I have noticed this interesting behavior several times now.

Many a time I scour Math SE looking for a particular query but fail to find anything even after several rounds of rephrasing and reformatting my query. I finally accept defeat and decide to ask a question myself, so I type in my question title and lo and behold! there are one or two relevant results under "Questions that may already have your answer". But then pasting and searching for same title text(equation markup included) yields no results.

So if Math SE has the ability to do a better markup-aware search feature which can return more relevant results, why is it not implemented for the normal search function? Is it an issue of computational cost?

If it matters, this is what I found while asking this but did not show up in search.

  • 22
  • 16
    Case in point! xD – Akay May 24 '17 at 18:03
  • 21
    Curiously none of the meta variants that Martin links to have a meaningful (IMHO) answer from the developers, yet they were (IMHO trigger-happily) closed as dupes of each other. – Jyrki Lahtonen May 24 '17 at 20:20
  • 11
    @JyrkiLahtonen Yes it is very odd that none of them have a proper answer. I for one consider this issue to be quite an annoyance, and only learnt quite recently that I should use the 'ask a question' in order to conduct more meaningful searches. – mrnovice May 24 '17 at 20:39
  • 13
    It's time to fix this. It's becoming really annoying. – Zaid Alyafeai May 27 '17 at 3:48
  • @ZaidAlyafeai Not really – four_lines May 27 '17 at 22:45
  • I think it is great. And like any search your mileage will vary depending on your wording. – SDsolar May 28 '17 at 13:59
  • 4
    @TheGreatDuck If you are privy to inner workings of the two search engines please elaborate in an answer! The shortcomings of the regular search engine are IMHO a major obstacle to smoother working of the site. – Jyrki Lahtonen May 29 '17 at 8:42
  • @JyrkiLahtonen I don't know the inner workings of the search engine. I was just making a remark. It's obvious that the search uses your whole post and not just the title. I don't claim to know how it does it. It just seems obvious to me that it would be using the whole page to run a search of some kind. – The Great Duck May 29 '17 at 14:45
  • @JyrkiLahtonen to be honest, I was pretty much just speaking out of my butt. It seems like everyone should be able to deduce that the body (and not just the title) factors into the search. I have honestly no clue though. For all I know, it only works by using the title. – The Great Duck May 29 '17 at 23:14
  • 8
    @TheGreatDuck Actually, the results returned after just typing in your question title with markup(no need to type in the body sometimes - YMMV) are better than than the normal search. – Akay May 30 '17 at 5:30
  • @Akay does normal search allow markup? That could be the issue. – The Great Duck May 30 '17 at 18:30
  • I'm probably going to go through the actual boxes in my browser later and see if I can isolate the code differences within the browser. It might just be a different interface they use or something. – The Great Duck May 30 '17 at 18:31
  • 3
    Many times (perhaps an average of more than once a day), I'll see a question that I know is similar to something else I've seen, and to find it (to post as a comment) I'll simply use one or more reasonable sounding key words along with the word "Stackexchange" in a google search. I've tried searching the site itself, but this seems way worse, so I just google . . . – Dave L. Renfro May 31 '17 at 18:37
  • 3
    @DaveL.Renfro Since you have mentioned searching the site, I'll add that if I see a question which was likely repeated many times, then I often find it by choosing a tag (or several tags) and then displaying frequent tab (which contains questions with most links). And for formulas Approach0 is often very handy. I will also add link to this question and this collection of meta posts. – Martin Sleziak Jun 1 '17 at 3:38

You must log in to answer this question.

Browse other questions tagged .