tea.mathoverflow.net - Discussion Feed (Classic questions)

Andrew Stacey comments on "Classic questions" (5617)

Andrew Stacey — Wed, 19 May 2010 00:19:29 -0700

To be precise, it is the mathematical part of the preview that is slow. I can always tell when the "preview math" button has gotten reset for me (it's usually "below the fold" so I don't notice beforehand) because as soon as I type a dollar sign then everything slows right down.

But of course, speed isn't the only reason for using the system I use. The fact that I'm used to Emacs and it's shortcuts and abilities means that composing length posts in Emacs is much easier for me than writing them in a text box in a browser.

Anton Geraschenko comments on "Classic questions" (5610)

Anton Geraschenko — Tue, 18 May 2010 19:09:35 -0700

The preview on MO is done locally. I can't think of any reason it would be slow. There's a timeout set so that if it takes more than three or four seconds to generate the preview, it won't bother, so the preview might slow down when you're writing a pretty long post. But even then it shouldn't interfere with typing into the input box and you can update the preview by just stopping typing for a couple of seconds.

Harald Hanche-Olsen comments on "Classic questions" (5608)

Harald Hanche-Olsen — Tue, 18 May 2010 11:59:07 -0700

The preview is remote? I though it was all locally processed. I certainly haven't seen much of a speed issue.

Andrew Stacey comments on "Classic questions" (5607)

Andrew Stacey — Tue, 18 May 2010 11:53:57 -0700

The problem with that solution is that the remote preview is rrrrrrrrrrrreeeeaaaaallllllyyyyyyyyy ssssssslllllllllooooooooowwwwww. Downloading and installing the scripts locally makes the whole thing much faster. (This is what I do for nLab entries: I use the "It's all text" firefox extension to easily transfer a "text box" to an emacs session. Then I load in the latex bindings for emacs, except that I tell it to call "maruku" (the ruby version of markdown, which has itextomml built in) instead of latex and "firefox" instead of xdvi. Then I just do the normal "compile-and-view" cycle that I do for LaTeX documents until I'm happy with it, whereupon the "It's all text" extension reloads it back into the browser and I sent it off to the nLab for processing.)

Harald Hanche-Olsen comments on "Classic questions" (5605)

Harald Hanche-Olsen — Tue, 18 May 2010 10:16:48 -0700

You can always write an answer right here on a MO page and see the preview as you write. Remember, you don't have to submit it! Just copy the text and paste into a local text file on your own computer to save as a draft for later polishing. In fact, I would always do this anyhow for submissions of nontrivial size, just to have a backup in case the web browser crashes or the MO backend misbehaves.

Andrew Stacey comments on "Classic questions" (5604)

Andrew Stacey — Tue, 18 May 2010 07:47:00 -0700

Probably not. You never know, someone might already have done it for you

Mariano comments on "Classic questions" (5603)

Mariano — Tue, 18 May 2010 06:49:52 -0700

A translator from a minimal variant of LaTeX to markdown+jsmath should not be thathard to write, too.

Andrew Stacey comments on "Classic questions" (5602)

Andrew Stacey — Tue, 18 May 2010 05:58:39 -0700

You could always install markdown and jsmath on your own computer and do a "write + compile" cycle using that. The difficulty with this is that there are many variants of markdown so you'd need to be sure to get the right one. Actually, you could do it with the javascript markdown that Anton wrote since that was designed to be the same as on MO.

So, here's how it could be done:

Write your question/answer in a text editor
Save it to a file and add some HTML wrapper tags around it
Add a javascript header to load the wmd.js and jsmath.js (which you should download from somewhere)
Look at the resulting file in your browser

Obviously, that's just an outline. In particular, one would need to tweak the javascript a bit and select the HTML wrapper stuff appropriately. But that would only need to be done once.

gilkalai comments on "Classic questions" (5601)

gilkalai — Tue, 18 May 2010 05:41:11 -0700

Thanks, Harry
One other features that can be useful is to have a way to write drafts of questions and answers so it will not be necessary to edit them in real time. (I think Wikipedia has such a feature). Is this possible?

Harry Gindi comments on "Classic questions" (5600)

Harry Gindi — Tue, 18 May 2010 00:10:44 -0700

@gilkalai:

Nope, not at the moment. This has been requested by something like half of the people on meta, but as of now, there is no way for the staff to actually implement that change (even if they wanted to. If I remember correctly, Anton and Scott were both opposed to this change, but I'm not positive, and I'm sure that they will tell you themselves. Either way, I'm sure one of them can supply you a link to an old discussion of this precise issue.)

gilkalai comments on "Classic questions" (5599)

gilkalai — Mon, 17 May 2010 22:57:56 -0700

I agree that there is no reason to have the "big list" problems pop up in the front page more than now. In fact, I have a related question. Right now any time a question or answer is edited it automatically appears in the front page. Is there a way to restrict it only to major editing, either by letting it depends on the number of characters changed or by asking the editor if the change justifies presenting the question in the front page?

Harry Gindi comments on "Classic questions" (5598)

Harry Gindi — Mon, 17 May 2010 22:57:39 -0700

Scott and Qiaochu have hit the nail on the head. In fact, if you look at many of the older big-list questions, you can see that they've been closed for the reasons cited above. MO is mainly for the substantive questions. Big-list/soft-questions are merely a nothing more than a sideshow for the main event.

Mariano comments on "Classic questions" (5597)

Mariano — Mon, 17 May 2010 21:07:33 -0700

@Qiaochu: I haven't really checked, but: is that a significant problem?

Qiaochu Yuan comments on "Classic questions" (5596)

Qiaochu Yuan — Mon, 17 May 2010 17:28:53 -0700

In particular, most people often don't bother to read all of the past answers and end up duplicating entries, which gets really annoying.

Scott Morrison comments on "Classic questions" (5595)

Scott Morrison — Mon, 17 May 2010 16:40:38 -0700

Oh dear. I'd far prefer to go deleting these questions than making them more prominent! :-)

Qiaochu Yuan comments on "Classic questions" (5591)

Qiaochu Yuan — Mon, 17 May 2010 14:34:40 -0700

When we've talked about this situation, most people have had the opposite reaction: these questions often accumulate too many low-quality answers from new participants and get bumped when they shouldn't be. That's why we've been closing them.

It's also already really easy to find these questions: you can either click on big-list or you can rank by votes.

vipul comments on "Classic questions" (5590)

vipul — Mon, 17 May 2010 14:16:44 -0700

Now that MO's been around for about seven months or so, a lot of questions of general interest (typically, labeled "community wiki", "big-list", "soft-question", etc) have accumulated. These typically garner about 15-50 votes, get a bunch of 20-25 answers, and then die out. (For instance: common errors in mathematics, funny things in mathematics papers, trends in what mathematicians should know, selecting problems for students ...)

It would be good to have these questions accessible and perhaps even randomly shown on the front page occasionally, so that the questions don't get buried and new participants can quickly see these questions and contribute answers. This way, the answer lists to these questions can keep improving. It also might encourage participation because these questions can often easily be answered based on personal experience and don't require on-the-spot mathematical thinking.

Along these lines, I thought of something like a "classic" tab that displays questions that satisfy some or all of these criteria: (i) are community wiki and/or big list (ii) got at least x votes (where x could be 15 or 20 or 30) (iii) got at least y answers (where y could be 5 or 10 or 20) (iv) got at least z votes (where z could be 1K or 2K), and (v) is at least two weeks old (because recent questions are displayed in the front anyway).

This is just a very rough idea and I don't know about its suitability or appeal to MO users.