« first day (1312 days earlier)      last day (388 days later) » 

2:20 AM
Jun 3 at 5:28, by Martin Sleziak
What to moderators think about this proposal: Should we add external searches in some help pages about search? (Posted 6 months ago, score +14.)
Jun 17 at 9:57, by Daniel Fischer
@MartinSleziak As long as it's still intelligible, not all is lost. I've asked the others for their opinion on it. Maybe some will reply.
Jun 29 at 6:25, by Martin Sleziak
@DanielFischer I will just add that I have mentioned this also in search/Approach0 room. The author of the website warned that it is still under development - you can see the reply in full there.
@DanielFischer Any word from other mods on this suggestion: Should we add external searches in some help pages about search?
 
 
14 hours later…
4:27 PM
@MartinSleziak I'd prefer to link to a meta post where this information is collected. On the one hand, that'd be more flexible. On the other hand, it side-steps the question what exactly is recommended. (I am not keen on discussing which search engines exactly should be listed there. If it's a meta post everybody that think that one is missing can just add it.)
 
So we should create some kind of FAQ-search post on meta? (Or, possibly, find a suitable one among already existing questions.)
 
@MartinSleziak yes I think this would be a good idea.
 
Why, after quite some time now, have the SE developers not addressed the rather poor search available directly on a given site? Surely, it could be improved? and if they can't improve it, then the SE developers ought to be talking outside search engines and what they'd have to offer for SE.
 
5:03 PM
@quid ^^^ Or my question may be somewhat naive. Improving the SE "search" feature would go a long way to more readily identifying, for askers, nearly identical questions as theirs, before needing to ask (hence cut down on duplicate questions), and for other users attempting to search for a duplicate of a question already posted, and improved search of any given site's preexisting questions would be able to return results, and hopefully, more readily be closed as dupes before answerers come.
 
@quid I imagine that the question could be rather short - I made a draft here, I'll be grateful for suggestions/improvements.
I suppose that both the question and answer would be CW - it is supposed to be maintained by community. And I think it might be useful to have separate answer for built-in search (describing a few most common search modifiers and explaining where to find more), separate answer for Approach0, et.c
 
@amWhy I am vaguely under the impression that some things were done to improve it. Getting good and easy to use search is I think not at all easy. I suspect the answer is as for most such things that it is a trade-off between time and cost involved and what it brings in return.
For example, the search functionality for MathSciNet seemed pretty rudimentary to me, too (I did not check recently if it improved). One could also ask why the AMS does not do more about it while they charge quite a lot for access to MathSciNet and one cannot even parse it with external tools.
Then "math aware" search is somewhat niche and even more tricky. I think "Zentralblatt" and maybe Springer more broadly tried to do something in that regard. I did not really follow it in detail though.
Possibly if somebody would reach out to SE specifically they might help some project. I think they did/do provide funding to MathJax.
 
5:19 PM
@quid I'm sorry if I came across as "intolerant" of the current site search. That's not what I meant. But much of this site is by and for developers, programs, etc., so I'd expect SE to be able to handle a site-sensitive search for MSE, and to have a step-up on AMS...
I realize it wouldn't be easy, not if copying and pasting text from an unformatted post which itself was copied and pasted from a pdf reader. But at the very least, if I encounter "T(n)=3T(n/2)+Θ(n)" in a question about it, I should be able to format it in mathjax, place in the search window, and get back any earlier posts containing the same, already formatted. I can imagine mathjax may present a problem, (though the folks at Google aren't encumbered in allowing for it and recognizing hits)
And I am completely fine with SE not wanting to spend resources to improve search. But in that case, they need to be working on an alternative, like hosting a search engine that can do so.
Anyway, @quid. This is not a top-of-the-list concern of mine. So I did not try to make it sound like an impending emergency. I just added my thoughts, given the conversation here. Perhaps they are not/were not welcome.
On a different note: Yikes, I just imagined the day when the userenters a chatroom they've never entered before, or hadn't already entered it a week or so ago., say @qued, with an auto comment posted on their behalf: Please welcome @qued. @qued is a new chatroom contributor!)
 
5:38 PM
@amWhy I am not sure (in a literal sense that I don't know) that the balance is like this. On the example you give, I thought this works more-or-less. Though it does not handle parenthesis smoothly it seems to me.
@amWhy it is not that they are not welcome, they are welcome as far as I am concerned. However, the form in which they are expressed to me reads as rather critical and implying some form of negligence (if not incompetence). I take it this was not intended, but that's how I would read the initial comment(s).
 
@quid I apologize if that's how it read. I do get impatient because the quality of the search feature has long, long, long been a problem stated and discussed on this site, anyway.
Finally, don't miss my fifth comment in that chunk; Meant to be humorous, and that's all.
 
@amWhy well it certainly did come up a few times on the main meta too :-)
 
6:10 PM
@quid Wow, pretty recently too!!
 
 
5 hours later…
11:37 PM
@MartinSleziak I like the suggestion, and agree with quid about it being better as a meta post. The fact that we are talking about external sources makes it somewhat important to be easily modifiable at will (external sources may have issues, become unusable, new external sources may come up etc).
 

« first day (1312 days earlier)      last day (388 days later) »