# Tapestry **Entity class:** Collaborative filtering system / Xerox PARC ## Definition and history Xerox PARC's Tapestry treated human annotations and stored messages as queryable inputs to a reader's filter in 1992. [[articles/Intelligence Inside|Intelligence Inside]] identifies its unusual design property: a reader could write and inspect the rule selecting which existing material would arrive. An item might remain in the repository until a trusted person's reaction satisfied the reader's query. The annotation became a first-class part of the selection mechanism rather than a hidden aggregate score. ## Mechanism and limits Tapestry's filter could be rerun over stored material when later annotations changed what qualified. This separated the time of publication from the time of delivery, making social response an explicit input to information flow. The mechanism is a predecessor to collaborative recommendation while preserving a degree of user control and inspectability that later selection systems need not provide. ## Relationships **Research setting:** [[wiki/Xerox PARC|Xerox PARC]]. **Method:** [[wiki/Collaborative Filtering|Collaborative Filtering]]. **Successor comparisons:** [[wiki/GroupLens|GroupLens]] and [[wiki/Ringo Recommender|Ringo Recommender]]. **Principle:** [[wiki/Reader-Inspectable Selection|Reader-Inspectable Selection]]. **Article route:** [[articles/Intelligence Inside|Intelligence Inside]]. ## Sources [Xerox PARC, *Using Collaborative Filtering to Weave an Information Tapestry* (1992)](https://www.bitsavers.org/pdf/xerox/parc/techReports/CSL-92-10_Using_Collaborative_Filtering_to_Weave_an_Information_Tapestry.pdf); [[articles/Intelligence Inside|Intelligence Inside]].