This "question" concerns the relationship between SO and tex.SE, in particular what the existence of the latter means for questions about TeX on SO.

First, my assumptions (yes, I'm a mathematician ...):

  1. SO is for programming questions, at a level appropriate to someone employed as a programmer (note that I do not say that one has to be employed as a programmer to ask a question on SO).

  2. tex.SX is for TeX-related questions, at pretty much any level since (to be honest) the barrier for asking a reasonable question is actually lower than the barrier for knowing what TeX is.

  3. The existence of tex.SX is a Good Thing and Officially Sanctioned by The Management. I've read a few comments on SO that go against this, but my assumption is that the people at the top are in favour of the SO daughter sites. I would add that I (as a non-programmer) find asking a TeX question on tex.SX a much, much more pleasant experience than on SO (but this isn't the place to explain that further).

  4. There are TeX-related questions on SO and the vast majority of them are not explicitly about programming.

    I include the word "explicitly" there because there is a level at which any question about TeX is about programming, since TeX is a programming language. But if SO spreads its net that wide then the majority of the new SE sites would be under its aegis and not have a reason to exist.

Prior to the existence of tex.SX, TeX-related questions seem to have been allowed on SO, but under sufferance. Which leads me to the first part of my question:

Does a generic TeX-related question have a place on SO?

Let me explain that "generic" a bit. My opinion is that for a TeX-related question to have a place on SO then it has to be a programming question that happens to involve TeX. Note the order there. A TeX question that happens to involve programming does not (in my opinion) belong on SO. So my question above is really asking if I have that right or not. If not, the rest of this probably won't make any sense, but I'll carry on regardless.

Assuming that I'm at least in the right ballpark on this, there is the further question as to what to do with TeX-related questions on SO that ought to really be on tex.SX. My understanding is that questions can be migrated, if enough people vote to do so and if a "pathway" has been opened between the sites (sounds like something out of Dr Who ...). This is all well and good, and my question could be just "Can we have such a pathway?" except for my next assumption:

  1. The best people to decide whether or not a TeX-related question belongs on tex.SX or on SO are the tex.SX people.

Quite apart from the knowledge issue (how many SO people with enough rep to vote to close actually a) know what TeX is, and b) monitor the TeX-related tags?), there's also the matter of incentives. There is no incentive for SO people to migrate TeX-related questions. They might be a little irritating to have on SO, but there's not enough to really be worth bothering about (unlike, say, homework questions). On the other hand, we over at tex.SX have all the incentives to go and grab questions from SO! So, the main part of my question:

How can I, as a high-ranking tex.SX user but a nonentity on SO, best encourage questions that ought to be migrated to be migrated?

Note that I say "encourage". I don't want to have a system whereby I just say "Migrate this one" and it happens. There should be checks and balances. But I would like something a bit stronger than just leaving a comment along the lines of "Come and ask this on tex.SX" since, again, there's no real incentive for the questioner to migrate it themselves. What would be nice would be that I could flag a question for moderator attention saying, "I recommend that this be migrated to tex.SX" whereupon the moderator could see my tex.SX reputation and use that as evidence for approving the migration.

(Severely edited to try to make this more focussed to make it clearer (many thanks to Grace Note for helping me clarify my own thinking).)

share|improve this question
2  
I agree with the migration path being necessary. But I believe in a philosophy that migration should happen not because a question can be addressed better elsewhere, but because it does not belong on the current site. I know too little about LaTeX to say where that line lies, so I'm requesting that you take simple care when selecting the questions that need to be migrated, and only select the ones which are off-topic to Stack Overflow. This might be all of them, perhaps, and I know you will know better than I will as to what belongs where. –  Grace Note Nov 18 '10 at 17:37
    
@Grace Note: I've deleted all my comments now as well. Thanks for the discussion - it helped me clarify what I meant to say (as opposed to what I actually said). I've rewritten the question to try to make it much more focussed. Please let me know if anything still isn't clear. –  Andrew Stacey Nov 18 '10 at 18:48
    
I'd upvote you if I hadn't done so already. I'm leaving that initial comment just to kinda... "supplement" your final paragraph, not for you but for those who read it. You're welcome, and thanks for enduring the bumbling of this Meta user here. –  Grace Note Nov 18 '10 at 18:55
2  
I wonder if there's a more general question here, about having two kinds of migration systems: in addition to the currently extant system, which is for sending irrelevant questions to elsewhere, this post seems to ask for the creation of a new system, for gathering relevant questions from elsewhere. Basically, TeX.SX isn't a special case—this sort of thing will only become more relevant as the numer of Stack Exchanges increases. –  Antal S-Z Nov 20 '10 at 4:49
add comment

2 Answers

I very strongly agree with the intent behind this proposal, and the actual proposal seems reasonable as well. As a StackOverflow user with sufficient privileges to vote for closure, and with some fair knowledge of TeX, I don't see many questions in the "latex" tag that seem signficiantly programming-related. Out of the top 75 on the list, I see 2 that definitely have programming nature, and 1 that I think is a "maybe".

There are currently over 1,700 questions tagged "latex" on StackOverflow, and extrapolating from that sample of 75, probably at least 1,500 of them should be migrated. Compare that to the not-quite-3000 questions on the TeX.SX site, and it's apparent that this is a significant fraction of the LaTeX questions in the StackExchange universe. I think that's a problem that merits solving.

I would disagree with one of your assumptions, though -- I don't think the TeX.SX people are necessarily the people who have the best idea whether a question should be migrated. It seems to me that the question really reduces to a simple matter of "Is this TeX-related?" (which is in general obvious) and "Does this have sufficient programming content to merit being on StackOverflow?" (which is a value judgement, and one that StackOverflow people can most accurately answer). This is because the question of migration needs to respect the OP's intent, and so it is not one of "where does this best fit", so much as "is this off-topic on the site where it was asked, and if so, does it belong somewhere else?" Off-topic is a local definition, and I think is the much harder question.

Edit: As Hendrik Vogt points out in comments, migrating the existing questions has some severe problems, which are probably not reasonably surmountable, so let's take that off the table. Analyzing recent questions, I see that the balance is not quite so split -- 64 "LaTeX"-tagged questions on StackOverflow from Jan. 1 to today (Jan. 29), and 599 in the same time on TeX.SX. Nonetheless, most of those 64 should be migrated, I think, and that's still a fairly sizable number of questions.

share|improve this answer
1  
I'm just visiting from tex.sx. One urgent plea: Please do not consider migrating hundreds of old questions to tex.sx. All this should essentially be about new questions. –  Hendrik Vogt Jan 28 '11 at 20:55
3  
Hendrik: Ok, but can you say why you think that should not be considered? I would have thought it would be a good thing to have the whole set of questions in one place, but I'm not strongly attached to that opinion. –  Brooks Moses Jan 29 '11 at 5:26
3  
@Brooks: Please don't forget to use "@" before my name. (I just found your question because I checked back.) Yes, it wouldn't be bad to have everything in one place, but there are a few severe problems. 1. Migrated questions appear at the top of the active questions. Now what would happen to the really recent questions if you put 1000 old ones on top? 2. There are many questions that have been asked on SO and on tex.sx, so we would get tons of duplicates. 3. The reputation on answers/questions looks totally different on the two sites, so the top voted questions on tex.sx would be buried. –  Hendrik Vogt Jan 29 '11 at 21:06
    
@Hendrik: All good points; I agree. –  Brooks Moses Jan 30 '11 at 0:28
    
@Hendrik I believe that #1 and #3 are a problem with migration code, not with the concept that they should be migrated. (For #3, are you saying that upvotes on one site get carried to the other site during migration? I wasn't aware of this.) –  Phrogz Apr 8 '11 at 17:45
    
@Phrogz: #1 can't really be dealt with: If you migrate tons of old questions and they do not bump to the top, then they don't get seen, and they'll just be sort of clutter in the archive. For example, if a question just carries the "latex" tag, then it will come as "untagged" to tex.sx (since a "latex" tag there is as useful as a "programming" tag on SO). Concerning #3: Yes, if you sort questions over at tex.sx by "votes", then the third from the top is a migrated question. Asked on tex.sx, it would just be a very average question. –  Hendrik Vogt Apr 8 '11 at 19:08
    
@Hendrik Questions with valid tags are desirable, but not required for a search engine (this site's or Google's) to help someone find it. IMO old questions should be sorted by last non-migration activity. They do not 'clutter' the archive - the archive is the jewel-filled chest that gives increasing value to the StackExchange sites. Would you rather have someone come to tex.sx, search for a question, find no match, ask it and have people answer it...or have them search and immediately get the satisfaction of finding an appropriate answer that happened to have been migrated from another site? –  Phrogz Apr 8 '11 at 20:16
    
@Phrogz: I'd rather have them ask a fresh question, really. One reason is that you still have all the dupes, and noone will see that they're dupes since they don't get to the top. I do see your point, but I don't see a good solution so far. –  Hendrik Vogt Apr 8 '11 at 20:20
add comment

Here is the most comprehensive "Does LaTeX belong on Stack Overflow" question from before TL&F came on line. My answer explains in detail why I thought that LaTeX was an appropriate subject for SO.

It also expresses my new position which is that LaTeX questions are still OK on Stack Overflow, but are better on TL&F.

I am not opposed to a migration path, because there are many LaTeX question that don't really have the programming nature. But I do oppose mass migration of all existing question and any general claim that LaTeX does not belong on SO.

share|improve this answer
1  
It probably won't surprise you to learn that I disagree with your answer there. I think that the factor that you do not address is the position of the questioner. I would hazard a guess that the majority of TeX questioners are completely ignorant of the fact that TeX is Turing complete and regard "Use \newcommand\Foo{\foo}" as black magic. In addition, I would not be very surprised to learn that programmers keep an eye on languages outside their speciality, but I would be surprised to learn that TeX was included in these. (ctd) –  Andrew Stacey Nov 18 '10 at 19:55
4  
(ctd) So I would say that TeX questions do not belong on SO because neither the TeX-users nor programmers regard TeX as a programming language. The fact that it actually is, is irrelevant when compared to this. –  Andrew Stacey Nov 18 '10 at 19:56
2  
@Andrew, I would hazard a guess that the majority of TeX questioners are completely ignorant of the fact that TeX is Turing complete - I'm not sure, but I think that's not true on SO. If you look at the Latex answers of Alexey, you'll find most of them pretty programmatic. –  Charles Stewart Nov 19 '10 at 11:55
    
@Charles: Two possible responses. The first: "one swallow does not make a summer", the second: "I guess I should get some data before theorising, but I am a mathematician so it's hard to go against my training.". –  Andrew Stacey Nov 19 '10 at 21:31
    
@Andrew: To clarify: I'd be happy to see people sort through the SO [latex] questions and migrate the NPR ones---I'd even spend close votes to help. But judgment should be used, rather than just programatically grabbing everything with a [latex] tag on it. The thing is this determination is harder than it looks...the asker does not always know at first if the problem will require non-trivial programming, and then there is the matter of limits: does "Get PackageFoo from CTAN, define a couple of lengths and wrap these commands up in a \newenvironment..." count as programming? –  dmckee Nov 19 '10 at 21:43
3  
@dmckee: One thing that I want to make very clear is that I am most definitely not in favour of any automatic system. What I would like is to swing the balance a bit more in favour of the daughter site rather than the mother. A few extra TeX questions probably doesn't bother you too much, but those few are vital publicity for us. On your last sentence, I would argue that if the asker does not know that the problem will require non-trivial programming, then the question is not an SO question - even if it does! It is not the answer that should determine the correct home (ctd) –  Andrew Stacey Nov 19 '10 at 21:52
1  
(ctd) but the question and the questioner (by which I mean that certain qualities of the questioner should be considered, not the questioner's opinion on where it should be answered). On MathOverflow, I've come to realise that the questioner is actually a big part of the question, and so if the questioner doesn't think of TeX in a programmatic way, then it isn't a programming question even if it blatantly looks like one. –  Andrew Stacey Nov 19 '10 at 21:55
2  
@dmckee: Let me outline my proposal again: I (as a tex.SX mod) spot a question that I think ought to be on tex.SX. I flag it. The fact that I'm a tex.SX mod means that someone from SO quickly takes a look at it. The SO person (may be a mod, may be a high ranker, may be more than one person) decides whether or not the question belongs on SO or not. If not, it gets migrated to tex.SX. If yes, I get a response outlining why (so I know better for next time). That way, the expertise and the incentives are all in the right places. –  Andrew Stacey Nov 19 '10 at 21:57
add comment

You must log in to answer this question.

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