20
$\begingroup$

I'm prompted to start this thread by a comment on a recent question about why this site has the homework policy that it does. As I said in that answer, the ecological niche that this site occupies is rather different to the one that Mathematics Stack Exchange does, and a lot of this difference is directly attributable to the existence of MathOverflow.

However, I realize that the details of why this situation came about are unknown to many users who joined after much of the development happened*, so it's worthwhile to go over some of that history. Plus, today marks a bit of a grim anniversary in that history, so maybe it's a good time to do some reflecting.

So:

How come there's only one Stack Exchange physics Q&A site, while mathematics gets two?

I've written an answer with the essential timeline of the history, along with some links to relevant landmarks, and I'll describe from a (very) high level some of the choices different communities made along the way. However, I think it would be nice if people actually involved in those decisions could add their perspective on how we came to where we are at the moment.


* Using some kludgy informal queries, I make it that from the users that posted ≥0-score posts in the last year, half joined in the last 18 months, and more than 75% within the past 3 years, which is rather later than much of the relevant history.

$\endgroup$
  • $\begingroup$ As an aside, I've noticed that the noise level at physics.SE has risen immensely as of late. It's almost as bad as is the noise level at stackoverflow.com. Perhaps almost all of the key questions that don't require a book as an answer have been asked. $\endgroup$ – David Hammen May 1 '17 at 20:13
  • 2
    $\begingroup$ @DavidHammen, I don't think that will ever be the case. Physics is ever growing, and there are new fields being explored. When we thought at the end of the 19th century that physics was all fixed up and perfect, quantum mechanics came along, and then relativity, and now we're trying to make it so the two fit together, and then we've got specific areas like quantum computing, and much of cosmology is questions not answers, and...I could go on. $\endgroup$ – heather May 6 '17 at 22:07
21
$\begingroup$

I'll take this in roughly historical order, and to the extent that I'm aware of things and can remember them, though there is a definite danger that I'm telling the stories as we'd like to reimagine them rather than as they actually happened, so please go check the original sources and make up your mind from them. This obviously can only be a pretty partial list, so if you have additions or corrections please add them in (just try to keep it neutral).



  • The first science site using the Stack Exchange engine was MathOverflow, which started on 28 September 2009 as Wikipedia tells it; this Meta MathOverflow question has more details. MathOverflow was started by a set of postdocs and graduate students, who set out to create a meeting place for professional research mathematicians. Some interesting reads about it are the official announcement on the blogosphere, or this slightly later piece on the AMS notices. From about one year into its tenure, there's pieces on The Atlantic and Mercury News that capture a lot of the zeitgeist, I think.

    On the more technical side, MathOverflow was started as a completely separate entity to Stack Exchange; instead, MathOverflow licensed the software from SE for its own use, a practice which SE did for a short time after it started (other instances including moms4mom and chiphacker, which became Electronics SE), and which stopped after they set up area 51 as a mechanism for creating non-SOFU sites within SE; see this SE blog post for more details. MathOverflow migrated to the Stack Exchange 2.0 platform, managed by SE itself, in mid-2013, without really changing much in how the site is run from what I can tell.

    Some interesting bits and pieces regarding MathOverflow are this, this and this meta questions. Also, note that the pre-migration Meta MathOverflow, which ran on an old-school PHP bulletin board, is now preserved as tea.mathoverflow.net, and if one wants to delve into deep MathOverflow history that is the place for it.

  • An important bit of context from the time around the creation of MathOverflow is the first Polymath Project, which started in January 2009 taking off from Tim Gowers' blog, and which closed successfully within a couple of months. I don't know to what extent the MathOverflow founders had been thinking of starting the site before this, but it certainly showed that there was a large community of mathematicians online ready to take a more hands-on approach than the (also rather active) existing blogging community of the time.


  • Mathematics Stack Exchange started a good deal later than MathOverflow, on 27 July 2010, and it came up through the area 51 mechanism. A good place to have a look at how the conversation looked like at the time it was founded is the proposal's Area 51 page, which contains example questions and some discussion - and particularly this 2011 question on the one-site model vs the two-site model. Have a look at the definition tab for those goodies (I find Robert Harvey's comment here to be particularly illuminating on how things developed), and at the beta tab for more stats. Tooltips have more precise dates.

    There is also a lot of relevant discussion on a thread on the pre-2010 Meta Stack Exchange (a meta site for the Stack Exchange 1.0 software and family of sites), which is preserved in the Wayback Machine.




  • I also want to mention another pair of sites that follow the two-site model that have been successful within the Stack Exchange network, and which we probably don't explore as analogues nearly as much as we should - Computer Science and Theoretical Computer Science, the Area 51 proposals for which are here and here respectively.

    Keep an eye on the timings, though: TCS was proposed in June 2010, went to public beta in August 2010, and it graduated from beta in November 2010; the plain CS site was proposed in September 2011, started public beta in March 2012, and it graduated four years later in January 2016. Thus, this is arguably another case where the 'hard' site came before the 'soft' one (bad descriptors, but you know what I mean), though obviously the elephant in that room is Stack Overflow, which isn't computer science as such but has a strong bearing on that discussion.

    Physics and computer science are very different in many respects, but this is nevertheless an interesting model to study.


  • Over the years there have been definite chafings at the perceived low level of questions on this site, as shown e.g. in this thread - there's probably more out there to be found, though.

  • There is also PhysicsOverflow, and I think I will leave the telling of this story to its organizers. I will point to the comments section of Luboš Motl's blog as the place for the initial discussions (more specifically on this thread and its comments), and its initial announcement on the MathOverflow meta, as well as the thread What is Physics Overflow and how is it linked to Physics.SE? on this meta. I do feel it's important not to underestimate how much of the motivation for the foundation of PhysicsOverflow was purely on moderation concerns rather than shooting for a research-oriented site, but I'll let them tell their story.

    PhysicsOverflow is currently active, though as far as I understand it they still import a fair amount of this site's content through our CC BY license. I really encourage readers to visit and form their own opinion.

    I do want to touch on traffic as regards PhysicsOverflow, though, because I've seen several people make claims along the lines of "it's OK if we litter Physics Stack Exchange with drudge-level homework because if researchers get offended they can go to PhysicsOverflow much like mathematicians can go to MathOverflow". PhysicsOverflow reports its traffic on its statistics page; in the past year those show an average of about 25 questions per month posted directly to PhysicsOverflow (and about 10 questions per month imported from Physics Stack Exchange), which is about the rate that Theoretical Physics had before it shut down. For comparison, MathOverflow gets about 40 questions a day, to 650+ on Mathematics and ~100 on this site.

$\endgroup$
  • 2
    $\begingroup$ Correction: prior to the migration, meta at MathOverflow was called meta.mathoverflow.net (being a PHP bulletin board, not sharing accounts with the site). The renaming to "tea" was made necessary by the migration, since a standard per-site meta was a part of SE software by then. relevant discussions $\endgroup$ – user60480 Apr 25 '17 at 21:36
  • 1
    $\begingroup$ Good research! I wish it could be extended by the remembers of our old site members, not using the site any more, on different reasons. $\endgroup$ – peterh Apr 25 '17 at 22:31
  • 1
    $\begingroup$ +1. This is useful history, saving recent users the trouble of trawling Meta sites to find out the facts. However, it does not (I think) explain why this site has the homework policy which it does, nor why you do not appear to view Physics Overflow as a suitable outlet for PSE users to ask and answer post-grad level questions. If PO has little traffic, this is the collective responsibility of its active and potential users. Your concern seems to be that PSE is being swamped by low-level questions, but the homework policy applies to all levels. $\endgroup$ – sammy gerbil Apr 27 '17 at 16:00
  • $\begingroup$ @sammygerbil I think PO is not a suitable outlet because it has a number of deep flaws independent of any of this; PO's growth is the responsibility of the people who want it to grow but I don't see why you're tasking its potential users with that. I'm not concerned that PSE is being 'swamped' - I think the current situation is just fine, but if people want things to change then the onus is on them to show that it can be done sustainably and addressing all the relevant concerns. $\endgroup$ – Emilio Pisanty Apr 27 '17 at 16:13
  • $\begingroup$ This specific thread is a response to comments like this one that draw a false equivalence between the development histories of MSE and PSE: if you actually look at what happened, there is pretty much absolutely no part where those histories are alike. You specifically asked for the differences in the environment - here they are in all their glory. $\endgroup$ – Emilio Pisanty Apr 27 '17 at 16:18
  • 1
    $\begingroup$ @EmilioPisanty The comment to which you refer does not mention development history. I was referring to the current situation, current environment. I am not aware of the deep flaws of PO which you allude to, nor the moderation concerns which started PO. These seem more like the environmental differences to which you refer. But the only explicit criticism you seem to make is low traffic. $\endgroup$ – sammy gerbil Apr 27 '17 at 16:48
  • 4
    $\begingroup$ @sammygerbil I have plenty of criticisms to make of PO, but this is not the place for it (and much of it is already public and linked to anyway, as is the moderation differences between PSE and PO - and I should say that arguing about PO without understanding those differences isn't particularly constructive, given their historical importance). I find personally that a five-minute skimming of PO makes it blindingly obvious that it is not at present equivalent to MO in any way, and claiming that it is is deeply disingenuous. If you find PO useful, good for you - just be aware that others don't. $\endgroup$ – Emilio Pisanty Apr 27 '17 at 18:02
  • 1
    $\begingroup$ @EmilioPisanty I do not use PO, I doubt I could answer any questions there. Many of its (few) active users are (or were) active on PSE also. I am not claiming PO is equivalent to MO, only that it provides an outlet for asking and answering research-level physics questions without the intrusion of the drudgery of homework questions which you appear to be complaining about in your answer to Why are homework-related questions accepted on maths SE more than here? $\endgroup$ – sammy gerbil Apr 27 '17 at 21:43
  • 4
    $\begingroup$ @sammygerbil I guess we're talking past each other, so please clarify your point. Yes, it provides a possible outlet for (a fraction of) research-level physics - but so what? Are you here to argue about its quality as such an outlet? (If so: this answer does not explain why I personally don't view it as a suitable outlet because this answer is not the place. Similarly, this answer doesn't explain this site's HW policy because it does not set out to do so.) Have you got something concrete to propose? Or do you just want to repeatedly bang the drum marked "PO exists", and nothing else? $\endgroup$ – Emilio Pisanty Apr 27 '17 at 22:38
  • 1
    $\begingroup$ @EmilioPisanty No I am not promoting PO or defending its quality, only pointing out its existence as an outlet and an opportunity. The purpose of my comment thread is to point out that your question and answer do not (for me, probably also other newcomers) explain why the history and ecological niche of MO/MSE versus PO/PSE have resulted in a different attitude towards homework-like questions. Interesting as the history is, I do not find that it addresses this issue. $\endgroup$ – sammy gerbil Apr 27 '17 at 22:59
  • 3
    $\begingroup$ @sammygerbil "If PO has little traffic, this is the collective responsibility of its active and potential users." No. It is not the responsibility of the 'potential' users. I have not the slightest responsibility for that site. It is wholly the problem of the people who run it. There is no in onus on the users of Physics Stack Exchange to gin up activity for them and no reason for us to consider the existence of a non-affiliated site in setting policy for Physics Stack Exchange. We don't base our rules on the existence of Quora or Physics Forums, either. $\endgroup$ – dmckee Apr 28 '17 at 3:32
  • $\begingroup$ @sammygerbil It doesn't because that wasn't its primary purpose. Nevertheless, there is strong support in the primary sources already linked to (most strongly the Area 51 proposals at definition stage) for the thesis that MSE was started to house the homework-like and other sub-research-level questions that MO did not accept. From a historical perspective, the outlier as regards homework is MSE, not this site. However, if you want more from this answer than its primary purpose but you're not willing to look at the primary sources, there's no point in further discussion. $\endgroup$ – Emilio Pisanty Apr 28 '17 at 13:03
  • $\begingroup$ A quick question - you state that "there was a discrete policy change announced on Graduation, site closure, and a clearer outlook on the health of SE sites that makes it much less likely that a small and slowly-but-consistently growing SE beta site will be shuttered for lack of traffic" - does this mean that if Theoretical Physics or a "Physics Overflow" or what-have-you was proposed on Area 51 now, it might survive? $\endgroup$ – heather May 6 '17 at 22:38
  • $\begingroup$ @heather If it had the same momentum that the original TP site did, then I tend to think that yes, it would survive, at least through a much more extended beta period than the original did. However, I would also tend to doubt that you could get that much participation a new time 'round, partly because I think the level of PSE has gone up since 2011 (at the time there were complaints that PSE was 'only high-school physics', which no longer ring true - though of course there's other problems), and partly because it'd be hard to gather people willing to give it a go. $\endgroup$ – Emilio Pisanty May 6 '17 at 22:52
8
$\begingroup$

The ''grim anniversary'' in the history of Physics SE that Emilio alluded to is the shutting down of TheoreticalPhysics SE in April 2012 by the Community Manager of Stack Exchange. Between this announcement and the actual closing act, there was quite some discussion on Physics SE and on the blogs (some of which are no longer accessible; the links are dead) of some participants. In one of these one can read:

The reason that TP.SE was formed was largely due to wanting somewhere for research level stuff without all of the noise.

As a member of TheoreticalPhysics SE wrote:

We have succeeded in maintaining high quality site which has become a reliable resource of correct and useful information, which is sometimes not that easy to obtain. We had top people coming to the site and contributing high quality content. Quality, rather than quantity, is of course something that is not easily measurable, but I think it has value nonetheless. Traffic was an issue, but for a highly technical site the pattern of traffic, including how it changes with time, should be evaluated differently.

But although the officials had said in a more general context:

As long as your site shows steady progress and continues to make the Internet a better place to get expert answers to your questions, it will march on.

and this is quoted in the shut down notice, ''making the Internet a better place'' is interpreted from the point of view of the SE marketing perspective only: Commercial viability requires high enough traffic rates and these were not maintained by TheoreticalPhysics SE during its beta existence.

Thus the reasons for shutting down the site have more to do with the standards of SE for a site to be tolerable in a commercial enterprise than with the quality of the contributions to the site. In the thread ''Why did Theoretical Physics fail?'' the SE officials didn't question the fact mentioned in the thread opening that

the quality of questions (and answers) was actually good.

But SE required growth and high traffic. (Typical argument: ''Physics [SE] was already a bigger and more established site when it launched.'') In the words of one of the active members of Theoretical Physics SE,

SE closed a Jazz club for not being the same as a Madonna concert.

The actual closing message on TheoreticalPhysics SE meta mentioned that

the questions and answers posted here will be made available for download and re-use by anyone who wants them.

There was an immediate desire to migrate the site outside the SE framework. See ''Where do we go from here?'', but for practical reasons it took nearly two years to restart the site as PhysicsOverflow. All questions and answers from TheoreticalPhysics SE and its meta site were migrated to PhysicsOverflow, and the site continues under improved rules. In particular, PhysicsOverflow treats comments on the same level as questions and answers (any length, anytime editing, fully disclosed history, bookmarking). PhysicsOverflow also imports upon request research level posts from Physics SE and MathOverflow; new answers of questions from there are usually reported back on the original site.

PhysicsOverflow hasn't become a Madonna concert with all its noise but it preserves the level, quality, and noise-free enjoyment the Jazz club once called TheoreticalPhysics SE offered to the physics research community. Reflecting the high level, the number of questions, answers, and comments per month by registered users is highly fluctuating, with long term averages around 25, 25, and 75, respectively. These numbers are comparable to those of Theoretical Physics SE at the time of its shut down and do not include the posts of anonymous users or unregistered users imported from elsewhere, which add about 30% to the other numbers.

In view of Emilio Pisanty's comment

There is also PhysicsOverflow, and I think I will leave the telling of this story to its organizers.

let me add that a timeline of PhysicsOverflow can be found here. Due to unrelated events, the relations with Physics SE have been characterized by some tensions, as reflected by the very unusual ratio of upvotes and downvotes in the promotion ad (visible to high rep users only). Being quite active on both sites I'd be happy if these tensions would give way to a friendly relationship of the same sort as it exists between Mathematics SE and MathOverflow.

$\endgroup$
  • $\begingroup$ Thanks for this perspective. I do find it important to remark that the SE focus on traffic should be seen with a past-tense lens, though - there was a public policy shift after TP was closed, indicating that this would not happen at present, and there's plenty of signs (such as the new astro site's continued beta) that it is the case. $\endgroup$ – Emilio Pisanty Apr 30 '17 at 16:35
  • $\begingroup$ That said, I would like your opinion on how you would compare PO with TP in terms of active user base (both size and how many TP users are now active on PO) and topical scope (actual active topics as opposed to in-principle on topic). I found TP at the time to be very limited in its scope compared to the breath of theory topics in physics - would you say that the scope has broadened significantly? $\endgroup$ – Emilio Pisanty Apr 30 '17 at 16:42
  • $\begingroup$ @EmilioPisanty I got the impression from my observations, that the SE company did not really change their attitude towards smaller more specialized communities. While it is true that all sites that have successfully reached the status of public beta can happily live on, new communities that look potentially too small for SE standards are now shut down earlier already in private beta as it happend to OpenScience or even before that as it occured to the Neuroscience Area51 Proposal. $\endgroup$ – Dilaton Apr 30 '17 at 16:56
  • $\begingroup$ I would also like your comments on the decision to host the site outside of SE. As a contrast, the old astro site was closed at the same time as TP and for the same reasons, and there is now a stable beta in its place. Do you personally share the views on SE 'censoring' expressed in the PO history thread you linked to? $\endgroup$ – Emilio Pisanty Apr 30 '17 at 17:00
  • $\begingroup$ @EmilioPisanty: Because of the 2 years break, we managed only to get back a fraction of the former activists on TP to contribute regularly on PO. Looking at the total user list physicsoverflow.org/users (which includes all TP users and all imported Physics SE users), we currently have 32 users with >1000 rep, of which (judging from memory; it is difficult to get detailed statistics) 17 displayed significant activities since PO revived TP. 4 of these 17 were not active before 2014. This leaves a fraction of 13/28, slightly less than a half of the frequent users of TP continued on PO. $\endgroup$ – Arnold Neumaier Apr 30 '17 at 17:05
  • $\begingroup$ As to the topics represented on PO in the last two years, I guess that roughly half is quantum field theory and a quarter string theory. If I remember correctly this was not too different from TP; I don't remember how the remaining quarter was distributed on TP. Thus now as then, the distribution of topics is very skewed, natural for a very broad field in which only a small number of experts contribute answers. These experts determine how people perceive the site and which kind of questions are being asked. We have very few questions from experimental physics or from solid state physics, say. $\endgroup$ – Arnold Neumaier Apr 30 '17 at 17:11
  • 1
    $\begingroup$ As to censoring, I am not a friend of harsh judgment on either side. I'd like to let the past rest in peace. Accusations don't help and people not satisfied in some place sooner or later find another one suiting them better. I had tried (in vain) to make suggestions to alleviate the rough tides around the Physics SE elections in fall 2012. $\endgroup$ – Arnold Neumaier Apr 30 '17 at 17:23
  • $\begingroup$ Note also that I was not involved in the creation of PO; I became active only shortly after the public beta started and the originators tried to get back the old TP crew. That PO was hosted outside SE was at that time already a fact. I joined the moderators of PO only a year later. $\endgroup$ – Arnold Neumaier Apr 30 '17 at 17:26
  • $\begingroup$ @ArnoldNeumaier Thank you for all those responses. $\endgroup$ – Emilio Pisanty Apr 30 '17 at 18:42
  • $\begingroup$ @EmilioPisanty even though the current actual active topics on PO may look rather limited, any questions about all experimental and theoretical physics topics, including astronomy and for physicists relevant mathematics, are wellcome as long as they are graduate-level and above. $\endgroup$ – Dilaton May 1 '17 at 11:51
  • $\begingroup$ BTW I very much second Arnold's wishes to leave past tensions between Physics SE and PhysicsOverflow rest in peace now and would be more than happy to see a respectful and professional relationship between the two communities getting established as it exists between Mathematics SE and MathOverflow in the long run. $\endgroup$ – Dilaton May 1 '17 at 11:56
  • 1
    $\begingroup$ @Dilaton I'm aware of the stated scope for PO. The question was about the topics that are actually active, have a sufficient number of of users, and therefore where tough questions have a good chance of getting answered (over, say, PSE). TP had some severe limitations in that regard and I hadn't seen a PO push to expand past them, hence the question. From Arnold's response, it looks like PO is roughly where TP was at its closure - which was unfortunately of rather limited use outside QFT. It's nice that you welcome other topics but what matters is whether questions get answered. $\endgroup$ – Emilio Pisanty May 1 '17 at 12:53
  • $\begingroup$ @EmilioPisanty yes, from time to time people ask about our scope as for example here, because just from looking at the front page the most active topics are similar to what TP.SE had indeed. There has not (yet) been a well organized larger campaign to broaden the range of active topics, though from time to time we try to seed activity in not so active tags by importing appropriate questions say about quantum optics, experimental physics, and others. But this did not yet pan out obviously ... $\endgroup$ – Dilaton May 1 '17 at 13:40

You must log in to answer this question.

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