hypothesis / hypothesis/product-backlog

multiple freetext terms should constrain search results

Open
#81 0 comments 0 reactions 0 assignees View on GitHub
activity pages Elasticsearch
Dominant language
No language data
Stars
122
Forks
7
PR merge metrics
No merged PRs in 30d

Description

### Problem you are trying to address with this feature

Facet-based lozenges combine in a way that progressively constrains the result set. Free-text based lozenges, however, do the reverse. It's inconsistent and -- in most cases, I claim -- unexpected and not optimally useful.

There's a related issue relevant here. A search for "gold star" -- in quotes -- [finds an annotation](https://docs.google.com/document/d/14c_vlwwtdFOAaTUIrT9B7D5Vpa-6UJ-U3dGXsVUp2vM/edit#heading=h.37lnj8445501) that doesn't contain the exact term. Why? If the answer is that the search engine is fuzzy, then we can discuss why and for whom. Ideally we'd find out, from users, how many expect fuzziness and how many don't, pick a default setting accordingly, and make the alternative available.

Whatever the conclusion, if users want fuzzy they have it now, and can have it for any of the terms in a multiple-freetext-term query, by doing each individually. The benefit of combining that effect is, I claim, far less than the benefit of meeting the expectation we set that added facets always constrain. Again this is a thing we can test with users when there's enough experience with activity pages to do so, but meanwhile I believe always-constrain is the right default.

### Your solution

Make free-text lozenges AND together as facet-based ones do.

Contributor guide

No contributing guide indexed for this repository

Research direction

Start by reviewing how free-text and facet-based lozenges currently combine on activity pages, then read the linked annotation about the quoted “gold star” search. Define done as multiple free-text lozenges combining with AND semantics like facet-based lozenges, while clarifying the expected fuzzy-search behavior before implementation.

Written by the indexing model from the issue text.

Assessment

Domain
search
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.