# What do you think is the largest problem facing Math.SE today (July 2015)?

I think it can safely be said that Math.StackExchange (and MathOverflow and perhaps some of the sister sites on the SE network) are the best resources for (English-speaking) people with objective math questions on the web. There are other math sites, such as those mentioned in Useful Mathematical Fora (and some of them, like AoPS, provide extremely valuable resources to the greater mathematical community).

But MSE is not perfect. For instance I can think of 10 problems facing MSE right now without working too hard:

1. There is a learning curve here. Many new users do not know (or read) How to ask a good question? and correspondingly many new users first experience with the site is getting downvoted. We desire questions written in MathJax, which itself has a learning curve (although we seem pretty good about editing it in for new users). Some might ask Are we too trigger-happy about deleting (relatively) new posts?
2. Even for established users, there can be disagreements. Sometimes, someone notices that Someone is wrong on the internet. Tensions fly, and sometimes it's hard to Be nice. How many people find MSE a comfortable place to work and interact? We are a community.
3. Roughly speaking, there are 3 people who ask a question for each 1 person who answers a question each quarter. [This and other statistics are at this Community Blog Post]. Further, very many users ask many questions and relatively few (by comparison) users answer many questions. How many find MSE to be a place that gets answers quickly and effectively?
4. For those that do ask and answer, are we voting in a way to both incentivize good behaviours and to distinguish between good and bad content? Should everyone need to reread Vote early, vote often and vote more? [This was another theme of the same Community Blog Post].
5. MSE is not just about getting answers to new questions. It is also about serving as a repository of good questions and answers from the past. Unfortunately, searching for past questions can be really, really hard since search engines don't index on LaTeX. While having informative titles and believing in our search engine of choice is our working solution, let's be frank: searching for math online (and offline) is an awful experience.
6. Along similar lines, we have a growing base of Abstract Duplicate questions. Some users diligently maintain our List of Generalizations of Common Questions (so that we have canonical questions/answers like Single Variable Calculus Reference Recommendations), but many do not.
7. The site is growing rapidly, and for whatever reason we have an extremely high questions/tag ratio compared to other sites. This can mean that organizing newly asked questions into digestible chunks is hard. Does the typical user have a way to view and find questions/answers that intrigue them? This is especially desirable for users to peruse questions of their interest and for power-users who answer very many questions.
8. The meta-tags and are increasing in popularity. These tags are a bit against the organizational grain of the site. See Would a tag for "check-my-proof" questions be useful? and linked meta threads. (While this isn't actually a pressing issue, I conjecture that in time will come back and burninate these tags).
9. There is a contentious (for lack of a better phrase) downvote-for-the-sake-of-deletion campaign from some users. This prompted meta threads such as Under what circumstances is it appropriate to delete a question that has received a good answer? and Is it appropriate to downvote answers for the sake of deleting a question? Roughly speaking, we might hope that users edit poor questions that receive good answers into good questions for the longevity of the site (or more reasonably, those who answer poor questions might edit them into good questions; this behaviour is incentivized by some relatively recently added badges). But in fact every once in a while, some upvoted answers are deleted as some users remove poor questions.
10. This site is being moderated by 9 semi-tyrannical volunteers who sometimes make mistakes and who each have fluctuating commitments. Historically, there has been friction between moderators or perhaps between some of the mods and some particular users (although right now it happens to be that the mod team gets along with itself extremely well).

You might find other problems. That's what this post is for!

What do you find to be the largest problem facing MSE today? Why is it a problem, and is there something to be done about it?

• Some might think it would be a good idea to restrict to solvable problems. But I think sometimes it's good to get a holistic understanding from the community. For instance, I think that (1) having a math search engine, (2) having many more people vote much more often, and (3) having new users read the how-to-ask-a-good-question guide would make the site run tremendously smoother. But the first is a white whale, the second is a golden fleece, and the third is at best a dark horse. – davidlowryduda Jul 6 '15 at 16:45
• re 9: I think the aspect of down-votes to delete is not really a main issue anymore. The conflicts these days are about usual deletions. Further, you link twice to the same post. Finally, to use the loaded word "campaign" seems not optimal to me. – quid Jul 6 '15 at 17:24
• Do you find it loaded? I meant it in exactly its definition: work in an organized and active way toward a particular goal, typically a political or social one. It's profoundly well-organized, and my opinion towards it isn't well-established either for or against. I suppose that those who complain about it usually reference it by a campaign. but I attribute no implicit value to the word. – davidlowryduda Jul 6 '15 at 17:30
• It is the wording used by those that complain about it, thus at least in this context it is loaded. I also would not know how it is "profoundly well-organized." Is it organized at all? A couple of occasional chat messages can hardly be described as organized. Anyway, while my delete votes are considerate (at least I claim so) they are in no way organized. – quid Jul 6 '15 at 17:39
• Regarding #2, I find MSE itself fine (even if there are occasional problems), it's on meta that things get out of hand often. (This is probably compounded by the fact that when things go bad on main, meta is often the next step.) – Najib Idrissi Jul 6 '15 at 18:39
• You're missing scope. Many of the points in your list, I believe, arise mainly due to a fundamental disagreement about the scope of MSE. Some people believe that MSE is a place for everything under the sun, regardless of quality, (others') ethical concerns, and usefulness, up to and including doing people's (graded!) homework for them. Others believe the scope should be more limited or that being so all-inclusive is counter-productive. – user14972 Jul 6 '15 at 19:22
• If I had an upvote each time someone asked us to evaluate an integral of the form $\displaystyle\int_0^\infty\frac{x^k}{1+x^n}dx\ldots$ Oh, wait, I sort of do! :-$)$ – Lucian Jul 6 '15 at 20:28
• Needs more cowbell. – Asaf Karagila Jul 6 '15 at 21:59
• @Hurkyl oh, that's very true. That's a very big point for me to have left out. – davidlowryduda Jul 6 '15 at 22:34
• I must confess, the sheer volume and frenzy on main is starting to discourage me from contributing as actively as I once did. It used to feel fun; now writing here feels like a chore. – J. M. isn't a mathematician Jul 11 '15 at 5:13
• @Guesswhoitis. My antidote to frenzy is the search parameter created:2d. I get a static snapshot of what was posted two days ago (using UTC), then slice and dice by tags, score, answered status... no frenzy, just a set of questions patiently waiting for an answer. – user147263 Jul 11 '15 at 5:59
• What about integrals question ? I feel that this is the only subject (modulo some questions ) here that one doesn't need to provide context, work to get lot of upvotes. On the other hand, for these questions, answers with only a closed form and/or numerical values should not be called as an real answer (my opinion). – user169373 Jul 25 '15 at 11:10
• @Guesswhoitis. Migration of poorly asked question from Mathematica doesn't help... in the last 90 days, 37% of migrations from MMA were rejected by voters, which is higher than from any other site with nontrivial number of migrations. – user147263 Aug 4 '15 at 19:54
• A moderator whose profile picture consists of an angry Japanese military general of some sort can hardly be considered only semi tyrannical :) – uniquesolution Nov 7 '15 at 10:34
• Approach0.xyz really seems to solve the problem #5. – Workaholic Dec 6 '16 at 10:13

I would say that the OP has identified what I feel is the site's main challenge in Point #5: it is difficult/darn near impossible to search on questions and answers within this site. Many other difficulties, e.g., duplicates, high # questions/answer, infighting (e.g., is that really a duplicate?), etc., are a result of this challenge.

If the site could index searches over MathJax, then many questions of the sort "evaluate this sum/integral/product" needn't be posed. Now, I am not sure how good such searching would have to be to eliminate duplicates for the non-lazy, but even something requiring an exact match would be an improvement.

Think about it. Many of the problems posed daily in the most popular tags are essentially duplicates. However, because of the poor searching facility, the average user has little chance of finding that out. We tend to rely on our gold-badge-holders to use their unilateral powers to sort out duplicates.

If someone can institute a MathJax search facility that is transparent to the user, then many other problems will go away.

• Many of the problems posed daily in the most popular tags are essentially duplicates - You mean like this one ? :-$)$ – Lucian Jul 6 '15 at 20:41
• SE is unlikely to develop a math search engine themselves, it is just too far from their core business and expertise. The best chance to get SE to add site-specific features is to suggest an external solution that could be integrated. I don't know if something like this exists in this case, though. – user9733 Jul 7 '15 at 10:52
• @MadScientist this goes somewhat in this direction trac.mathweb.org/MWS On this site though it might be especially tricky as the TeX-nical quality of the math is quite uneven and MathJax is (ab)used in various ways. – quid Jul 7 '15 at 11:54
• @quid That project looks pretty abandoned – user9733 Jul 7 '15 at 12:06
• @MadScientist perhaps I did not chose the best link and I really do not know if there is still active development. But the thing is used on zbMATH zbmath.org/formulae which is a main bibliographic database for math. Also see kwarc.info/node/12372 and more generally kwarc.info which seems quite active. (Sorry for the double post.) – quid Jul 7 '15 at 14:13
• One possible solution would be to improve the tagging system - perhaps add "sub-tags" - for instance, "$L^p$ spaces", "non-measurable sets", "sigma algebras", etc. could be sub-tags of measure theory. Clicking on a tag could then lead you to a page listing all the sub-tags, which could then link to the questions. – Prahlad Vaidyanathan Jul 20 '15 at 5:04
• I totally agree this is the main problem, and a huge one. Has anyone ever proposed to institutionalize the List of Generalizations of Common Questions (it would be better to change the title in Frequently Asked Questions, IMHO) and make it more organized/accessible to newcomers? – Jack D'Aurizio Jul 20 '15 at 10:16
• I guess that user who regularly visit meta must have noticed it. But perhaps it is worth mentioning also here this relatively recent post: Announcing a third-party search engine for Math StackExchange. Maybe after some time we will see how this project works out. – Martin Sleziak Sep 15 '16 at 15:58
• @Nemo: I agree. I have seen this behavior from many folks who should know better. When I suspect a dup, I find a solution that I did to the problem and post a link in the comments. That way, I answered the question without posting a dup solution. And if my solution gets a little more notice then at least it is my original solution. – Ron Gordon Jan 12 '17 at 14:22

A great problem is the lack of specialists in advanced subjects.
And that problem is getting worse. In algebraic geometry (my field) some of the best experts have left the site : mature prestigious mathematicians like Matt Emerton or Qing Liu, and very promising younger people like Asal Beag (who deleted his account).
This is catastrophic because many hard questions, at the graduate level say, will consequently not be answered here, nor on MathOverflow if they are not at research level.

Edit
To prevent misunderstandings, let me emphasize that I find many questions in the fields I'm concerned with on this site remarkably interesting, mature, original and intelligent.
They attest to the infectious dynamism, curiosity and enthusiasm of the askers.
I have found answering these questions the best possible use of my time and the best way to learn mathematics: it is an underappreciated truth that the best way to understand a subject is to explain and teach it.
What I deplore is the departure of the outstanding answerers evoked above.

• I think that MO is receptive towards grad level questions, provided they are well written and it's clear that the OP is interested in learning, not just receiving an answer. – Asaf Karagila Jul 7 '15 at 10:15
• Of course, this differs from area to area. From what I can tell, there are enough expert set theorists active on the site. General topologists, too, although graduate-level questions in general topology seem to have become an extreme rarity here. – user642796 Jul 7 '15 at 10:15
• What do you think is the reason for the (perceived) lack of specialists? – quid Jul 7 '15 at 10:19
• As an aside, we might guess at natural reasons for specialist shortages: 1) specialists are likely to be academically very busy IRL and 2) such users may find that writing high-quality answers for upper-level questions very energy consuming and 3) questions catching the fancy of a particular specialist may occur at such low relative frequency that the specialist gets out of that habit of listening and participation drops. – rschwieb Jul 7 '15 at 15:28
• I've thought about this. It's very easy to find easy low-level problems on the site. It's relatively hard to find interesting high-level problems on the site, especially if they're a bit old. I bet this seems much harder for a new visitor to the site --- to fresh eyes, the typical page is full of early undergrad and high school problems. Would you agree? – davidlowryduda Jul 7 '15 at 16:47
• I agree that this is a problem. I try to use my votes to reward questions and answers in more advanced topics (within the tags that I'm semi-familiar with). But that is just a drop in the ocean. BTW Asal Beag Dubh means "Little black donkey" in Irish, so it is safe to assume it was an alias. Also this. All the users you listed are missed. – Jyrki Lahtonen Jul 7 '15 at 19:23
• Thanks a lot for your very interesting and much appreciated comment, @Jyrki. – Georges Elencwajg Jul 7 '15 at 20:44
• I'm curious as to why those people you mentioned left. – user21820 Jul 12 '15 at 7:27
• @rschwieb, I could probably try to actively search for questions in just the two tags I am most interested in, but even then, the results are polluted by people who couldn't be arsed to tag properly, which has the net effect of putting me off. Which brings me back to my comment in the question: I came here because posting was fun, but now it feels like drudgery. – J. M. isn't a mathematician Jul 29 '15 at 14:02

The biggest problem (at least of those not mentioned so far) is that there is no optimization for pearls. This has a some different facets:

• Many of the questions we receive are just special cases of broader concepts.

Calculating a specific integral is unlikely to be helpful to future visitors unless that problem comes from a standard text (so that the same problem is likely to have been at least recommended to thousands of students). Unfortunately, even in this case future visitors are unlikely to find these questions because of both the difficulty in searching for mathematical formulae, but also the fact that we don't insist that users provide sources for their questions.

At the same time, we spend approximately zero time providing extremely good Q&As for more general concepts/techniques. An awesome question and answer(s) about integration using partial fractions would be a much better resource than hundreds of specific questions that no-one looks at.

• We don't edit questions nearly as much as we should, and when we do these edits are often incomplete. Instead of simply introducing $...$ MathJax delimiters, we should be spending time improving the organisation, English, etc. of questions to make them as good as possible.

Remarkably, of the 20-odd users who have posted at least 500 answers so far in 2015 and had not already received it in 2014, not a single one has earned the Illuminator badge. As of this writing only Davide Giraudo received the Illuminator badge this year, and has posted about 200 answers so far in 2015.

• Our love of "HINT" answers means that many of our answers are intentionally incomplete. While there are some users who have the ability to provide worthwhile hints, this is a very small minority among those who try. As such, many answers are really poor, cryptic references to potential answers. But even good hints made with the intention to directing the OP to the answer should be filled out at some point. (Well, IMHO.)

Even when questions achieve a high score, it seems that their pearliness is still highly questionable. It is perhaps illustrative to browse the questions found when searching is:question created:2015-01-01..2015-12-31 score:50.. on various Stack Exchange sites:

• I'm like... THIS close to the Illuminator badge! – Asaf Karagila Jul 7 '15 at 10:08
• I disagree about hints. I think that any reasonable (not just good, or great, hints) should be left as hints. If anything, in the long run hints can be as useful and perhaps pedagogically more useful than just writing the answer. If I were to transform every hint into a pedagogically complete answer, I'd probably spend the next couple of years in editing. Frankly, I'd rather spend this effort into writing a book about set theory (intro, intermediate, advance, all three?). – Asaf Karagila Jul 7 '15 at 10:12
• I think the "HINT:" style is just terrible. By contrast, I have no problem with terse answers that leave out details. – quid Jul 7 '15 at 10:23
• I prefer hints to be in the comments, leaving the answer box for actual answers. I also prefer when the OP follows up in a hint in the comments, but this happens all too rarely. – Tobias Kildetoft Jul 7 '15 at 10:24
• To be fair, even on SO with its millions of users, only 30 illuminator badges have been awarded. I agree with everything else. – Najib Idrissi Jul 7 '15 at 10:25
• @Asaf: You might be able to give good hints, but then users who are decidedly not as able begin to give their "hints". I used to think hints were great. Then I became a mod and began to see a lot of crappy hint "answers", and wondered why. I believe it because higher rep users (who often post hints) are inadvertently acting as role models, and so others try to copy their style of answering. If answers were answers, I think we'd see a lot fewer crappy answers. Peruse our shortest hint answers for a taste. – user642796 Jul 7 '15 at 13:31
• Integration by partial fractions; how and why does it work? does exist, written by a former user who cared about some of the things on your list. – user147263 Jul 7 '15 at 16:32
• Indeed, I also used to like hints. But not anymore. @1999: In fact, it is also included on meta.math.stackexchange.com/questions/1868/… as a canonical best answer. – davidlowryduda Jul 7 '15 at 16:45
• A big problem with hints is that its providers are sometimes unable to give a complete solution themselves. This is clear when they admit they can't give details after being challenged or when they give hints for proving statements which happen to be false. – Georges Elencwajg Jul 8 '15 at 7:18
• I’ve already upvoted his comment, but to make it explicit, (to no one’s surprise) I agree with @Asaf about hints. I also disagree to some extent with Arthur’s call for more thorough editing of questions. I will edit them to make them understandable, but as an answerer I find it helpful to see the original wording: it can give me an idea of the asker’s level of knowledge and facility with English. – Brian M. Scott Jul 10 '15 at 0:53
• Off-topic: search parameter created:2015 is easier to type and does the same thing. – user147263 Jul 12 '15 at 0:46
• @GeorgesElencwajg: Yes there is that unfortunate tendency to make mental errors when trying to give hints. The solution is to actually write out a full solution before cutting out parts that one feels the asker can feel in or even cutting it further down to hints. – user21820 Jul 12 '15 at 7:10
• @Arthur: Some of those short hints are actually pretty good, and others are reasonable, if skimpier than I would offer. In any case, the fact that some people overestimate their ability to give hints isn’t really much of an argument against hints per se, any more than the fact that some people overestimate their ability to give answers is much of an argument against answers per se. – Brian M. Scott Jul 13 '15 at 2:58
• Am I the only one who thinks the Illuminator badge is somewhat of a rat race (in the sense of: more of a rat race than the average SE badge)? "Edit and answer 500 questions (both actions within 12 hours, answer score > 0)." Sometimes a good answer takes more than 12 hours to find and write up (often aided by the fact that the original version of the OP is unclear and needs author input to even become a question), while a good edit can be made right away. – darij grinberg Jul 18 '15 at 23:38
• Another point about "HINT", sometimes users give an answer with this and doesn't reply in any, real, question by the OP to get the final result. And I feel that if we show our work, a "HINT" answer is inappropriate. – user169373 Jul 25 '15 at 11:14

I think that regarding point #5, there is a disconnect between what MSE purports to be and what its users (both new and established) want it to be.

As mentioned by mixedmath, the official policy of the site is the goal to become a "repository of good questions and answers". But that doesn't fit with the new user who wants a quick homework answer, nor it fits with the long established behavior of asking "what have you tried" or answering with hints.

• I strongly agree with this. Even more, I would say that that this is a wonderful technology for interactive discussion amongst folks interested in mathematics at a wide range of levels. It is much better as a discussion forum than tools (like usenet) that preceded it. By contrast, it really contributes very little as a "repository" as compared to say, Wikipedia or the library. I would say it's much worse as a repository, in fact, as many answers (even accepted ones) are just plain incorrect and the overall quality is very uneven. I think math.se would be well advised to focus on it's strength. – Mark McClure Jul 16 '15 at 15:21
• @MarkMcClure The site has about 100,000 visitors daily. Of them, about 500 ask a question, and another hundred or so answer questions. What do you think the rest are here for? – user147263 Jul 16 '15 at 16:46
• @1999 Arguing on meta, of course. – Najib Idrissi Jul 16 '15 at 17:26
• @1999 Downvoting? I guess in a futile effort to organize this site? I noticed a post this morning asking a general question on generating functions. How could anyone here possibly provide a better explanation than that found in Concrete Mathematics or Generatingfunctionology? To answer your question directly, I suspect that the vast majority of the visits are drive-bys with most visits lasting a few seconds - a common occurrence around the web. – Mark McClure Jul 17 '15 at 0:04
• @MarkMcClure That's not a particularly useful question, "intuition behind blah is blahblah". Stack Exchange is designed to cover the long tail of specific questions that do not warrant a Wikipedia article. An example of Banach spaces for which the Banach-Mazur distance is not reached... the proof of compactness of an embedding between Hölder spaces... (taking random examples from functional analysis). None of them individually are of interest to many... but there are many of them. There's a reason why this site has traffic of order 100,000 while, say, History of Science and Mathematics is of order 100. – user147263 Jul 17 '15 at 0:26
• @1999 I agree - that question is not particularly useful on this site. That's one reason why this site isn't particularly useful as a general repository. More generally, I don't see that this site has added anything particularly novel in this area. The questions you suggest are much more specific and lend themselves more readily to a discussion forum. This site is great for that! – Mark McClure Jul 17 '15 at 0:44
• @Mark: Any competent teacher with a reasonable amount of experience knows (or should know) that there is enormous variation in the explanations that individual students find helpful. I agree that the explanation of generating functions in Concrete Mathematics is very good, but I know from experience that it isn’t sufficient for some students. It’s entirely possible that someone here can provide an explanation that works better for the person asking the question. – Brian M. Scott Jul 19 '15 at 6:23
• @Mark: This site is more useful as a general repository than you seem to think: it regularly turns up in Google searches on various mid-level mathematical topics. – Brian M. Scott Jul 19 '15 at 6:24
• @Brian: Google search results depend on your online activity, your geographical location, and so on. – Asaf Karagila Jul 19 '15 at 8:54
• @Asaf: I know. That doesn't change the fact there's a lot of useful information there, and that it is accessible, albeit probably with different levels of ease for different users. (By the way, my Kindle thinks that Asaf ought to be A sad.) – Brian M. Scott Jul 19 '15 at 9:02
• @Brian: That's true. – Asaf Karagila Jul 19 '15 at 9:07
• @BrianM.Scott Thanks. I always appreciate your input. I don't mean to imply that math.se is completely worthless as a repository; there is clearly information here to be found. I'm simply saying that it is a better discussion forum than a repository, in spite of the fact that this goes against it's stated objectives. That's a primary reason math.se is able to "provide an explanation that works better for the person asking the question"; the discussion provides feedback on the person's background. – Mark McClure Jul 19 '15 at 10:17
• @BrianM.Scott I'm also saying that these two objectives are at odds and the fact that the "repository" objective is held above the "discussion" objective damages the sites effectiveness as a discussion forum. The fact that some folks dislike shorter hint answers and that some folks downvote simply because it helps to organize the site both make some sense if you want a repository but, I think, are harmful to a discussion site. – Mark McClure Jul 19 '15 at 10:18
• @Mark: I don’t think that they’re entirely at odds, but they certainly do impose different priorities, and I agree with the rest. – Brian M. Scott Jul 19 '15 at 18:53

The way this site interacts with new users seems to often lead to one of the following outcomes:

• a new user posts a poor question and receives an answer nonetheless (even if the question is later closed).

• a new user posts a poor question and is downvoted into oblivion.

The first outcome encourages dishonest askers and the second outcome discourages honest askers.

Largely, I think the problems stem from the somewhat parallel systems of downvoting and closevoting. The latter system works efficiently and effectively, and I hope that question triage will further improve things by reducing the visibility of questions that will ultimately be closed (reducing the prevalence of the first outcome). The former system is more troublesome - once cast, downvotes don't tend to get reversed, regardless of the author's future actions - which serves to discourage authors who do make improvements, but do too little against authors who don't - downvotes on question sort of act like closevotes that don't work as well, but are more widely accessible.

• Wouldn't it be nice could some agreement be found that poor question are not answered right away and in turn they are not down-voted right away? Do you think a compromise even only among the more meta active users on this could be found? – quid Jul 7 '15 at 1:08
• @quid I think it would be necessary to change the mechanics of the site before any compromise can be made. Triage, if it works as intentioned, may do a lot of good, since the people who even see poor questions will be more invested in the site (i.e. sufficient rep & active reviewers) - not to say such users aren't part of the problem, but they're not all of the problem either. I would like to see the effect of downvotes on questions rethought too - users are right to mark poor quality questions for the system, but, unfortunately, the system uses that data in a nonconstructive way as it is. – Milo Brandt Jul 7 '15 at 1:37
• @quid: I don't think you can reach this level of understanding on a site this big. Then you run the risk of people ignoring community "standard", and doing thing their own way. – Asaf Karagila Jul 7 '15 at 5:54
• @Asaf and Meelo: while I do think that what I propose would be nice, I am aware that it will not work things being as they are, and a significant part of the point of the comment is as part of the discussions on my answer. On the remark that the site is big though, I would reply that I restricted to the "more meta active users" which is really not that big a group. Would there be some consensus it would trickle down over time. A problem is that casual users get completely mixed signals what is and is not appreciated on the site. – quid Jul 7 '15 at 10:30

Not enough meta involvement.

I don't know how many users regularly check the meta site, read the discussions and vote on issues, but it can't be enough. Just looking at the fact that most posts on meta get just a little attention. Not plenty of votes, not plenty of views.

We need to encourage more users to read questions on meta more regularly. Both to learn about new and repeating issues, and to be slightly more involved with this site (which will have positive implications on moderator election as well).

• Dear Asaf, what an amusing contrast : I have just concluded in view of recent developments that I should reduce my involvement on meta to zero, except for technical questions on the mechanism of the math.stackexchange site. – Georges Elencwajg Jul 7 '15 at 11:56
• I do not see any contrast at all. Rather I see an example of the problem described. – quid Jul 7 '15 at 12:21
• @Georges: I wrote to someone on meta.MO recently, I don't know many professional anything with penchant for drama. So I understand your comment. Unfortunately, the fact most people ignore meta-related discussions is bad. I would be okay with that, if people who actively follow (I'm not even saying participate in anything more than silent voting) meta would be granted extra rights, like electing moderators. Your comment is a direct connection between my answer and quid's answer. – Asaf Karagila Jul 7 '15 at 12:32
• I don’t read it unless something that crops up in the sidebar on the main site drags me over, or we’re in a moderator election season. Frankly, I think that many of the non-technical questions are a complete waste of time, and that some of the people whose names I see here just about every time I drop in are busybodies. (That emphatically does not include you. Or, perhaps surprisingly, the Man of Many Names.) – Brian M. Scott Jul 10 '15 at 1:09
• @BrianM.Scott I think it is not an uncommon phenomenon that one can be easily annoyed when one finds one's on characteristics in others. Perhaps surprisingly I am not out for a fight here but I think it might do you some good to reflect on the fact that you at the time you wrote this comment, also commented in a confrontational way on a completely resolve, friendly and tangential debate that did not concern you at all. This type of behavior is just the one I found described when looking up "busybody." – quid Jul 10 '15 at 10:14
• @quid: I did no such thing. I commented on precisely the issue in question, giving one of the most important reasons for my relative lack of participation here. I certainly did not comment on any issue that is completely resolved. And I’m more than fed up with your holier-than-thou pose, which is very much at odds with your actual behavior. – Brian M. Scott Jul 10 '15 at 18:17
• @BrianM.Scott at the time (minutes apart) you wrote the comment on this post you also wrote in this very thread another comment. I was talking about that comment. – quid Jul 10 '15 at 18:33
• I agree with @BrianM.Scott. There are just way too many wannabe politicians on meta, trying to make rules for everything, no matter how trivial it may be. – Mister Benjamin Dover Jul 14 '15 at 10:21
• @Ben: Unfortunately these rules come organically from interactions people have on the site (both main and meta). – Asaf Karagila Jul 14 '15 at 11:11
• @Ben: meta.math.stackexchange.com/a/21061 something without drama. Well, mostly. – Asaf Karagila Jul 22 '15 at 22:12
• Personally, I think it is a good thing if some people are not involved in meta. If some users concentrate more on their contributions to the main and are not distracted by bureaucracy and issues related to running the site, it is a good thing. If there are some users willing to spend their time in some meta discussions which might be, in the end result, useful, that is fine two. The slogan is: meta means murder. Shamelessly stolen from this answer, which is where I saw it for the first time. – Martin Sleziak Oct 23 '15 at 11:46

Solution verification and proof critique.

This is something I go back and forth with. On the one hand, we should encourage people to post their own proofs, and it is a good thing when you can post a proof, and knowledgeable people will read it and comment.

But (1) this has very little archival value as presented, if people are so bent on creating a repository of questions and answers. (2) "Yes" is sometimes a sufficient answer to the question "is my proof correct?" and there's very little to add; sure you can post an almost identical proof written with your own style, preferences and idiosyncrasies, but it's not quite a proof verification, is it now?

So I'm honestly wondering, what should be the fate of proof verification and proof critique questions. I feel that this is an often ignored problem on this site. It probably deserves its own thread, though. Much like all the other problems posted on this page.

• I have strong opinions on this, but I also think this is something that not so many others are worried about. In short (and because I do not want to create another thread that demands thought now), I have wondered whether these posters should just phrase the question as a question, and then write their (attempted) answer as an answer, and have others say what they say. A possible downside is that wrong answers get downvoted, and many deleted downvoted posts lead to people losing answer abilities. This is just a quick thought. – davidlowryduda Jul 8 '15 at 0:03
• You can't lose answering abilities. Just saying. Also, as far as the Trogdor universe, if I recall correctly only peasants are on fire (the rest of the burninated folks die off pretty fast). – Asaf Karagila Jul 8 '15 at 0:13
• I had to make sure, but it is possible to lose answering abilities (at least temporarily). I recall seeing this happen, and when I asked some other mods (from other sites) they recall it happening. But when I just looked to find someone blocked from answering, I didn't find anyone. [I did find people who were throttled, and people who were blocked from asking. But not answering]. – davidlowryduda Jul 8 '15 at 0:25
• Are they really of such little archival value? If a proof is correct and well written, it can be a valuable item (depends on what's being proved, of course -- but the same holds for all questions). From the Google point of view, there isn't much difference between proof being in question box vs in answer box. If the proof is not correct, there will probably be a correct one added. The discomfort of posting short answers is something for answerers to deal with; I don't find it impossible to add a couple of lines. Overall, this doesn't quite rise to the level of the largest problem. – user147263 Jul 8 '15 at 2:14
• @mixedmath: Oh, you mean answer ban. I guess you can lose that privilege. I see what you mean now about downvotes. – Asaf Karagila Jul 8 '15 at 4:45
• @1999: Admittedly, answerers make up only 1/3rd of this site (or even less if you consider regular contributors), but I don't see how something which is essentially a problem of all answerers is not a large problem of the site. – Asaf Karagila Jul 8 '15 at 4:47
• re 2: in my experience normally for such questions when they are asked here there are plenty of things that can and should be said beyond "yes." For more details see this meta-answer I wrote some time ago. – quid Jul 8 '15 at 13:28
• @quid: As I wrote, I agree that this is often the case, but it's not uncommon to find a reasonably written proof. Since the question is about proof verification, I sometimes find it in bad taste to write my proof (unless I feel that the OP lacks any style). So to a proof that is written reasonably well, and I have no stylistic comments, what would I write in an elaborate answer? – Asaf Karagila Jul 8 '15 at 16:22
• I’m in favor of such questions. I’ve not actually counted, but I think that at least half the time there’s some infelicity that could be commented on, some shortcut or other improvement that could be suggested, or the like. If not, though, I’ve no problem simply leaving a comment and an upvote. – Brian M. Scott Jul 10 '15 at 1:13
• @Brian: I don't have a problem with the concept. I am always very happy to see people who actually tried to do the work. It's just that... I'm not sure what to do with these questions after the basic critique has been given and received. Do we close them as duplicates (e.g. how many nearly identical proof verifications of basic set theoretic questions have we seen? Many, that's how many. After the basic critique has been given, should they stay open)? Do we leave them? Do we edit them somehow? It makes me a bit uneasy. – Asaf Karagila Jul 10 '15 at 13:26
• @Asaf: I don’t mind leaving them open; it doesn’t matter much to me whether the useful information is in the question or in an answer. My real preference, I suppose, would be for the asker to transfer the argument to an answer, but I see no practical way to make that work in general. – Brian M. Scott Jul 10 '15 at 18:11
• Re: your first point. There are users (who shall go unnamed) that post questions with the intent simply to answer them. I get that this technically is not in violation of site rules but it really seems against the spirit of the site. Occasionally, someone comes across something cute that they want to share and a Q/A format is reasonable; however posting question and answer after question and answer feels like it's just a way to post homework problems (or something of the sort). Blogs are best suited for this and I'd really like us to amend the rules to omit such behavior. – Cameron Williams Jul 15 '15 at 4:15
• The suggestion given in this answer was: "Close as a duplicate after the OP had sufficient feedback." – Martin Sleziak Oct 13 '15 at 11:59
• My biggest problem with proof-verification questions (apart from the few crank cases of check-my-proof-of-Famous-Conjecture) is that they're usually asked about homework, which means that there's a teacher somewhere who is getting paid specifically to attempt to understand the proof and tell the asker what problems (if any) there are with it. It seems like a waste to tie up the site's volunteer resources simply so the asker can have his curiosity satisfied a bit earlier than he would otherwise in the due course of things. – hmakholm left over Monica Nov 7 '15 at 13:59
• @Henning: While I am willing to help my students review their proofs from time to time, I doubt that I'd be that willing if all 120 of them came to me with these requests every week. And most people I know in academia will be willing to help on occasion, but not on a constant basis. For most students the feedback comes from homework graders, who often have significantly less experience in writing proofs let alone critiquing them, and even then many times the grading is partial (you only grade some questions, over a limited time frame); or in exams, which is essentially "too late". – Asaf Karagila Nov 7 '15 at 14:02

A main problem of the site is that there seems to be too little interest in finding compromise solutions (a policy on deletion, for example), and most energy is spent to establish one's own point of view or dually to denigrate and belittle opposing views.

A related problem, augmenting the former, is that even in the case where some consensus is found, some will not feel bound by these agreements and will continue to do things their way.

This creates a climate of confrontation that makes finding compromise solutions even more difficult.

• One has to wonder if the downvotes come from people who honestly disagree, or just people who feel [rightfully?] offended as supposedly singled out as a problem on this site? – Asaf Karagila Jul 6 '15 at 22:32
• This post is way too vague; I don't know who you're talking about, whether you're referring to main or meta, or what ideal you think we're straying from - and it's not been my experience that the site has an overly confrontational culture, so some explanation would be very helpful if I am to see what you're saying. The post seems more to be taking a jab at some unnamed users than illustrating a problem. – Milo Brandt Jul 7 '15 at 0:09
• @Meelo it is not taking a jab at some unnamed users, but rather at some named behavior. If you did not notice or experience it, good for you. What I am talking about is for example: voting four times on the same post or calling others' comments/opinions 'utter bullshit', refusing to revise or reask question that was badly received, but also insisting to delete a post that is not worse than thousands of other posts when somebody cares about that post specifically, etc. – quid Jul 7 '15 at 0:19
• For an example of what would not be an "ideal" but at least something: on MO there used to be at some point massive conflicts over deletions. So some discussions took place and a compromise was reached, between those that did not want anything deleted and those that wanted many things deleted, that was followed without too many issues for quite some time. (Then things took still another turn.) In my observation, here, there is not much reasonable discussion what should and should not be delete to be had at all. This is not to say that MO is somehow "better" but this thing was resolve @Meelo – quid Jul 7 '15 at 0:36
• @quid So, to paraphrase you, would it be fair to say that the problems are somewhere along the lines of meta discussions not being able to reach consensus (due to behavior obstructive to the goal) and users not treating the communal consensus as binding (for lack of a better word)? (If so, I really think you should edit the ideas from your comments into your post - I get wholly different impressions from the post and the comments as it stands) – Milo Brandt Jul 7 '15 at 1:09
• @Meelo yes, basically. I tried to reformulate it to make this more clear. – quid Jul 7 '15 at 1:20

I know that I already wrote meta involvement. But I feel this is a specific issue that merits a separate answer.

Tag management [and maintenance] involvement.

I think that more users should be involved in the tag management issues. And not just the big ones (namely, the dissolution of the homework tag and other similar occasions). But the smaller ones. There is a very few people who bother to check out and weigh in on the majority of minor tagging issues. And this is a problem.

The top three [ordered by activity] answers on the tag management thread are all issues raised by me. And after several days, they still haven't received any substantial feedback (if any at all).

Tags are particularly important, because tags are more or less the only tool we have to consistently sift through the sand and highlight/ignore topics of interest/disinterest. If someone chose to tag a topology question with and nothing else, then all those people who follow are less likely to see the question.

So the fact people are less involved with tag management actively contributes to questions not being seen by the right people, or getting a proper answer. It also means that we are creating a repository of knowledge, rather than a database of knowledge. Databases have some sort of internal order structure, and bad tagging is the destruction of this structure.

And in addition, it really bothers me that a lot of established users interact with a question without retagging it. Sure, it happens to most people that they are not sure how the tags should be modified, but sometimes it's quite obvious and people still ignore it (e.g. a user who answers set theory related questions, but rarely retags a question about intersection from to ).

Retagging, and in general tags management is an important part of exposure, searching and classification of questions (and answers). And I wish more people would take it more seriously.

• I wish there was a better system for creating tags. Right now, anyone over 1k rep can create one, and they don't deprive themselves of it. I used to look at the new tags page often, but it was very depressing and felt like fighting an uphill battle -- people just create random, very often useless tags, sometimes with spelling errors, and it doesn't get fixed for days. I wish more people got involved. – Najib Idrissi Jul 23 '15 at 7:34
• To quote meta.SE: "Communities count on the 10k users to do these reviews from time to time". I feel it's part of the overall issue that relatively very few users over 10k rep are interested in moderation. – Najib Idrissi Jul 23 '15 at 7:36
• @Najib: this old feature request comes to mind. – Asaf Karagila Jul 23 '15 at 8:10
• @NajibIdrissi while I agree in principle, I think the emphasis placed for tags on 10k users is outdated. Everybody can see the new tags on the page you linked and the 10k tool does not really add anything to this page; however this page did not always exist and earlier it was hard for non-10k user to follow tag-creation. – quid Jul 23 '15 at 9:02
• Good point though – Mister Benjamin Dover Jul 26 '15 at 13:30

I have asked one question, and answered a small handful. To me, the biggest issue is downvoting questions from first-time posters. I try to avoid downvoting, unless the issue is really obvious (i.e. copied homework). MathJax takes some time to learn, and I'm still not all that good (though I try to fix posts when I see them). That's why if the OP doesn't use MathJax, I don't care, I just edit the post.

If I downvote, I always try to leave a comment as to why I think it deserves it. If the OP is trying to understand the question, based upon clarifications in the comments, I am willing to go ahead and give an answer. I think we should ask questions of new posters, and if they make it clear they aren't trying to understand, then downvote the question.

Sometimes, it's frustrating when I need to edit one character (two dollar signs to turn it to math formatting), and because I have <2000 rep, I can't do that. :( I think that limitation should be taken off, although it should be a theoretical/historical issue for me in the future.

• Lol. Less than 15 minutes after I made this post, somebody +1'd and accepted my answer. So I made it over the hill!:-) Lol, forgot the \$in my post. Thanks Joel! – FundThmCalculus Jul 7 '15 at 1:43 • At least part of this answer is false now. – Asaf Karagila Jul 7 '15 at 9:25 • "when I need to edit one character" in general I would refrain from making tiny edits like this... There's always room for more improvement, take that opportunity to make more substantial edits (tags, title...). – Najib Idrissi Jul 7 '15 at 9:36 • @JoelReyesNoche my comment was meant in a playful way. But the serious nucleus is that this number just should not be typeset in math mode as far as I know: "Yesterday 23 students came to my office hour to ask if$23\$ is a prime number." (The first is not math, the second is and the typography should highlight this.) – quid Jul 7 '15 at 14:03
• As it's my post, I appreciate the feedback. :) Don't worry, I'm not offended by people editing my posts, even for something as small as that, because the goal here is to improve the post for everyone. In reality, who gets the "credit" is less important than that the post is as good as it can be. – FundThmCalculus Jul 7 '15 at 14:25
• @quid, I agree. FundThmCalculus, sorry for the mess. – Joel Reyes Noche Jul 7 '15 at 22:19
• No problem. :) I like a little humor and tongue-in-cheek, helps keep the atmosphere a little lighter. – FundThmCalculus Jul 8 '15 at 2:26
• @quid: That’s entirely a matter of personal preference. I prefer to set almost all numbers in math mode. – Brian M. Scott Jul 10 '15 at 1:17
• @BrianM.Scott well, I suppose, for some following most any convention is a matter of personal preference. However, mainly, I would like to point you to the context of all this: the first comment by OP suggests to me that upon the edit, they thought they made an error by not putting it into math-mode. I wanted to clear this potential misconception. (In fact you seem to agree that there is no obligation to set this in math mode.) That I do not consider it a priority that this is not set in math-mode can be inferred from the fact that I did not roll-back the edit. – quid Jul 10 '15 at 9:14
• In case anyone cares, here is a reference from a site close by Should running-text numbers be typeset with math mode?. – quid Jul 10 '15 at 9:38

The biggest problem facing this site is: the quick downvoting, deletion, and general hostility towards newcomers and beginners.

To maintain a healthy online community, there must be a constant flux of newcomers who join the site and stick with it long enough to progress and replace advanced users who leave. The current trend is to create higher and higher barriers towards new user participation, which is stopping this cycle right at the start.

I see math.stackexchange in a similar situation as wikipedia circa 2007-2008, when the deletionism and beureaucracy craze there was just beginning to take off. Now wikipedia is really struggling to gain and retain users, it's a big problem for them. That's where math.stackexchange is heading.

• I, for one, rather have no new users than a steady influx of low-quality help vampires. It is not valid to put the burden only on the cmmmunity; the average newcomer has also changed over the years, resulting in a decreased percentage of retained quality contributors. – Lord_Farin Jul 10 '15 at 7:11
• I don't see any hostility towards newcomers and beginners. The hostility, if any, are towards low quality posts. – user99914 Jul 10 '15 at 7:26
• Since this is not a kindergarten, I don't see the need to fully educate people. If someone is unable to take cue from real life, and at least take a cursory examination of a community before jumping in, that's their problem. If someone is not willing to look around and see what are the community norms, it's their problem. Sure, learning MathJax and all is nice and important, but that's the icing on the cake. I want people who post new questions to at least put some basic effort into understanding what a reasonable question should look like. If they can't do that, well... – Asaf Karagila Jul 10 '15 at 13:22
• "I see math.stackexchange in a similar situation as wikipedia circa 2007-2008, when the deletionism and beureaucracy craze there was just beginning to take off. Now wikipedia is really struggling to gain and retain users, it's a big problem for them." These are two statements about wikipedia: one, that some "deletionism and b(e)ureaucracy craze" began to be prevalent there circa 2007-2008 (hence, presumably, that it was absent or significantly lower before that date and that it has remained prevalent since?); two, that a big problem for wikipedia nowadays is to "gain and retain" .../... – Did Jul 12 '15 at 9:50
• .../... contributors. Are these two statements considered folklore and/or documented and/or quantified somewhere? I am asking this because, from the fragmentary knowledge I may have of WP, a move did occur around the time you mention, but from a status of global non reliability of WP pages (including some annoying clear-cut cases of lunacy) to a much greater reliability, to the point that, now, one can actually learn some subjects from WP, and/or refer students to it. – Did Jul 12 '15 at 9:53
• This is getting a little off topic from math.se, but the shift is well established in the history of wikipedia. The main principles that have shifted are primarily related to WP:IAR, WP:BURO, WP:N, and a few others. For the rise of bureaucracy and rule heavy culture and it's effect on new user participation, see the Slate article "The Unbearable Bureaucracy of Wikipedia". For the changing balance of deletionism and inclusionism you can look at the wikipedia article on the subject for a history. Verifiability (WP:V) has always been a core principle and it is mostly orthogonal to these issues – Nick Alger Jul 12 '15 at 10:06
• Another excellent article on the subject is the MIT Technology Review's featured article, "The Decline of Wikipedia" – Nick Alger Jul 20 '15 at 0:22
• @NickAlger Thanks for the Slate and the MIT Technology Review papers on the subject. (But if ever you add other references, please use @.) – Did Jul 20 '15 at 7:15
• @Did You might also be interested to see this recent academic article on the wikipedia new user retention problem, the 2007-2008 shift in participation, and it's relationship to the introduction of a rule-heavy culture. – Nick Alger Jan 1 '16 at 7:11

I assume MSE is meant to be a meritocracy, but it is not. The point/reputation/privilege system might well be designed to reward enthusiastic new members for their efforts in answering math questions, and to encourage them to further improve their skills. But in reality the points for a "nice answer" hardly ever reach them. For newcomers it is very difficult to get even a single up-vote, even though the answer might well be excellent.

By contrast, there is an elite at MSE who constantly praise and upvote each other (even the most trivial comments) and who are able to protect their practices due to the near-moderator privileges of themselves and their friends.

Many newcomers will be discouraged by this system and leave MSE for good.

• Instead of downvoting your borderline-rant answer, let me point you to this comment. Also, we see in the new users tab that indeed many users manage to accumulate reputation. It can be a bit hard to get started, but with a bit of perseverance you can surely get there. To those who don't want to spend that effort, I say: Your loss. – Lord_Farin Jul 9 '15 at 16:20
• "there is an elite at MSE who constantly praise and upvote each other (even the most trivial comments) and who are able to protect their practices due to the near-moderator privileges of themselves and their friends." Names! We want names! – Did Jul 9 '15 at 16:57
• I have recently gained some reputation points, but this has not changed my opinion about the site at all. – M. Wind Jul 9 '15 at 17:02
• A meritocracy doesn't award people for their efforts and encourage enthusiasts. It awards quality of effort (not quantity). – Matt Samuel Jul 10 '15 at 0:56
• @Matt: Unfortunately, there are people whom I'd rather have no vote to close or reopen, but due to sheer number of posts have reached beyond the 10k barrier for "literally everything" to be accessible. So quantity does play a role, if one is sufficiently motivated to gain points. – Asaf Karagila Jul 10 '15 at 13:23
• People who get lots of upvotes are far more likely simply be people who tend to write things others like to upvote rather than part of a conspiracy. – user14972 Jul 13 '15 at 11:53
• @Did were there not multiple users who were banned for this very issue? (Note past tense.) It seems to have been significantly alleviated in the last year or so. A couple years ago, I noticed that new users who had posted very good answers before high rep users often were not upvoted. It seemed like high rep users were voting each other up in droves. I get the reasoning: there is an implicit trust in an answer from a high rep user. I think this has been reversed quite a bit and I've seen new users get the credit they deserve. I think the site has definitely become more inclusive recently. – Cameron Williams Jul 15 '15 at 4:21
• This is just not consistent with the good experience I have had on this site. – ncmathsadist Jul 20 '15 at 19:11
• This answer doesn't agree with my experience on this site. – Nate River Jan 13 '16 at 11:24