に、本発明に係る第３の実施形態について説明する。
本実施形態における討論空間の構造は第１の実施形態と同様であるが、装置の構成および討論展望図の概念は若干異なるため説明する。
【０１０５】
「第３実施形態における討論展望図の概念」
本実施形態では、サービスおよび討論展望図の概念構成は図３１に示すようになっており、これは上記の第２実施形態と同様に、検索式（ｑｕｅｒｙ）と座標軸（ａｘｉｓ）に基づく部分討論空間へのアクセス手段をユーザに提供するものである。第２実施形態との違いは、討論空間関連生成手段３０が追加された点にある。討論空間関連生成手段３０は各討論ビュークライアント１１における討論展望図の状況を総合することで、ユーザ全員の興味の動向を検出し、さらに、それを各ユーザにフィードバックするものである。
この討論空間関連生成手段３０の導入により、多くのユーザが共通して興味を持っている討論の現状が浮き彫りにされる。さらに、このように多くのユーザの共通の興味のある討論空間が各ユーザに対して示されること、討論空間同士の交流が活発化し、一層討論が活性化するようになる。
【０１０６】
図３２は本実施形態の装置の構成図である。
第１実施形態に対して、討論空間関連生成手段３０がネットワーク５上に追加され、また、各討論ビュークライアント１０にそれぞれ討論空間関連連係手段１６が追加されている。討論空間関連生成手段３０と各討論空間関連連係手段１６が組合わさって討論関連付け手段を構成する。
討論空間関連生成手段３０は、各討論空間関連連係手段１６とネットワーク５を介して通信し、各ユーザの討論展望図を取得してそれらを総合し、その結果を討論空間関連連係手段１６に返すのものである。
【０１０７】
「討論ビュークライアントの動作」
図３３に本実施形態における討論展望図関連ＵＩ手段１３の動作を表現するＰＡＤ図を示す。
この図における第２実施形態との相違点は、討論空間関連情報受信通知イベントに対する処理が加わったことである（ステップｓ９−０２６）。このイベントは、討論空間関連連係手段１６が討論空間関連生成手段３０より討論空間関連情報を受信したときに発生するものである。
このイベントに対し、処理はステップｓ９−０２７に分岐する。ステップｓ９−０２７では、討論空間関連情報ｒを取得し、ステップｓ９−０２８でこれを記憶する。記憶された討論空間関連情報ｒは討論展望図再表示の手続き（ステップｓ８−００８、ｓ８−０１３、ｓ８−０２３）において利用される。なお、関連情報ｒのデータ構造は、関連する討論空間の討論空間ＩＤからなる集合である。
【０１０８】
また、もう一つの変更点は、ステップｓ９−０１４、ｓ９−０１５およびステップｓ９−０２４、ｓ９−０２５の追加である。これは、討論展望図が更新／修正された際に、討論空間関連連係手段１６に対し、討論空間の間の新たな関連情報を渡すものである。ステップｓ９−０１４またはｓ９−０２４では、更新／修正された討論ビュー討論空間の間の関連情報を検出している。本実施形態では、討論展望図上で討論空間を表現する図形が重なる部分の面積がある一定以上の大きさである討論空間は関連ありと判断される。もちろん、討論空間を表現する図形の重心間のユークリッド距離やその他の距離など、他にも様々な関連性の評価基準を利用できる。そして、ステップｓ９−０１５またはステップｓ９−０２５において、討論空間関連連係手段１６に討論空間関連情報ｒを渡している。
【０１０９】
「討論空間関連連係手段１６の動作」
図３４は第３実施形態における討論空間関連連係手段１６の動作を表現するＰＡＤ図である。
この動作は実行後すぐに、ステップｓ１０−００１において無限ループに入る。ループ内では、ステップｓ１０−００２でイベントを待ち、ステップｓ１０−００３では、受けたイベントが討論空間関連情報の送信である場合にステップｓ１０−００４に分岐する。このイベントは上述したように、討論展望図関連ＵＩ手段１３が発生する。
【０１１０】
ステップｓ１０−００４では、討論空間関連情報ｒを取り出し、ステップｓ１０−００５にてこれを討論空間関連生成手段３０へ送信する。討論空間関連生成手段３０との間で送受信される討論空間情報のデータ構造は、図３５に示すように、ここの情報を識別するための「討論空間関連情報ＩＤ」、情報の発信元を識別するための「発信元識別子」、関連する討論空間のＩＤ集合を含む「討論空間関連情報」を含んでいる。
一方、ステップｓ１０−００６では、受けたイベントが討論空間関連情報の受信である場合にステップｓ１０−００７に分岐する。このイベントは、討論空間関連生成手段３０から新たな討論空間関連情報を受信した際に発生する。ステップｓ１０−００７では、討論空間関連情報ｒを取り出し、ステップｓ１０−００８にてこれを討論展望図関連ＵＩ手段１３へ渡す。
【０１１１】
「討論空間関連生成手段３０の動作」
図３６は第３実施形態における討論空間関連生成手段３０の動作を表現するＰＡＤ図である。
この動作は実行後すぐに、ステップｓ１１−００１において変数ｓを空に初期化する。この変数は後述するように、ループ内で前回の処理における討論空間の関連情報を記憶するものである。
そして、ステップｓ１１−００２で無限ループに入り、ループ内では、ステップｓ１１−００３でイベントを待ち、ステップｓ１１−００４において、イベントｅが討論空間関連情報の受信であれば、ステップｓ１１−００５に分岐する。このイベントは、討論空間関連生成手段３０が各討論ビュークライアント１０の討論空間関連連係手段１６より討論空間関連情報を受信した際に発生する。
【０１１２】
ステップｓ１１−００５では受信した討論空間関連情報ｒを取得し、ステップｓ１１−００６でこれをもとに討論空間関連状況を再計算する。討論空間関連情報は、図３５に示したように関連する討論空間の集合の集合である。
討論空間関連生成手段３０は各討論ビュークライアント１０より発生する討論空間関連情報を集計するためのテーブルを持っており、本実施形態におけるこのテーブルは、図３７に示すように、各討論ビュークライアント１０毎に現在の討論空間関連情報がどのようになているかを保持していている。
【０１１３】
ステップｓ１１−００６では、討論空間関連情報ｒをこの表に反映させた後、この表全体から関連討論空間ＩＤ集合各々について、同一の集合がいくつ現れているか計算し、所定の数以上現れている集合は討論空間関連情報として取り出す。すなわち、多くの討論ビュークライアント１０で「関連有り」と判断されている討論空間集合の集合を取り出すわけである。この結果は変数ｓ１に代入される。
ステップｓ１１−００７において、再計算後の変数ｓ１が前回の関連討論空間情報を記憶している変数ｓと異なるかどうか検査され、異なる場合は、処理はステップｓ１１−００８に分岐する。ステップｓ１１−００８では、この新たな関連討論空間情報ｓ１を接続中の各討論ビュークライアント１０に送信し、そして、ステップｓ１１−００９で前回の関連討論空間情報ｓをｓ１に更新する。
【０１１４】
「第３実施形態の動作例」
図３８に第３実施形態の動作例を示す。
同図では、ユーザＡ、ユーザＢ、ユーザＣの３ユーザが本装置を使用中である。ユーザＢ、ユーザＣの興味の観点から見ると、「討論空間１」と「討論空間２」が関連性が高くなっている。討論空間関連生成手段３０は、これを集計し、ユーザＡにもこの集計結果をフィードバックする。ユーザＡの興味の観点では、「討論空間１」と「討論空間２」は関連しない状況ではあったが、図のように、新たに「討論空間２」との関連性が示される。これにより、ユーザＡは、現在他のユーザの興味の大勢は「討論空間１」と「討論空間２」との間に共通する「何か」にあることを知ることができ、新たに、思いがけなかった「討論空間２」への参加が促されることになる。
なお、この動作例では、図示の都合上、わずかユーザＢ、ユーザＣという２ユーザの興味の合致により、「討論空間１」と「討論空間２」が関連有りと判断されるように表現しているが、本来この判断には、もっと多くのユーザが必要であることは言うまでもない。
【０１１５】
【発明の効果】
以上説明したように、本発明は、部分討論空間の導入による適合討論部分の構造保存的抽出機能を基本機能とし、さらに、分散情報検索技術の導入により、広域分散環境下に散在している討論空間全体に対する現状把握および推移追跡機能をユーザに提供することができる。これにより、ユーザは広域分散環境下に散在している討論空間全体の状態把握および推移追跡を行うことができ、特に、意見の発せられた場所やそれまでの経緯の中で討論の現状を把握出来るようにする検索結果の構造およびユーザ・インターフェースが提供されることから、ユーザは討論の現状把握および推移追跡をより簡便かつより正確に行えるようになる。
【０１１６】
また、本発明は、ユーザの動向総体をフィードバックすることにより、「多くのユーザが興味を持っている討論空間」という、個人の興味を超越した「全体の興味の在処」という新たな観点からの現状把握機能をもユーザに提供することができる。これにより、ユーザ個人が興味を持っている討論空間と、多くのユーザが興味を持っている討論空間とが関連付けられることから、ユーザは討論の現状把握および推移追跡を全体動向を視野に入れたより大局的な観点から行えるようになる。
【図面の簡単な説明】
【図１】 第１の実施形態に係る電子的討論支援装置の構成を示す図である。
【図２】 第１の実施形態に係る討論空間の論理的構造を示す概念図である。
【図３】 第１の実施形態が提供するサービスの概念図である。
【図４】 討論空間のデータ構造の一例を示す図である。
【図５】 討論空間のデータ構造の一例を示す図である。
【図６】 討論空間のデータ構造の一例を示す図である。
【図７】 討論空間のデータ構造の一例を示す図である。
【図８】 論理演算「＊」の意味を説明する概念図である。
【図９】 論理演算「＋」の意味を説明する概念図である。
【図１０】 論理演算「＃」の意味を説明する概念図である。
【図１１】 討論空間検索要求の構成例を示す図である。
【図１２】 討論空間検索要求の構成例を示す図である。
【図１３】 討論空間検索制御手段の動作を表現するＰＡＤ図である。
【図１４】 討論空間検索制御スレッドの処理内容を表現するＰＡＤ図である。
【図１５】 検索状況記憶手段のデータ構造の一例を示す図である。
【図１６】 討論空間検索手段の動作を表現するＰＡＤ図である。
【図１７】 検索連携手段の動作を表現するＰＡＤ図である。
【図１８】 第１の実施形態の実施形態のユーザインターフェース画面を示す図である。
【図１９】 第１の実施形態の実施形態のユーザインターフェース画面を示す図である。
【図２０】 討論展望図関連ＵＩ手段の動作を表現するＰＡＤ図である。
【図２１】 討論展望図構成手段の動作を表現するＰＡＤ図である。
【図２２】 検索要求生成手段の動作を表現するＰＡＤ図である。
【図２３】 第１の実施形態の動作例を示す図である。
【図２４】 第１の実施形態の動作例を示す図である。
【図２５】 第１の実施形態の動作例を示す図である。
【図２６】 第２の実施形態が提供するサービスの概念図である。
【図２７】 第２の実施形態のユーザインターフェース画面を示す図である。
【図２８】 第２の実施形態における討論展望図関連ＵＩ手段の動作を表現するＰＡＤ図である。
【図２９】 第２の実施形態の動作例を示す図である。
【図３０】 第２の実施形態の動作例を示す図である。
【図３１】 第３の実施形態が提供するサービスの概念図である。
【図３２】 第３の実施形態に係る電子的討論支援装置の構成を示す図である。
【図３３】 討論展望図関連ＵＩ手段の動作を表現するＰＡＤ図である。
【図３４】 討論空間関連連携手段の動作を表現するＰＡＤ図である。
【図３５】 討論空間関連情報のデータ構造を示す図である。
【図３６】 討論空間関連生成手段の動作を表現するＰＡＤ図である。
【図３７】 討論空間関連状況のデータ構造を示す図である。
【図３８】 第３の実施形態の動作例を示す図である。
【符号の説明】
１・・・ＷＷＷサーバ、 ２、４・・・電子文書記憶手段、
３・・・ＤＢＭＳ、 ５・・・ネットワーク、
６・・・ＤＢＭＳゲートウエイ、 １０・・・討論ビュークライアント、
１１・・・ＷＷＷブラウザ、 １２・・・討論実行手段、
１３・・・討論展望図関連ＵＩ手段、 １４・・・展望図構成手段、
１５・・・討論空間検索要求生成手段、 １６・・・討論空間関連連係手段、
２０・・・討論ビューサーバ、 ２１・・・討論空間検索制御手段、
２２・・・検索状況記憶手段、 ２３・・・討論空間検索手段、
２４・・・検索連携手段、 ３０・・・討論空間関連生成手段、[0001]
BACKGROUND OF THE INVENTION
The present invention is an electronic debate for presenting and analyzing the current state of debate according to the user's interests and intentions in conjunction with an electronic debate system that debates by accumulating and browsing opinions in the form of electronic documents. The present invention relates to a support device.
[0002]
[Prior art]
In order to carry out joint work, a forum for discussion is indispensable. Conventionally, various electronic discussion systems have been considered.
Known techniques for electronic discussion systems include the following.
[0003]
"Network News System"
This is one of the electronic discussion systems built on a global scale, in which electronic documents are stored in multiple distributed news servers, and electronic documents are stored in buckets between servers. By using the relay method, electronic discussions can be performed on a global scale. Users express their opinions by inputting their opinions into the system in the form of electronic documents, and the opinions are handled in units of related opinions called news groups. In this system, there are various news groups according to the purpose and taste of various discussions, and the user can freely bystand and participate in discussions by selecting a news group.
[0004]
"WWW (World Wide Web)"
This is a system that can store and browse electronic documents on a global scale. In this system, it is possible to transmit not only browsing but also user input to a server by using a form. In recent years, there are countless "electronic bulletin boards" and "electronic conference rooms" using this function.
Although these electronic discussion systems are functionally diverse, users input opinions in the form of electronic documents, which are stored in the server, and further stored in the server in response to user requests. The basic function of discussing by browsing the opinions and adding new opinions is common.
[0005]
"Groupware products"
Many groupware products for supporting collaborative work are also used, and one of typical products of this type is Lotus Notes (trade name). This product is designed to perform collaborative work by storing and browsing the products obtained during the group's activities in the database in the form of electronic documents. This product is equipped with various services such as e-mail and document database for collaborative work. Of course, it also provides services for electronic discussions.・ It has a method of organizing opinions similar to the system.
[0006]
As described above, these conventional techniques provide a basic function for inputting an opinion, organizing the input opinion in some form, and browsing the organized opinion.
[0007]
Here, as the first issue in the electronic debate system, the debate is sometimes very complicated, and it is widely developed across various debate spaces so that it will continue to change on a large scale and dynamically. Therefore, there is a need for a means to accurately grasp the current state of the discussion.
In addition, as a second problem in the electronic discussion system, there is a need for a method in which a user searches for a discussion space according to his / her own interest from a variety of existing discussion spaces.
[0008]
For such a problem, first, as a first method for dealing with the first problem, there is a method using an overview of the discussion space. The above-mentioned prior art displays this structure by displaying an organized discussion space structure and attributes of opinions, and providing an interface that allows the user to roughly grasp the flow of discussion by viewing the display. It is addressed. For example, the reference structure is displayed in the form of a pseudo tree structure called “thread”, or the attributes of opinions such as the subject and the name of the speaker are displayed. The user can roughly grasp the current state of the discussion by seeing this.
[0009]
In addition, as a method of dealing with the second problem, there is a method of using the subject of the discussion space. In the above-mentioned conventional technology, a subject is usually defined for the discussion space, and the user participates in the discussion space with the expectation that the discussion in the discussion space is almost the subject. Addressed by operational technology.
[0010]
Furthermore, as a second method of dealing with the first problem, there is a method of using information retrieval based on the user's interest together. In the invention described in Japanese Patent Application Laid-Open No. 8-110892, it is possible to easily collect a large group of related electronic documents by retrieving, collecting and displaying artifacts (electronic documents) according to the user's interest from a database. Can be handled. Further, according to the invention described in Japanese Patent Laid-Open No. 9-6667, it is possible to search for a keyword that operates in conjunction with a network news system or the like and specifies a search range for an opinion that suits the user's interest.
[0011]
Furthermore, as a third method for dealing with the first problem, there is a method in which the evaluation by the user is fed back to each user to direct the discussion. In the invention described in US Pat. No. 5,317,683, a system that allows an examiner to select several items or add comments to the contents of a certain electronic document is introduced, and this evaluation is totaled. As a result, a point to be discussed later about the electronic document is automatically calculated and returned to the user. In other words, an agenda for the subsequent discussion is automatically generated, whereby the current state of the discussion is grasped and directed in accordance with the overall trend of each user.
[0012]
[Problems to be solved by the invention]
However, in the first method to deal with the first issue, it is necessary for the user to manually select and examine the discussion contents with reference to the rough structure and attribute information of the discussion, which is very useful for grasping the current situation of the discussion. There was a problem of cost.
[0013]
Further, in the second method for dealing with the first problem, the user's manual operation is somewhat reduced by using the information retrieval technique together.
However, since the invention described in JP-A-8-110892 is an invention for collaborative design support, a matching artifact is searched from a single artifact group in which dependency management is performed on a specific design model. Is. For this reason, there is no concept of multiple discussion spaces, and it is difficult to apply to the purpose of executing discussions in multiple discussion spaces and searching for such discussions. Does not work.
The invention described in Japanese Patent Application Laid-Open No. 9-6667 organizes a plurality of discussion spaces or searches for a suitable opinion from a plurality of discussion spaces. According to the present invention, a file for listing opinions of search results and a file for storing each opinion of search results linked from this file are obtained as search results. However, in such a search result structure, it is possible to directly obtain the structure information of which discussion space each relevance opinion included in the search result originally belongs to, and what other relevance opinion it belongs to. Cannot be expressed.
[0014]
In this way, with the prior art methods based on these information retrievals, it is possible to trace again how the retrieved opinions were described in the original discussion space, and what other opinions were accompanied by. It is difficult.
In discussions, it is important to have a place where an opinion is generated and the process leading to the statement of the opinion. Like these prior arts, collecting only opinions that match the user's intentions and interests makes discussions easier. It is not enough for grasping the current situation and tracking changes.
[0015]
In addition, the third method of dealing with the first problem can only deal with the preparation of a discussion about a single electronic document, and cannot evaluate the entire plurality of discussion spaces.
Also, the method of dealing with the second issue is a method that relies only on the operational skill that discussions will be conducted in accordance with the subject of the discussion, and the discussion content is not necessarily in the discussion space even now and in the future. There is no guarantee that it will fit the purpose. In addition, in the case of network news systems and groupware products, the logical location of the discussion space (for example, news group names and discussion database names) is clear, but on the World Wide Web. In the case of a discussion space, there are a myriad of discussion spaces distributed around the world, and in the first place it is difficult to find a discussion space with the subject desired by the user.
[0016]
The present invention has been made in view of the above-described conventional circumstances, and an object thereof is to provide a user with an automatic current state grasping and transition tracking function for the entire discussion space scattered in a wide-area distributed environment.
As described above, an important point in grasping the current situation and tracking changes in the discussion is that the place where the opinion was issued and the background up to that point can be reflected in the current situation grasping and transition tracking. The object of the present invention particularly includes the achievement of this problem as an important item.
Furthermore, it is possible to grasp the current state of discussions that have not yet been converted to opinions, by reflecting the current state and transition tracking of the current state according to the user's individual interests and intentions, and reflecting the overall trend of the user in the current state grasp and transition tracking. It is to be. The object of the present invention includes providing the user with such a more advanced current situation grasping and transition tracking function.
[0017]
[Means for Solving the Problems]
In order to achieve the above object, the present invention is based on a mechanism for extracting a relevant part of a discussion based on a search expression for searching a discussion space. Further, the present invention accepts input from a user and obtains an extraction result. By constructing an interface that can be visualized and the discussion can be traced from the visualized result, the present situation grasping and tracking function of the discussion is provided.
At that time, rather than simply collecting relevant opinions, by extracting relevant opinions with a structure divided for each discussion, it is important to grasp the current situation in the discussion and follow the transition, where the opinions were generated and The background can be reflected in the current status and tracking.
[0018]
In the electronic discussion support apparatus according to claim 1, a search expression for a discussion space (discussion management data and an electronic document group expressing an opinion organized by the discussion management data) is obtained from a plurality of discussion spaces. The electronic document extraction function unit extracts a matching part as a search result having a unique structure described below. This search result is obtained as a collection of a plurality of data in the form of search result management data, in which relevant opinion groups are divided and aggregated for each discussion space to which the opinion belongs. As a result, it is possible to obtain a search result in a form in which a certain opinion is described in which discussion space, and in addition, what other suitable opinions exist in the discussion space.
[0019]
The electronic discussion support apparatus according to claim 2 is provided with user interface means. Thus, the user can input the search formula according to his / her interest / intention, search the discussion space, and receive the display of the search result.
The electronic discussion support apparatus according to claim 2 is configured in a client-server format, the user interface means is configured by a client apparatus, and the electronic document extraction function unit is configured by a server apparatus.
[0020]
In the electronic discussion support apparatus according to claim 4, particularly when there are a plurality of electronic document extraction function units, a search expression is obtained by operating in cooperation with each electronic document extraction function unit. Search cooperation means for relay transfer between them, and the user interface means has a function of receiving search results from a plurality of electronic document extraction function sections and displaying them. Thereby, even in a discussion space dispersed in a wide area environment, it is possible to search and grasp the current situation all over.
[0021]
The electronic discussion support apparatus according to claim 5 further comprises a search formula storage means for permanently storing the search formula, and the electronic document extraction function unit performs a search based on the search formula in the search formula storage means. The display is repeated according to the repetition of the search by the electronic document extraction function unit. Thereby, even if the discussion space suitable for one's interest and the suitable part in them change along a time course, the user can track the change without effort.
[0022]
In the electronic discussion support apparatus according to the sixth aspect, the user interface means receives the input of the evaluation function for evaluating the search result, and displays the search result based on the result of evaluation using the evaluation function. Thereby, the user can see the search result more analytically, and can grasp the current situation more accurately than the result.
[0023]
In the electronic discussion support apparatus according to claim 7, when a search result of a plurality of users in use is examined and a set of discussion spaces having a high similarity is extracted in common by many users, the high similarity is obtained. A discussion-related generation unit that determines that a group of discussion spaces having a certain degree is related, and notifies each user interface unit as a determination result with the set of discussion spaces, and further includes a user interface unit Then, based on the judgment result of the related discussion space notified by the discussion related generation means, the display showing the relationship between the discussion spaces is performed. As a result, it is possible to reflect the overall trend of the user in grasping the current state of the discussion space, and it is possible to grasp the current state from the new viewpoint of “the location of the entire interest” that transcends the individual interest of the user.
[0024]
In the electronic discussion support apparatus according to the eighth aspect, the user interface means displays the search results as character strings or graphics, and displays these search results in a form prioritized by the evaluation function. Thereby, the user can grasp the search results visually in an easily understandable order.
[0025]
In the electronic discussion support apparatus according to the ninth aspect, the user interface means displays the search result as a character string or a figure in a multidimensional graph constituted by the axis of the value of the evaluation function. Thereby, the user can grasp the search result visually and in a more analytical manner.
[0026]
In the electronic discussion support apparatus according to the tenth aspect, the user interface means displays a list of relevance opinions corresponding to the search result in response to a user operation on the display expressing the search result. Thereby, the user can grasp not only information about where the discussion space is suitable but also what kind of opinion is present in the discussion space.
[0027]
In the electronic discussion support apparatus according to claim 11, the user interface means receives a user operation on the display expressing the search result, and the discussion of the discussion space in the discussion space corresponding to the search result is selected. Are displayed separately. As a result, the user can not only grasp the information about where the discussion space fits, and what kind of relevance opinion exists in the discussion space, but also how the relevance opinion is based on. It is possible to grasp whether the remark has been made.
[0028]
In the electronic discussion support apparatus according to the twelfth aspect, the user interface means displays an electronic document describing the contents of the opinion in response to the user's operation on the display expressing the opinion. Thereby, the user can smoothly transition from grasping the current state of the discussion and tracking the transition to grasping the content of the opinion, and can easily grasp the current state of the discussion and track the transition.
[0029]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments of an electronic discussion support apparatus according to the present invention will be described with reference to the drawings.
Note that each embodiment described below is an electronic document written in HTML (Hyper Text Markup Language) format using WWW (World Wide Web) technology, which is one of the world-wide environments for handling hypertext type data. The present invention is applied as a device that cooperates with an electronic discussion support system in a network environment in which a plurality of databases storing various types of electronic documents are widely distributed.
[0030]
FIG. 1 shows an example of the configuration and usage of the electronic discussion support apparatus according to the first embodiment.
As shown in the figure, the electronic discussion support apparatus according to the present embodiment includes a plurality of WWW servers 1, electronic document storage means 2 connected to each WWW server 1, a plurality of database management systems (hereinafter referred to as DBMS). 3. The electronic document storage means 4 connected to each DBMS 3 is connected by a network 5 and is implemented in a network environment in which electronic documents can be stored in a distributed manner.
The WWW server 1 cooperates with a database management system gateway (hereinafter referred to as a DBMS gateway) 6 connected thereto, so that an electronic document under the management of the DBMS 3 is also an ordinary Web document. It can be handled in the same way.
[0031]
A plurality of discussion view clients 10 and a plurality of discussion view servers 20 are connected to the network 5, and these discussion view server 20 and discussion view client 10 are the main components of the electronic discussion support apparatus according to the present invention. It is.
Each discussion view client 10 is provided with a WWW browser 11, and a user interface provided through each WWW browser 11 is used for a user to use such a network system.
These facilities are not special, and are generally widely used technologies, and the present embodiment is constructed on these existing technologies in order to effectively use existing assets. Of course, in the realization of the present embodiment, these maintenances are not indispensable, and may be replaced by other techniques that perform the same function.
[0032]
Each discussion view client 10 is provided with a discussion execution means 12, which operates in cooperation with the corresponding WWW browser 11 and allows users to access the discussion space. Provide opinion browsing and additional services. This part can also be easily realized by using any electronic conference system implemented on the normal WWW technology.
[0033]
The discussion view server 20 generates a discussion view for the electronic document group and the discussion space distributed and stored on the network 5 and provides them to the user. In order to realize this function, the discussion space search control is performed. Means 21, search status storage means 22, discussion space search means 23, and search cooperation means 24 are provided.
The discussion space search control means 21 and the discussion space search means 23 are combined to constitute an electronic document extraction function unit, and the search status storage means 22 constitutes a search expression storage means. The linkage unit 24 constitutes an electronic document extraction linkage unit. The detailed configuration of each part will be described later.
Each discussion view server 20 has a management target range specific to the WWW server 1 and the DBMS 3. That is, one discussion view server 20 does not process all electronic documents on the network 5, but a plurality of discussion view servers 20 cooperate to process electronic documents on the network 5.
[0034]
The discussion view client 10 provides an interactive interface to the user by being combined with the discussion execution means 12 and the WWW browser 11. The discussion view client 10 operates in conjunction with the normal WWW browser 11 and communicates with the actual user. For exchange, the WWW browser 11 serves as a window.
The discussion view client 10 includes a discussion view map related UI unit 13 as a discussion view front end, a view map configuration unit 14 as a discussion view configuration unit, and a discussion space search request generation unit 15. The detailed configuration of each part will be described later.
In the present embodiment, the discussion view server 20 and the discussion view client 10 are realized as a program executed by a computer, but their entire configuration or partial configuration may be realized by some hardware.
[0035]
"Structure of the discussion space in the first embodiment"
The concept of the logical structure of the discussion space in this embodiment is shown in FIG.
The discussion space is a management unit in which zero or more opinions that can be associated are grouped together. Further, this discussion space can add a relationship between a plurality of discussion spaces. The discussion space is realized as an electronic file having a special data structure on a database (electronic document storage means 2 and 4 managed by the WWW server 1 and DBMS 3), and details of the data structure will be described later.
[0036]
An opinion is a reference to an electronic document existing on the network 5. In this discussion support apparatus, the user creates an opinion in the discussion in the form of an electronic document, and the electronic document storage means 2, 4 on the network 5. You can express your opinion by storing it. The storage location of the electronic document that becomes the opinion does not necessarily exist in the electronic document storage means 2 and 4 in which the discussion space in which the opinion is described exists, and may be located anywhere. In the present embodiment, the position of the opinion entity is expressed using a URL (Uniform Resource Locator) generally used in the WWW technology.
Arbitrary attribute information can be added to each discussion space and opinion, which will be described in detail in the discussion space data structure described later.
[0037]
“Concept of Discussion View in the First Embodiment”
Prior to describing the detailed operation of this embodiment, the concept of the discussion view in this embodiment will be described.
As described above, the present invention provides the user with a service for grasping the current state of the discussion space and tracking the situation in the future. FIG. 3 shows a conceptual configuration of such a service in the present embodiment.
[0038]
On the user side, with respect to the discussion space, a search expression (query1, query2,...) Based on an attribute of interest is generated and sent to the discussion view server 20. Details of the structure of the search formula will be described later.
The discussion view server 20 uses the search formula received from the user to search the discussion space that is managed by each discussion view server 20, and specifies and specifies the portion of the discussion space that matches the search formula. A part of the discussion space is returned to the user as a search result with a unique structure described in the basic configuration. This unique data structure is hereinafter referred to as a partial discussion space.
Unless there is an instruction from the user, the search formula is permanently recorded in the search status storage means 22 of the discussion view server 20, and a re-search is performed as needed to provide the user with a search result. By such processing, the user can track and confirm the transition of the entire discussion space. The discussion view server 20 copies the retrieval formula received from the user to the other discussion view servers 20 so that a plurality of discussion view servers 20 distributed on the network 5 operate in a distributed and cooperative manner. The search result is provided to the user individually from each discussion view server 20.
[0039]
The user can grasp the current state of the entire discussion space by using the search result. As shown in FIG. 3, for example, a display entitled “Discussion perspective map for query 1” is output from the user interface. The In this discussion perspective map, icons representing partial discussion spaces that match the search formula of each discussion space are shown. The same applies to the display entitled “Discussion perspective map for query 2”.
Furthermore, the area (icon) that represents the partial discussion space is also an entrance to the partial discussion space or discussion space. For example, if you place the pointer on the icon that represents the partial discussion space in the figure and click the mouse button The user interface is configured such that a partial discussion space corresponding to this icon is displayed through a cooperative operation with the discussion execution means 12. Details of operations related to implementation examples and functions of this interface will be described later.
Note that the specific form (shape, color, etc.) of the function as the partial discussion space and the entrance to the discussion space depends on various user interface policies and designs. In the present invention, these specific designs are used. Is not specified.
[0040]
"Definition of discussion space"
Here, a formal definition of the discussion space will be described later as preparation for defining the meaning of the discussion search formula in the present embodiment. This is one of the formalizations of the discussion space in the usual electronic discussion system.
The discussion space d is defined as d = <idd, Pd, Rd, Ad>, and is composed of four terms. The discussion space d corresponds to the discussion management data in the basic configuration.
Here, idd is a discussion space ID of the discussion space d, and is a set of idd ∈ discussion space IDs (a set of discussion space IDs is a set of all ID numbers that can be taken as ID numbers of the discussion space). . This is one of the attributes of the discussion space described below, but is described separately in consideration of the convenience of processing and notation.
[0041]
Pd is an attribute set of the discussion space d, and is a set of the form Pd = {<n, v> | n∈any set of character strings, v∈any set of character strings∪a set of real numbers}. This is a set of pairs <n, v> consisting of an attribute name n consisting of an arbitrary character string and a value v of that attribute.
Rd is a related set of the discussion space d, and is a set of the form Rd = {<id, v> | id∈discussion space ID set, v∈arbitrary character string set}. This is a set of a pair <id, v> of the ID number id of the discussion space to which the discussion space d is related and the character string v representing the type of the relationship.
[0042]
Ad is an opinion set Ad = {a1,..., An} consisting of opinions ai belonging to the discussion space d.
Opinion a is defined as a = <ida, Pa, Ra, Ba>, and is composed of the following four term groups. In this embodiment, it is assumed that attribute information of an opinion, related information to another opinion, and an electronic document representing the text of the opinion are associated in this way. It doesn't matter how you do it.
Here, ida is an opinion ID of opinion a, and is a set of idaε opinion IDs (a set of opinion IDs is a set of all ID numbers that can be taken as IDs of opinions).
[0043]
Pa is an attribute set of opinion a, and is a set of the form Pa = {<n, v> | n∈any set of character strings, v∈any set of character strings∪a set of real numbers}. This is a set of a pair <n, v> consisting of an attribute landmark n consisting of an arbitrary character string and a value v of that attribute.
Ra is a related set of opinion a, and is a set of the form Ra = {<id, v> | id∈set of opinion ID, v∈set of arbitrary character string}. This is a set of a pair <id, v> of an ID number id of an opinion related to the opinion a and a character string v representing the type of the relationship.
Ba is an opinion text, and the form of Ba may be any form. In the system used in the present embodiment, Ba is a URL indicating the electronic document of the opinion entity, as will be described later.
[0044]
"Data structure of discussion space"
4, FIG. 5, FIG. 6, and FIG. 7 show the data structure of the discussion space. This data structure is an example of the expression for the formal definition of the discussion space described above.
These data structures are stored in one of the electronic document storage means 2 and 4 in the form of an electronic file, and are managed by the corresponding specific discussion view server 20.
The “discussion space management table” having the structure shown in FIGS. 4 to 7 is the highest-level data structure handled by the discussion view server 20, and each line of the discussion space management table represents one discussion space. The discussion space management table is composed of four groups of “discussion space ID”, “attribute list”, “related list”, and “opinion list”, each of which is the “discussion space ID” of the above-described formal definition. ”,“ Attribute set (in discussion space) ”,“ related set (in discussion space) ”, and“ opinion set ”.
[0045]
“Discussion space ID” is an identification number of the discussion space uniquely determined on the network 5. There are various methods for generating a unique number on such a network. In the present embodiment, a unique number is generated using a host name and a random value. Of course, other methods may be used.
As shown in FIG. 4, the “attribute list” is a reference indicating a list of attribute names and attribute pairs related to each discussion space. The “attribute list” includes any attribute name and its attribute name. You can attach attribute values.
[0046]
As illustrated in FIG. 5, the “related list” is a reference that indicates a list that includes a pair of discussion space IDs and related types of other discussion spaces related to each discussion space. You can attach any other discussion space and its associated content related to one discussion space.
As shown in FIGS. 6 and 7, the “opinion list” is a reference that designates an “opinion table” that is a table for managing the opinions given in each discussion space.
[0047]
6 and 7 illustrate this “opinion management table”, and each row of the “opinion management table” corresponds to one opinion.
As shown in the figure, the opinion is composed of four sets of “opinion ID”, “attribute list”, “related list”, and “opinion reference”, each of which is an “opinion ID” of the formal definition described above. , “(Opinion) attribute set”, “(opinion) related set”, and “opinion text”.
[0048]
The “opinion ID” is an identification number of an opinion uniquely determined on the network 5. There are various methods for generating a unique number on such a network. In the present embodiment, a unique number is generated using a host name and a random value. Of course, other methods may be used.
As shown in FIG. 6, the “attribute list” is a reference indicating a list composed of attribute names related to each opinion and a pair of the attributes. The “attribute list” includes an arbitrary attribute name and its attribute. Can be priced.
[0049]
As illustrated in FIG. 7, the “related list” is a reference indicating a list including pairs of opinion IDs of other opinions related to each opinion and their related types, and a discussion in the “related list”. You can attach any other discussion space related to the space and its related content.
The “opinion reference” is an indicator indicating the position of the entity of the electronic document that is an opinion, and a URL is used in this embodiment. Thereby, the electronic document on the electronic document storage means 2 managed through the WWW server 1 and the DBMS gateway 6 operating in conjunction with the WWW server is designated.
[0050]
In the present embodiment, the logical structure of the discussion space shown in FIG. 2 is expressed by the data structure as described above. Note that the generation / maintenance of such a data structure is outside the scope of the present invention, and thus description of the generation / maintenance of such a data structure is omitted.
That is, the gist of the present invention is how to construct a service for providing a discussion perspective map for a data structure generated and maintained in such a structure.
[0051]
"Structure of discussion space search formula"
In this embodiment, the discussion space retrieval formula has the following configuration.

[0052]

Attribute name :: = Element of a set of arbitrary character strings
Attribute value :: = Arbitrary string set element | Arbitrary constant set element
Upper limit :: = constant | system constant
Lower limit value :: = constant | system constant
Association type :: = Elements of a set of arbitrary character strings
Keyword :: = elements of a set of arbitrary character strings
[0053]
The above contents express the grammar of the search expression for the discussion space in this embodiment. For example,
dispace-attrib (total number of opinions, [0,1000]): article-attrib (speaker, {Kou, Otsu}), article-attrib (speaker, {丙, Ding})
Is a search expression.
The meanings of the partial discussion search predicate and the full discussion search predicate are defined as follows.
[0054]
article-attribute (opinion attribute name: [lower limit, upper limit]) =
{<Iddi, Pdi, Rdi, {aj | aj ∈ Adi, <n ′, v> ∈ Paj, n ′ = n, l <= v <= h}> | di∈D}
However, n = opinion attribute name, l = lower limit value, h = upper limit value, D = a set of all discussion spaces under discussion view server management. This represents “a set of discussion spaces having an opinion set in which only opinions whose attribute value indicated by the opinion attribute name n is between the lower limit value and the upper limit value are extracted”. For example, it can be used for the purpose of extracting only opinions from a certain time to a certain time. The upper limit value and the lower limit value can be arbitrary real values, and special constants prepared by the system can be used. For example, various constants such as a constant representing the current time and a constant representing a time n hours before the current time can be used. A discussion space having an opinion set in the form of an opinion set having a certain property is called a “partial discussion space”.
[0055]
article-attribute (opinion attribute name: {attribute value 1,..., attribute value m}) = {<iddi, Pdi, Rdi, {aj | aj ∈ Adi, <n ′, v> ∈ Paj, n ′ = n, vεV}> | diεD}
However, n = opinion attribute name, V = {attribute value 1,..., Attribute value m}, D = a set of all discussion spaces under discussion view server management. This is a “discussion space having an opinion set in which only an opinion whose attribute value indicated by the opinion attribute name n is an element of the attribute value set {attribute value 1,..., Attribute value m}” is extracted. Represents a set. For example, it can be used for the purpose of extracting only opinions that the speaker is this person or that person.
[0056]
article-ref ({association type 1,..., association type m}) =
{<Iddi, Pdi, Rdi, {aj | ajjAdi, <id, v> εRaj, vεV}> | diεD}
However, V = {related type 1,..., Related type m}, D = a set of all discussion spaces under discussion view server management. This represents “a set of discussion spaces having an opinion set from which only opinions whose relation type is an element of {relation type 1,..., Relation type m}” are extracted. For example, it can be used for the purpose of extracting only opinions whose relation is “question” or “response”.
[0057]
article-body (keyword) =
{<Iddi, Pdi, Rdi, {aj | ajjAdi, Baj contains k}> | di∈D}
However, k = keyword, D = a set of all discussion spaces under discussion view server management. This represents “a set of discussion spaces having an opinion set in which only opinions including keywords are included in the opinion text”.
[0058]
dspace-attib (discussion space attribute name: [lower limit, upper limit]) =
{Di | di.epsilon.D, <n ', v> .epsilon.Pdi, n' = n, l <= v <= h}
However, n = discussion space attribute name, l = lower limit value, h = upper limit value, D = a set of all discussion spaces under discussion view server management. This represents “a set of discussion spaces in which the value of the attribute value indicated by the discussion space attribute name n is between the lower limit value and the upper limit value”. For example, the discussion start time can be used for extracting only opinions from a certain time to a certain time. The upper limit value and lower limit value can be arbitrary real values, and special constants prepared by the system can be used. For example, various constants such as a constant representing the current time and a constant representing a time n hours before the current time can be used.
[0059]
dspace-attribute (discussion space attribute name: {attribute value 1,..., attribute value m}) =
{Di | di∈D, <n ′, v> ∈Pdi, n ′ = n, v∈V}
However, n = discussion space attribute name, V = {attribute value 1,..., Attribute value m}, D = a set of all discussion spaces under discussion view server management. This represents “a set of discussion spaces in which the attribute value indicated by the discussion space attribute name n is an element of a set of attribute values {attribute value 1,..., Attribute value m}”. For example, it can be used for the purpose of extracting only the discussion space where the organizer is this person or that person.
[0060]
dspace-ref ({association type 1,..., association type m}) =
{Di | di∈D, <id, v> ∈Rdi, v∈V}
However, V = {related type 1,..., Related type m}, D = a set of all discussion spaces under discussion view server management. This represents “a set of discussion spaces whose related types are elements of {related type 1,..., Related type m}”. For example, it can be used for the purpose of extracting only the discussion space whose relation is “subordinate discussion”.
[0061]
Further, the meaning of logical sum (+), logical product (*) and logical difference (#) is defined as follows.
D1 * D2 =
{<Iddi, Pdi, Rdi, Adi∩Adj> | di∈D1, dj∈D2, iddi = iddj},
As shown in FIG. 8, this is a new form in which only the discussion space having the same ID included in both of the discussion space sets D1 and D2 is left with only the opinions included in both. This operation generates a set of discussion spaces.
[0062]
D1 + D2 =
{<Iddi, Pdi, Rdi, Adi∪Adj> | di∈D1, dj∈D2, iddi, = iddj} ∪ {di | di∈D1, dj∈D2, iddi ≠ iddj} ∪ {dj | di∈D1, dj∈D2, iddi ≠ iddj},
As shown in FIG. 9, this is a new discussion space in which the opinions included in both of the discussion spaces having the same ID included in both of the discussion space sets D1 and D2 are combined. This is an operation for generating a set of discussion spaces that is a combination of a set and a set of discussion spaces included only in one of them.
[0063]
D1 # D2 =
{<Iddi, Pdi, Rdi, Adi \ Adj> | di∈D1, dj∈D2, iddi = iddj} ∪ {di | di∈D1, dj∈D2, iddi ≠ iddj},
This is because, as shown in FIG. 10, regarding the discussion space having the same ID included in both of the discussion space sets D1 and D2, the opinion included in D2 is deleted from the opinions included in D1. This is an operation for generating a set of discussion spaces, which is a combination of a set of new discussion spaces of the form described above and a set of discussion spaces included only in D1.
[0064]
Further, the meanings of the qualifier (:) and the connector (,) are defined as follows.
D1: D2 is D1 * D2, which allows the discussion space set D2 to be limited by a logical product with D1.
D1 and D2 are D1 + D2, and by this, sets D1 and D2 of the discussion space can be connected by logical sum.
[0065]
"Structure of discussion space search request"
The configuration of the discussion space search request (query) described above will be described with reference to FIGS.
A normal discussion space search request has a configuration as shown in FIG. 11, and includes a search request ID, a processing result return destination, a search request propagation path, a search request type, and a search expression.
The search request ID is an identification number of a search request that is uniquely determined on the network 5 and can be generated using a host name, time, or random number.
The processing result reply destination is a target to which the discussion view server 20 sends a discussion view (discussion perspective map), and indicates the discussion view client 10 that has transmitted a discussion space search request. This processing result reply destination information can be uniquely determined on the network 5 using a host name, a port number, or the like.
[0066]
The search request propagation path is a delivery history when a discussion space search request is delivered among a plurality of discussion view servers 20, and is a list composed of the host names of the discussion view servers 20 through which the discussion space search request has passed.
The search request type is information indicating the type of processing, and a symbol indicating “search” is stored in the normal discussion space search request shown in FIG.
The search formula stores the discussion space search formula itself.
[0067]
The discussion space search request shown in FIG. 12 is a special discussion space search request for canceling the request for the discussion space search requested in the past, and the search request ID, processing result return destination, and search request propagation path are the normal discussion space. This is the same as the search request.
As the search request type, a symbol representing “cancel” is stored when the discussion space search is cancelled.
Further, in the cancellation of the discussion space search, a search request ID to be canceled is added instead of the search formula, and the discussion space search request to be canceled is specified by this ID information.
[0068]
"Operation of Discussion View Server"
The discussion view server 20 in this embodiment operates as follows.
FIG. 13 is a PAD representing the operation of the discussion space search control means 21. This discussion space search control means 21 receives a discussion space search request via the network 5, and performs a search request registration or cancellation process depending on the type of the request. In addition, the discussion space search control means 21 passes a search request to the search cooperation means 24 so as to cooperate with other discussion view servers 20, or newly creates a dedicated thread for performing a search in response to the search request periodically. It also performs processing such as generating.
[0069]
That is, as shown in FIG. 13, the discussion space search control means 21 executes the procedure “discussion space search control” at the start of execution, and this procedure immediately follows an infinite loop for event processing in step s1-001. to go into. In this loop, event e is acquired in step s1-002. If the acquired event e is “arrival of new search request”, that is, the discussion space search control means 21 receives the search request, step s1-003 is executed. From step s1-004.
[0070]
In step s1-004, the received search request r is extracted, and in step s1-005, whether or not this search request r is already registered in the search status storage unit 22 is used by using the search request ID included in the search request. Judgment. If the search request r is not registered in the search status storage means 22, the process proceeds to s1-006 and subsequent steps, and in step s1-006, the search request r is registered in the search status storage means 22, and The search status corresponding to the search request r is initialized. The search situation will be described later.
[0071]
In steps s1-007 and s1-008, the search request ID of the search request r is extracted, and a “search control thread” using the ID as an identification number is newly generated. In this process, a thread for performing a process written in a procedure “search control thread” to be described later is formed (the process by this thread is a process that is processed in parallel independently of other processes).
Further, in step s1-009, the search request r is passed to the search cooperation unit 24, thereby causing a cooperation operation between the plurality of discussion view servers 20 to be performed. The operation of the search cooperation unit 24 will be described later.
[0072]
On the other hand, when the acquired event e is “arrival of search cancellation request”, that is, when the discussion space search control means 21 receives the search cancellation request, the process branches from step s1-010 to the processing after step s1-011.
In step s1-011, the received search cancellation request r is extracted, and in step s1-012, the search request ID to be canceled is extracted from the search cancellation request r as a variable id. In step s1-013, it is determined whether or not the search status corresponding to the search request ID exists in the search status storage unit 22, and if it exists, step s1-013 to step s1-014 and subsequent steps are determined. Processing branches to processing, and in steps s1-014 and s1-015, the search status corresponding to id is deleted from the search status storage means 22, and the search control thread corresponding to id is deleted.
In step s1-016, the search cancellation request r is passed to the search cooperation unit 24 to perform a cooperative operation between the plurality of discussion view servers 20. The operation of the search cooperation unit 24 will be described later.
[0073]
FIG. 14 shows the processing contents of the “search control thread” procedure described above, and this search control thread is generated with a search request ID in a variable id.
First, in step s2-001, the search formula q is extracted from the search request corresponding to id by the search status storage means 22, and in step s2-002, the processing result reply destination is searched from the search request corresponding to id from the search status storage means 22. b is extracted, and an infinite loop for search control is entered in step s2-003.
In the loop, after waiting for a certain time (step s2-004), the discussion space search means 23 is searched using the search expression q to obtain the search result a (step s2-005). The process of updating the search status on the status storage means 22 (step s2-006) and returning the search result a to the processing result return destination b (s2-007) is repeated.
[0074]
FIG. 15 shows the data structure of the information stored in the search status storage means 22.
The search status storage means 22 stores a list structure based on a search request and a search result for the search request as shown in the figure, and a search result pair for the search request is searched for. Is the situation.
In the search request, the discussion space search request itself is stored, and in the search result, only the portion satisfying the search request in the entire discussion space is stored with the same data structure as the discussion space. The search result returned to the reply destination is transmitted to the reply destination using this form of data in an appropriate form.
[0075]
FIG. 16 is a PAD diagram representing the operation of the discussion space search means 23. This procedure “discussion space search” is called from the search control thread with the search expression q. First, in step s3-001, it is determined whether the search expression q is grammatically correct. This can be done in light of the grammar rules of the search expression described above. If it is not correct, an error is returned in step s3-002 and the procedure is terminated.
On the other hand, if it is correct, the variable a for storing the search result is linked to the empty list in step s3-003, and the pointer dp pointing to the row of the discussion space management table is pointed to the head of the discussion space management table in step s3-004. Initialize to.
[0076]
Then, in step s3-005, a loop that scans the rows of the discussion space management table from the beginning to the end is entered. Is done. For this determination, each attribute value of the “attribute list” linked to the row of the discussion space management table pointed to by the pointer dp, the value of the related type of the “related list”, and the qualifier “:” appearing in q: All discussion search formulas specified on the left side of "" are used.
If the row of the discussion space management table pointed to by dp matches even one of all the discussion search expressions appearing in the search expression q, it is determined that this discussion space is suitable. However, if the left-side value of the qualifier “:” is empty, it always matches. Note that the match of all discussion space search expressions is determined based on the meaning of the discussion space search expression described above. That is, the determination is made based on the match of the attribute value and the relation type value that each all discussion search predicate means and the meaning of the logical operation between these predicates.
[0077]
If the discussion space pointed to by dp matches, the process branches to step s3-007. In step s3-007, a variable q0 is substituted with a reconstructed search expression q in a form that extracts only the discussion search term that matches the row of the discussion space management table pointed to by dp. In step s3-008, a new empty list is generated and linked to the variable a0. Next, in step s3-009, a pointer pointing to the head of the opinion management table linked to the row of the discussion space management table pointed to by the variable dp is substituted into the variable ap.
[0078]
In step s3-010, a loop for scanning the rows of the opinion management table from the beginning to the end is entered, and in step s3-011, the opinion management table pointed to by ap or the opinion text traced from the opinion management table matches the search expression q. It is determined whether to do it. For this determination, each attribute value of the “attribute list” linked to the row of the opinion management table pointed to by the pointer ap, the value of the related type of the “related list”, the opinion text, and each discussion search of the search expression q0. The partial discussion search expression specified on the right side of the qualifier “:” appearing in the term is used.
If the row of the opinion management table pointed to by ap and the comment text to be traced match even one of the partial discussion search expressions appearing in the search expression q0, this opinion is determined to be suitable. However, if the right side value of the qualifier “:” is empty, it always matches. The match of the partial discussion space search expression is determined based on the meaning of the above-described discussion space search expression. That is, the determination is made based on the match of the attribute value, the relation type value, and the keyword that each partial discussion search predicate means, and the meaning of the logical operation between these predicates.
[0079]
If the opinion pointed to by ap matches, the process branches to step s3-012, and the line of the opinion table pointed to by ap is added to the end of the list a0. In step s3-013, the pointer ap indicating the row of the opinion management table is advanced, and the process returns to the top of the loop s3-10. This loop ends when it reaches the end of the opinion management table, and the process branches to step s3-014.
In step s3-014, if a0 is not an empty list, that is, if any relevance opinion is found, the process branches to step s3-015, and a search result is configured. In step s3-015, the line of the discussion space management table pointed to by dp is duplicated and linked to the variable al. In step s3-016, the opinion list of the variable al is connected to the variable a0 which is a list of relevant opinions. As a result, a new discussion space management table row al that represents a discussion space having the opinion list a0 consisting only of conforming opinions as an opinion list is constructed. In step s3-017, the row al of this new discussion space management table is added to the search result a.
[0080]
In step s3-019, the pointer dp pointing to the row of the discussion space management table is advanced, and the process returns to the top of the loop s3-005. This loop ends when it reaches the end of the discussion space management table, and returns a result a in step s3-019.
Thus, the search result in this example is in the form of extracting a suitable opinion group for each suitable discussion space and connecting it to the list.
[0081]
FIG. 17 is a PAD diagram representing the operation of the search cooperation means 24. The search cooperation control procedure in this figure is called from the discussion space search control means together with the search request r.
First, in step s4-001, the search request propagation path length of the search request r is equal to or less than a predetermined length (hop number limit: this is predetermined for each discussion view server), and the search request When the own host name is not included in the propagation path length, the processing branches off to step s4-002 and thereafter. The purpose of this condition determination is to prevent the search request from being traversed on the network 5 without disappearing indefinitely.
[0082]
In step s4-002, the local host is added to the search request propagation path field of the search request r to obtain a new search request r ′. In steps s4-003 and s4-004, the search request r ′ is transferred to each of the predetermined transfer destination hosts (this is predetermined for each discussion view server). The partner that receives the search request r ′ is the discussion space search control means of the partner host d.
[0083]
"Discussion View Client Behavior"
As shown in FIG. 1, in the electronic discussion support apparatus of this embodiment, the discussion view client 10 operates in cooperation with the WWW browser 11 and the discussion execution means 12. This discussion execution means 12 is a mechanism for reading opinions in the discussion space and placing new opinions in the discussion space, and is a part of the infrastructure of the electronic conference system to which the present invention is applied as an example. It is.
[0084]
FIG. 18 shows an example of a user interface viewed from the WWW browser 11 in the present embodiment, where F1 and F2 are outputs of the discussion execution means 12, and F3 is a user managed by the discussion view client 10. Interface.
In the state shown in FIG. 18, in the display area F3, the entrance to the current partial discussion space that matches the “search expression” is displayed with icons “discussion space A”, “discussion space B”, and “discussion space C”. The “discussion space A” is selected by the user, and the discussion space including the partial discussion space selected in the display region F3 is displayed in the display region F2. The portion indicated by the dotted line in the display area F2 is an opinion that is not included in the partial discussion space of the search results in the discussion space, and the portion indicated by the solid line is the search result. It is an opinion included in the partial discussion space. In this display area F2, the discussion execution means 12 compares the opinions included in the partial discussion space of the search result passed from the discussion view client 10 with all the opinions in the discussion space under the partial discussion space. It is generated with.
In the state shown in FIG. 18, the opinion shaded in the display area F2 is selected by the user, and the content of the opinion is displayed in the display area F1.
[0085]
The discussion view client 10 cooperates with the WWW browser 11 and the discussion execution means 12, and corresponds to the entrance of the partial discussion space when the user inputs a predetermined action such as double-clicking the partial discussion space icon in the display area F3. The WWW browser 11 and the discussion execution means 12 can be instructed to display the discussion space and the partial discussion space displayed in the display area F2 as shown in FIG.
Also, a plurality of “search formulas” can be provided, and as shown in FIG. 19, a plurality of search formulas can be displayed in the pop-up list PL and selected from them. Note that what is not currently displayed in the area F3 is always held and updated internally.
[0086]
FIG. 20 is a PAD diagram showing the operation of the discussion perspective related UI means 13.
First, in step s5-001, an infinite loop for waiting for an event is entered. In step s5-002, an event is waited for, and that event is substituted into the result e. The event is generated from the WWW browser 11 in accordance with the user's operation or when a search result is given from the discussion view server 20 via the network 5.
[0087]
Next, in step s5-003, if the event e is a request for access to the partial discussion space, the process branches to step s5-004. This event occurs when the user takes a predetermined action such as double-clicking the entrance of the partial discussion space on the display area F3 shown in FIG.
In this case, the partial discussion space specified by the user action is specified in step s5-004, and further, the partial discussion space is displayed to the discussion execution means 12 in step s5-005. As a result, the corresponding partial discussion space is displayed on the display area F2 shown in FIG. As a result, the user can quickly access the part of the debate in line with his / her interest.
[0088]
On the other hand, if it is determined in step s5-006 that the event e is a request for changing the display discussion perspective view, the process proceeds to step s5-007. This event occurs when the user changes the “search expression” currently selected on the pop-up list PL as shown in FIG.
In this case, in step s5-007, the search request ID for the selected “search expression” is extracted as a variable id. Here, the discussion view client 10 holds all the discussion space search expressions and the search results (discussion views) that have been designated by the user since the start. A key for identifying a plurality of search expressions and search results is based on a search request ID generated when a search request is issued.
In step s5-008, the search result v {id} corresponding to the search request ID is extracted from the search request ID and displayed again in s5-008. Note that v {id} is an associative array variable that stores the search result using the search request ID number as a key.
[0089]
On the other hand, if the event e is a discussion perspective map update request in step s5-009, the process proceeds to step s5-010. This event occurs when the perspective map constructing means 14 receives a search result.
In this case, in step s5-010, the search request ID is extracted as in step s5-007, and in step s5-011, the search request ID and the search result are combined and stored. Further, in steps s5-012 and s5-013, if the discussion view corresponding to the search request ID is being displayed, the display is also updated.
[0090]
On the other hand, if it is determined in step s5-014 that the event e is a discussion space search registration request, the process proceeds to step s5-015. This event occurs when the user enters a new search expression in the “search expression” field and presses the “register” button.
In this case, in step s5-015, the search expression q to be registered is acquired, and in step s5-016, the discussion space search request generation unit 15 is instructed to issue a search request for the search expression q.
[0091]
On the other hand, if the event e is a request to stop the discussion view client 10 in step s5-017, the process proceeds to step s5-018. This event occurs when the WWW browser 11 cancels the display of the discussion view client 10 or the browser 11 itself is terminated according to a user instruction.
In this case, in step s5-018, the discussion space search request generation unit 15 is instructed to issue a search cancellation request for all the discussion space search requests held by the discussion view client 10.
[0092]
FIG. 21 is a PAD diagram showing the operation of the perspective view constructing means 14.
This perspective map constructing means 14 monitors the network 5, operates when a search result is returned from the discussion view server 20, and immediately enters an event waiting loop at s6-001. In step s6-002, the event e is acquired. In step s6-003, if the event e is a response to the discussion space search result, an action is taken.
[0093]
Then, in step s6-004, the retrieval result a is extracted from the returned result, and based on this, in step s6-005, a discussion perspective view is constructed and the result is substituted into v. In the present embodiment, the discussion perspective view does not have a structure, and therefore, the processing is simply performed by arranging the entrances to the partial discussion space according to some geometry management policy. Of course, other configurations may be performed.
In step s6-006, the discussion perspective map-related UI means 13 is requested to update the discussion perspective map to v. As described above, this request is acquired as an event in the discussion perspective map-related UI means 13 and updated.
[0094]
FIG. 22 is a PAD diagram showing the operation of the discussion space search request generation means 15.
As shown in FIG. 22A, the procedure “issue discussion space search request” is called from the discussion perspective map-related UI means 13 with the search expression q. In step s7-001, a search request r is generated by adding a field to the search expression q as shown in FIG. 11, and in step s7-002, the discussion view server 20 set by the user is generated. This search request r is transmitted.
Also, as shown in FIG. 22B, the procedure “Issuance of Cancellation of All Discussion Space Search Request” is also called from the discussion perspective map related UI means 13. In step s7-003, a search cancellation request as shown in FIG. 12 is generated for all search requests held by the discussion view client 10, and transmitted to the discussion view server 20 set by the user.
[0095]
"Operation example"
FIG. 23, FIG. 24, and FIG. 25 show operation examples in the present embodiment.
It is assumed that the user's interest is in the persons “speaker A” and “speaker B” who are key in many discussions, and furthermore, it is desired to know the trends of these persons over the last three hours.
In this case, the user expresses his / her interest by a search expression and transmits it to the discussion view server 20 using the discussion view client 10. The search formula in this case is “article-attrib (opinion speech time, [3 hours ago, current time]) * article-attribute (speaker, {speaker A, speaker B})”.
[0096]
FIG. 23 shows an initial state in which a portion that matches the user's interest is specified, and the user can easily access the portion through the discussion perspective view displayed on the user interface.
Furthermore, in the present embodiment, the object of interest of the user is tracked over time, so that the latest situation of the discussion that changes as the discussion progresses can be easily grasped without the user struggling.
FIG. 24 shows a situation when a little time has passed from the state shown in FIG. 23. As the discussion proceeds, the partial discussion space that matches the user's interest changes. In addition, with the passage of time, a “discussion space 2” that has not been related to the user's interest until now appears as a new object of the user's interest.
FIG. 25 shows a situation when a little more time has passed than the state shown in FIG. 24, and the partial discussion space that matches the user's interest is changed.
According to the present invention, such a dynamically changing situation of the discussion can be accurately tracked without the user's operation, and can be easily accessed as necessary.
[0097]
Next, a second embodiment according to the present invention will be described.
The configuration of the apparatus and the structure of the discussion space in this embodiment are the same as those in the first embodiment. However, the concept of the discussion perspective map is slightly different, so it will be explained.
[0098]
“Concept of Discussion View in Second Embodiment”
The conceptual configuration of the service and discussion perspective view in this embodiment is as shown in FIG. 26. The difference from the first embodiment is that not only the search expression (query) but also the coordinate axis (axis) is used as the user's intention. This is possible. As the coordinate axis, for example, numerical information such as a time axis and the number of utterances, character string information such as a speaker name and a related type, and the like can be freely set.
With the introduction of the coordinate axes, the first embodiment has no structure in the discussion perspective view, but in this embodiment, the structure along the designated coordinate axes can be performed. As a result, the position of each partial discussion space in the discussion perspective map can be clarified, and the user can grasp the current state of the discussion more accurately.
In the present embodiment, unlike the first embodiment, the discussion perspective map constructing means 14 constructs a discussion perspective map based on the coordinate axes, and is partially modified. This changed part will be described below.
[0099]
"Discussion View Client Behavior"
An example of a user interface viewed from the WWW browser 11 in this embodiment is shown in FIG. Unlike the first embodiment, the interface design of the region F3 in FIG. 27 is changed. In the first embodiment, the entrance of the current partial discussion space that matches the “search expression” is expressed without a structure. However, in this embodiment, the entrance icon of the partial discussion space has two coordinate axes ( axis) is mapped onto a two-dimensional graph.
Selection of a coordinate axis means inputting an evaluation function for evaluating the search result and mapping it to the coordinate axis. Of course, since the essence of this embodiment is that the partial discussion space as a search result can be analyzed from various perspectives using some sort of evaluation scale called coordinate axes, the three-dimensional display and the multi-dimensional display can be reduced as an expression form. Various other forms such as those obtained may be used.
[0100]
FIG. 28 shows a PAD diagram expressing the operation of the discussion perspective view related UI means 13 in the present embodiment. The difference from the first embodiment in this operation is that the discussion view map redisplay procedure (steps s8-008, s8-013, s8-021) is changed to display the structure in accordance with the coordinate axes. Is a point.
In this embodiment, the attribute of the partial discussion space, the attribute of the opinion included in the partial discussion space, or a value obtained by evaluating them with some function is simply mapped and displayed on the coordinate axis designated by the user. In the present embodiment, since the graph is a two-dimensional graph, the region to be mapped constitutes a two-dimensional figure. An example of this display will be described later.
[0101]
The difference from FIG. 28 in the first embodiment is that a coordinate axis change request is added to the acceptable event e (step s8-19).
This event occurs when the user changes the coordinate axis by the coordinate axis pop-up list L as shown in FIG. In this case, the process branches to steps s8-20 and s8-21, and the discussion perspective map is redisplayed according to the new coordinate axes.
[0102]
"Operation example of the second embodiment"
29 and 30 show an operation example in this embodiment.
Suppose that the user's interest wants to know the trend of discussions over the last three hours, and also wants to know how the remarks are related.
In this case, the user first expresses his / her interest by a search expression, and transmits this to the discussion view server 20 using the discussion view client 10. In this case, the search formula is “article-attrib (opinion speech time, [3 hours ago, current time])”.
Then, the user determines the coordinate axis for displaying the discussion perspective map. In FIG. 29, “speech time is within the last 3 hours” is selected as the search formula, and “opinion speech time” and “opinion related type” are selected as the coordinate axes.
[0103]
The state shown in FIG. 29 is the initial state and shows the discussion situation at the time of 11:00. In the present embodiment, in the display of the discussion perspective map, a rectangle representing the entrance of the partial discussion space indicates the range of attribute values of the corresponding discussion partial space. For example, at the entrance to the partial discussion space of “discussion space 1” displayed as “Dspace-1”, the oldest speech time in this partial discussion space is 9:00, and the latest speech time is 10:30. It is shown that. Also, when looking at the relation type of opinion, it indicates that a question and an answer were made.
The state shown in FIG. 30 is a state when time has passed from FIG. At this point, it can be seen that the discussion in “Discussion Space 1”, which was only questions and answers in the previous time, has moved to a more productive discussion center such as answers, suggestions, and summaries. In addition, the discussion in “Discussion Space 3” had been proposed in the previous time, but at this point, only the questions are present, and it can be seen that the discussion has not progressed very well.
Of course, in the present embodiment, as in the first embodiment, a desired partial discussion space can be accessed using the display of the discussion perspective map.
As described above, in the present embodiment, by introducing the coordinate axis, the user can grasp the current discussion situation more accurately and appropriately to the quality.
[0104]
Next, a third embodiment according to the present invention will be described.
The structure of the discussion space in this embodiment is the same as that in the first embodiment, but the configuration of the apparatus and the concept of the discussion perspective view are slightly different, and will be described.
[0105]
“Concept of Discussion View in Third Embodiment”
In the present embodiment, the conceptual configuration of the service and discussion perspective map is as shown in FIG. 31. This is the same as in the second embodiment described above, which is a partial discussion based on a search expression (query) and coordinate axes (axis). It provides users with access to the space. The difference from the second embodiment is that a discussion space related generation means 30 is added. The discussion space relation generation means 30 detects the trend of interest of all the users by integrating the situation of the discussion perspective map in each discussion view client 11, and further feeds it back to each user.
The introduction of the discussion space relation generation means 30 highlights the current state of discussion that many users are commonly interested in. In addition, discussion spaces that are of common interest among many users are shown to each user, exchange between discussion spaces is activated, and discussion is further activated.
[0106]
FIG. 32 is a block diagram of the apparatus of this embodiment.
In contrast to the first embodiment, a discussion space related generation unit 30 is added on the network 5, and a discussion space related link unit 16 is added to each discussion view client 10. The discussion space relation generating means 30 and each discussion space relation linking means 16 are combined to constitute a discussion relation means.
The discussion space relation generating means 30 communicates with each discussion space relation linking means 16 via the network 5, acquires a discussion perspective map of each user, synthesizes them, and returns the result to the discussion space related linking means 16. belongs to.
[0107]
"Discussion View Client Behavior"
FIG. 33 shows a PAD diagram expressing the operation of the discussion perspective view related UI means 13 in the present embodiment.
The difference from the second embodiment in this figure is that processing for the discussion space related information reception notification event has been added (step s9-026). This event occurs when the discussion space related linking means 16 receives the discussion space related information from the discussion space related generation means 30.
For this event, the process branches to step s9-027. In step s9-027, the discussion space related information r is acquired and stored in step s9-028. The stored discussion space related information r is used in the discussion perspective map redisplay procedure (steps s8-008, s8-013, s8-023). The data structure of the related information r is a set of discussion space IDs of related discussion spaces.
[0108]
Another change is the addition of steps s9-014 and s9-015 and steps s9-024 and s9-025. This is to pass new related information between the discussion spaces to the discussion space related linking means 16 when the discussion perspective map is updated / corrected. In step s9-014 or s9-024, relevant information between the updated / modified discussion view discussion spaces is detected. In the present embodiment, it is determined that a discussion space having a certain area or more in area where the figure that represents the discussion space on the discussion perspective map overlaps has a certain size or more. Of course, various other evaluation criteria such as Euclidean distance and other distances between the centroids of the figures representing the discussion space can be used. In step s9-015 or step s9-025, the discussion space related information r is passed to the discussion space related link means 16.
[0109]
"Operation of discussion space-related linkage means 16"
FIG. 34 is a PAD representing the operation of the discussion space related link means 16 in the third embodiment.
This operation enters an infinite loop in step s10-001 immediately after execution. In the loop, the process waits for an event in step s10-002. In step s10-003, if the received event is transmission of discussion space related information, the process branches to step s10-004. As described above, this event is generated by the discussion perspective map related UI means 13.
[0110]
In step s10-004, the discussion space related information r is taken out and transmitted to the discussion space related generation means 30 in step s10-005. As shown in FIG. 35, the data structure of the discussion space information transmitted / received to / from the discussion space relation generating means 30 is “discussion space related information ID” for identifying the information here, and the source of the information is identified. “Discussion space related information” including an ID set of related discussion spaces.
On the other hand, in step s10-006, if the received event is reception of discussion space related information, the process branches to step s10-007. This event occurs when new discussion space related information is received from the discussion space related generation means 30. In step s10-007, the discussion space related information r is taken out and passed to the discussion view map related UI means 13 in step s10-008.
[0111]
"Operation of the discussion space related generation means 30"
FIG. 36 is a PAD diagram representing the operation of the discussion space relation generation means 30 in the third embodiment.
This operation initializes the variable s to be empty at step s11-001 immediately after execution. As will be described later, this variable stores information related to the discussion space in the previous process in the loop.
In step s11-002, an infinite loop is entered. In the loop, an event is waited for in step s11-003. If event e is reception of discussion space related information in step s11-004, the process branches to step s11-005. To do. This event occurs when the discussion space related generation unit 30 receives the discussion space related information from the discussion space related link unit 16 of each discussion view client 10.
[0112]
In step s11-005, the received discussion space related information r is acquired, and in step s11-006, the discussion space related situation is recalculated. The discussion space related information is a set of related discussion spaces as shown in FIG.
The discussion space relation generating means 30 has a table for collecting the discussion space related information generated from each discussion view client 10, and this table in this embodiment is as shown in FIG. It holds what the current discussion space related information is like.
[0113]
In step s11-006, after reflecting the discussion space related information r in this table, the number of the same set appears for each related discussion space ID set from the entire table, and more than a predetermined number appears. The set is extracted as discussion space related information. That is, a set of discussion space sets determined to be “relevant” by many discussion view clients 10 is extracted. This result is assigned to the variable s1.
In step s11-007, it is checked whether or not the recalculated variable s1 is different from the variable s storing the previous related discussion space information. If so, the process branches to step s11-008. In step s11-008, this new related discussion space information s1 is transmitted to each connected discussion view client 10, and the previous related discussion space information s is updated to s1 in step s11-009.
[0114]
“Operation Example of Third Embodiment”
FIG. 38 shows an operation example of the third embodiment.
In the figure, three users, user A, user B, and user C, are using this apparatus. From the viewpoint of the interests of the users B and C, the “discussion space 1” and the “discussion space 2” are highly relevant. The discussion space relation generating means 30 totals this and feeds back the totaled result to the user A as well. From the viewpoint of the interest of the user A, the “discussion space 1” and the “discussion space 2” are not related to each other, but a new relationship with the “discussion space 2” is shown as shown in the figure. As a result, the user A can now know that the majority of other users' interests are in “something” that is common between the “discussion space 1” and the “discussion space 2”. Participation in “Discussion Space 2” was not promoted.
In this operation example, for the convenience of illustration, it is expressed so that “discussion space 1” and “discussion space 2” are determined to be related to each other by matching the interests of two users, user B and user C. However, it goes without saying that more users are necessary for this judgment.
[0115]
【The invention's effect】
As described above, the present invention is based on the structure-conservative extraction function of the relevant discussion part by introducing a partial discussion space. It is possible to provide a user with a current state grasping and transition tracking function for the entire space. As a result, the user can grasp the state of the entire discussion space scattered in a wide-area distributed environment and track the transition, and in particular, grasp the current state of the discussion in the place where the opinion was made and the circumstances so far. Since the search result structure and user interface are provided, the user can grasp the current state of the discussion and track the transition more easily and accurately.
[0116]
In addition, the present invention provides feedback from the overall trend of users, which is a “discussion space in which many users are interested”. It is also possible to provide a user with a current state grasping function. As a result, the discussion space in which individual users are interested is linked to the discussion space in which many users are interested. It will be possible from a global perspective.
[Brief description of the drawings]
FIG. 1 is a diagram illustrating a configuration of an electronic discussion support apparatus according to a first embodiment.
FIG. 2 is a conceptual diagram showing a logical structure of a discussion space according to the first embodiment.
FIG. 3 is a conceptual diagram of services provided by the first embodiment.
FIG. 4 is a diagram illustrating an example of a data structure of a discussion space.
FIG. 5 is a diagram illustrating an example of a data structure of a discussion space.
FIG. 6 is a diagram illustrating an example of a data structure of a discussion space.
FIG. 7 is a diagram illustrating an example of a data structure of a discussion space.
FIG. 8 is a conceptual diagram illustrating the meaning of a logical operation “*”.
FIG. 9 is a conceptual diagram illustrating the meaning of a logical operation “+”.
FIG. 10 is a conceptual diagram illustrating the meaning of a logical operation “#”.
FIG. 11 is a diagram illustrating a configuration example of a discussion space search request;
FIG. 12 is a diagram illustrating a configuration example of a discussion space search request.
FIG. 13 is a PAD diagram representing the operation of the discussion space search control means.
FIG. 14 is a PAD representing processing contents of a discussion space search control thread.
FIG. 15 is a diagram showing an example of the data structure of search status storage means.
FIG. 16 is a PAD representing the operation of the discussion space search means.
FIG. 17 is a PAD representing the operation of the search cooperation means.
FIG. 18 is a diagram showing a user interface screen according to the first embodiment.
FIG. 19 is a diagram illustrating a user interface screen according to the first embodiment.
FIG. 20 is a PAD diagram representing the operation of the discussion perspective related UI means.
FIG. 21 is a PAD diagram representing the operation of the discussion perspective map constructing means.
FIG. 22 is a PAD diagram representing the operation of search request generation means.
FIG. 23 is a diagram illustrating an operation example of the first embodiment.
FIG. 24 is a diagram showing an operation example of the first embodiment.
FIG. 25 is a diagram illustrating an operation example of the first embodiment.
FIG. 26 is a conceptual diagram of services provided by the second embodiment.
FIG. 27 is a diagram illustrating a user interface screen according to the second embodiment.
FIG. 28 is a PAD diagram representing the operation of the discussion perspective view related UI means in the second embodiment.
FIG. 29 is a diagram illustrating an operation example of the second embodiment.
FIG. 30 is a diagram illustrating an operation example of the second embodiment.
FIG. 31 is a conceptual diagram of services provided by the third embodiment.
FIG. 32 is a diagram showing a configuration of an electronic discussion support apparatus according to a third embodiment.
FIG. 33 is a PAD diagram representing the operation of the discussion perspective related UI means.
FIG. 34 is a PAD diagram representing the operation of the discussion space related cooperation means.
FIG. 35 is a diagram illustrating a data structure of discussion space-related information.
FIG. 36 is a PAD diagram representing the operation of the discussion space relation generation means.
FIG. 37 is a diagram illustrating a data structure of a discussion space-related situation.
FIG. 38 is a diagram illustrating an operation example of the third embodiment.
[Explanation of symbols]
1 ... WWW server, 2, 4 ... Electronic document storage means,
3 ... DBMS, 5 ... Network,
6 ... DBMS Gateway, 10 ... Discussion View Client,
11 ... WWW browser, 12 ... Discussion execution means,
13 ... Discussion view map related UI means, 14 ... View map composition means,
15 ... Discussion space search request generating means, 16 ... Discussion space related linking means,
20 ... Discussion view server, 21 ... Discussion space search control means,
22 ... Search status storage means, 23 ... Discussion space search means,
24 ... Search linkage means, 30 ... Discussion space related generation means,

Claims (12)

Connected to a network having electronic document storage means for storing opinions in discussion as electronic documents associated with at least attribute information of opinions and information related to other opinions, and a plurality of users can exchange and view electronic documents. An electronic discussion support device for discussion,
The discussion management data including the Sanshoko indicating the entity or position of the entity of the electronic document according to the attribute information and the discussion space of at least discuss space is managed for each discussion space,
An electronic document extraction function unit that performs a search process according to a search formula for the electronic document and the discussion management data;
The electronic document extraction function unit searches a plurality of electronic document entities that match the search formula or a reference indicating the position of the entity, and extracts and extracts these search results for each discussion space , The search result management data including the extracted electronic document entity or a reference indicating the position of the entity is generated for each discussion space as the extraction source, and the plurality of search result management data is used based on the search formula. Electronic discussion support device characterized by search results. In the electronic discussion support device according to claim 1,
User interface means for accepting input of the search expression by a user and passing the search expression to the electronic document extraction function unit, and receiving the search result extracted by the electronic document extraction function unit and displaying the search result An electronic discussion support device characterized by comprising: In the electronic discussion support device according to claim 2,
The user interface means is provided in a client device operated by a user,
An electronic discussion support device, wherein the electronic document extraction function unit is provided in a server device connected to the client device via a network. In the electronic discussion support device according to claim 2 or claim 3,
A plurality of the electronic document extraction function units, each electronic document extraction function unit is provided with search linkage means for exchanging the search formula with other electronic document extraction function units,
The electronic discussion support apparatus, wherein the user interface means displays the search results received from a plurality of electronic document extraction function units. In the electronic discussion support device according to claim 2 or claim 3,
Search expression storage means for storing the search expression,
The electronic document extraction function unit repeatedly extracts the search result based on the search formula stored in the search formula storage means,
The electronic discussion support apparatus according to claim 1, wherein the user interface means repeats updating of the display every time a search result is extracted by the electronic document extraction function unit. In the electronic discussion support device according to claim 2 or claim 3,
The user interface means receives an input of an evaluation function for evaluating a search result by a user, and evaluates and displays the search result received from the electronic document extraction function unit by the input evaluation function. An electronic discussion support device. In the electronic discussion support device according to claim 2 or claim 3,
Extracting a set of discussion space having a high similarity between the discussion space included in the search results for a plurality of users and a set of discussion space with their high degree of similarity is common from the search results for a number of users When it is extracted, it includes a discussion space generation unit for notifying the user interface means of the determination result as a determination result with the set of the discussion spaces having the high similarity,
The electronic discussion support apparatus according to claim 1, wherein the user interface means displays a relevance based on the determination result notified by the discussion relation generation means. The electronic discussion support device according to claim 6,
The user interface means displays a character string or a figure representing each of the plurality of search result management data constituting the search result in the order of the value of the evaluation function, and supports electronic discussion apparatus. The electronic discussion support device according to claim 6,
The user interface means positions a character string or a graphic representing each of the plurality of search result management data constituting the search result in a graph of an arbitrary dimension with the value of the evaluation function as an axis. An electronic discussion support device characterized by displaying search results. In the electronic discussion support device according to claim 8 or 9,
The user interface means displays a character string or a graphic representing the search result management data, accepts an operation instruction for the display by the user, and holds the search result management data corresponding to the instructed display. An electronic discussion support apparatus that displays a character string or a graphic representing the entity of the electronic document or the electronic document to be referred to. In the electronic discussion support device according to claim 8 or 9,
The user interface means includes
The character string or graphic representing the search result management data is displayed, and an operation instruction to the display by the user is received, and the electronic document entity held by the search result management data corresponding to the indicated display or A character string or a figure representing the electronic document to be referred to is displayed together with related information between the electronic documents,
In the display, the character string or figure corresponding to the electronic document entity or the reference electronic document held in the search result management data is replaced with the character string or graphic not corresponding to the electronic document entity or the reference electronic document. An electronic discussion support apparatus characterized by displaying in a form different from a figure. In the electronic discussion support device according to claim 10 or 11,
The electronic discussion support apparatus, wherein the user interface means displays the content of the electronic document corresponding to the character string representing the entity of the electronic document or the electronic document to be referred to or the graphic.

Images