根據研究顯示，理想索引典應是模擬人類思路及記憶結構的線上圖形索引典，而 資訊科技的進步正提供其絕佳的發展機會。這個理想的索引典分為主題索引典和作者索引典 ，其中主題索引典是一語意網路構造之圖形索引典。在此主題索引典中，階層關係除保持原 有的廣義詞和狹義詞外，可更進一步細分為類屬關係、主題包含關係及整體部份關係，其中 只有類屬關係能將屬性直接由電腦自動傳給子類。同時也將聯想關係分成八種不同的關係， 因此空間認知可以發揮它的功能。也就是說，此索引典將提供處理不確定所需的複雜和多樣 性詞彙，它具有學習能力，能模仿人類思考，是一個擬人化的讀者自建之超索引典。至於作 者索引典，除了保留現有字母順序關係的陳列外，還可透過共用書目分析提供和該指定作者 研究領域類似的作者供讀者參考。

　　　　　This article presents ideal thesauri for both subject and author parts. The subject thesaurus is a graphic thesaurus with a semantic network, so the spread activation assumption can be applied. The hierarchical relationship is broken into three specific relations, so automated inheritance can be performed. The associated relationship is separated into several specific relationships, so spatial recognition can function well. This thesaurus will provide variety and complexity to deal with uncertainty, it will be a user-constructed, user-enhanced hyper-thesaurus that serves as an analogue of human mind. Concerning the author thesaurus, it is designed to provide all other authors which are alphabetically adjacent to a given author as well as other authors whose subject area are close to the given author.