15

Seriously, I was attempting to search for threads in the Photography Stack Exchange site and did a search for “HEIC” with sorting ordered by “Votes” and it seems the search logic — whatever that might be — is interpreting “HEIC” as “Hey” which is… Not really great. Screenshot below.

Are there exception lists for known acronyms and file extensions to improve sorting? Can HEIC be added to that if possible.

Screenshot showing “HEIC” being interpreted as “Hey.”

And as noted in a comment by animuson, adding quotes to “HEIC” in the search ends up giving the expected results; screenshot below.

But honestly adding quotes to a single word doesn’t seem like an intuitive solution for most people who will be casually searching for something like “HEIC.”

Screenshot showing adding quotes to “HEIC” provides desired results.

Improve this question
3
  • That looks like stemming is responsible, though I wouldn't have expected this term to be affected by stemming at all.
    – Mad Scientist
    Jul 6, 2020 at 16:23
  • 9
    I've never heard of an exception list for that system, and I don't believe one exists. But there has got to be a way to identify situations where Elasticsearch is changing what it is looking for and display a note at the top like Google does. "Showing results for {altered}; search instead for {original}" would be great there. Because searching for "HEIC" (with quotes) works perfectly.
    – animuson StaffMod
    Jul 6, 2020 at 16:23
  • 5
    This is a good post, but to be honest I just upvoted it for the title.
    – Cave Johnson
    Jul 6, 2020 at 17:02

1 Answer 1

Reset to default
11

Stack Exchange's search uses Elasticsearch with default 'stemmers' for English, which try to combine variations of the same word. It's possible to tune those stemmers but it's a lot of work and I guess it won't be done for individual cases like this one.

If I would encounter such a case, I'd immediately spot it (you can see the highlighted heys in the search results) and either

Improve this answer
1
  • 7
    Decent answer. But will say that given Stack Exchange sites are tech leaning in nature, and the world of tech uses acronyms all the time, I still wonder if some generalized way of prioritizing acronym meanings ahead of default Elasticsearch stemming logic has value. What’s great about these sites is there is a lot of casual language in play, but I believe the culture of acronyms should be accommodated as well.
    – Giacomo1968
    Jul 6, 2020 at 16:31

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .