There has been recent turmoil on StackExchange (relating to various issues) and there are thoughts rising that it might be better for the community to be more independent from the commercial side of StackExchange.

We may repeat that meta-discussion over here. We should/can ask ourselves what specific measures we can take in the statistics corner?

  • What can we do to become more independent from the community/people's source code of the old platform and become an open-source source model that is more decentralized and more controlled by the community, the community that is also creating the content?
  • Are there steps we can take that are specific for CrossValidated?

Any suggestions?

For example, there is a wide spectrum of possibilities:

  • An extreme case would be to create a spin-off as has been done before (see PhysicsOverflow)

  • As a softer version of a spin-off we could possibly disentangle our contributions from Cross Validated by posting the majority of our contributions elsewhere. This will create more diversity in the ownership of platforms by which the content is spread to the public/internauts.

    Posting shorter answers here and placing more in-depth content elsewhere may work because it is the length of posts that relates to the quality of the content, at least the score of the posts does relate to the length of the post

    post length

  • Ideally we should make a smaller step than fully separating (as proposed on meta by Scortchi). That is because disconnecting from StackExchange removes many advantages (like connections with other SE sites).

The goal of this post is that, based on the suggestions, we might start to develop a consensus solution that provides options for breaking the corporate monopoly, and unites the community, and prevent chaos and prevent separation, prevent splits in the community and prevent that we will start to move in different ways.

  • 1
    $\begingroup$ I am currently thinking about something like a system/platform that integrates blogrings related to Q&A by using some shared standard and being a distributed system rather than being centralized. A sort of inverse process how Usenet (distributed) turned into Google Groups (centralized). $\endgroup$ Oct 24 '19 at 14:28
  • 1
    $\begingroup$ (+1) Certainly worth discussing. I've lightly edited one sentence that seemed ambiguous to me; please check to see that I've captured your meaning. $\endgroup$
    – mkt
    Oct 24 '19 at 18:39
  • 11
    $\begingroup$ Recent events certainly have cast doubt on the future of SE as a reliable platform. However, the reason I contribute here, and not on reddit, Quora, Wikipedia, or whatever, is because of its (current) style. To win people over for a migration, there would have to be a lot of similarities between the new platform and CV. I would certainly encourage a more independent platform (+1), but the alternative must probably be compelling not only because it is more independent if we are to reach consensus. $\endgroup$ Oct 25 '19 at 5:05
  • 10
    $\begingroup$ I am with @Frans Rodenburg here. I can imagine someone setting up a blog and saying Let’s run this like CV using the style we like. But, but, but: something like CV requires a secure server, more software than blogs run on, and a set of people doing a lot of work. Do I misunderstand? I would like it to work but I can’t imagine it working. Despite despair at the SE company I (we) have much sunk capital here and I am highly reluctant to part company and start something new. Yes, I know that SE are betting on that attitude. $\endgroup$
    – Nick Cox
    Oct 25 '19 at 7:53
  • 1
    $\begingroup$ @NickCox I agree, it will be difficult. And indeed the PhysicsOverflow does not look as nice, and a hard STEXIT is undesirable for all parties. But possibly we can at least shift it a little. Several people here already have blogs. We could highlight them better and make more possibilities available for others. Create some standards such that the blogs are more alike and link better to Q&A here. Who knows in five/ten years from now there might be some open standard and an independent platform is more nearby. Then we are at least more prepared for it or made the first steps towards it. $\endgroup$ Oct 25 '19 at 8:04
  • 1
    $\begingroup$ Maybe I am too much desiring the old-Internet back? $\endgroup$ Oct 25 '19 at 8:06
  • 3
    $\begingroup$ There are many reasons why people fall in here and stay, but seemingly they overlap. I was aware of SE for some years before I joined; my main hesitation was that it looked so extraordinarily complicated. But I first fell into SO because some questions I could answer were not being answered and then later into CV, where there was enough low-hanging fruit to engage. Wikipedia goes all the way from astoundingly good to utterly awful, but it's not for me to participate. My limited sample of Quora and Reddit is a mix of dismay and disappointment, and very little else. $\endgroup$
    – Nick Cox
    Oct 25 '19 at 16:27
  • 5
    $\begingroup$ (ctd) But the mix here on CV of willingness to help and high standards for analysis, advice and exposition is deeply congenial. There are dozens of people whose identifiers I recognise and who I regard as virtual friends. I am unwilling to accept the distinct possibility that SE is doomed if community goodwill has been lost. $\endgroup$
    – Nick Cox
    Oct 25 '19 at 16:29
  • 1
    $\begingroup$ CV is a great statistics resource. There must be a way to fork it to some other place. Answering people own copyright, we can port the content elsewhere. I think it is inevitable that tech folks will migrate elsewhere. There is no reason to contribute under the new CoC. If they felt they can annoy 99% to please 0.01% then only change in management can fix it and it’s not happening. We need to separate tech content from identity politics $\endgroup$
    – Aksakal
    Oct 26 '19 at 2:52
  • 3
    $\begingroup$ @Aksakal I fear the "must" here is wishful thinking. CV is thousands of threads, the software that binds them, the hardware that hosts them, and even if what you say about copyright is true there is more than that to the legal basis. Even if SE collapses their dying phase may not include acts of charity letting communities float free. Beyond the present crisis what a community also needs is a framework for dealing with trolls, spammers and various kinds of obnoxious visitors, some but by no means all transient in their postings. That needs to carry over too. $\endgroup$
    – Nick Cox
    Oct 26 '19 at 8:54
  • 2
    $\begingroup$ @NickCox I think cv was dealing with all the issues you mentioned very well, it’s due to good work of moderators. New CoC is addressing the problem that didn’t exist here, and created a set of new problems. It’s a more general problem in online communities where liberal arts folks drag tech people into their mud wrestling matches. There had to be a way to abstain from their wars $\endgroup$
    – Aksakal
    Oct 26 '19 at 13:57
  • 1
    $\begingroup$ @NickCox, I also agree that forking would not be an easy ans straightforward operation $\endgroup$
    – Aksakal
    Oct 26 '19 at 20:02
  • 1
    $\begingroup$ I would very interested in a "statisticsoverflow" website if questions could be migrated. Otherwise as NickCox has said, we are throwing away hundreds (if not thousands) hours of work. @SextusEmpiricus: If you can afford the time, I would contact biostars' to learn more about their operating model. A working website/online community requires more than content creation. At the very least it requires hosting and it is totally unclearly who is going to foot the bill for millions of daily visits on a website. $\endgroup$
    – usεr11852
    Nov 10 '19 at 2:36
  • 1
    $\begingroup$ @usεr11852saysReinstateMonic I would say that hosting would be the least of problems. StackOverflow is big, but CrossValidated is already small. It is not unlike other free (ie. University or otherwise sponsored) websites. For me the question is still whether the community wants to relocate, and if they want to, then how do they want to. So for me personally, I am still standing on the lowest step, the blog-model. $\endgroup$ Nov 10 '19 at 8:00
  • $\begingroup$ Would you know a university willing to do this? I am not affiliated with one nowadays and I would be skeptical of a company. (I am not a fan of the blog idea, I cannot see how it would work.. Plus I find the majority of blogs somewhat preachy.) $\endgroup$
    – usεr11852
    Nov 10 '19 at 9:18

Solution: exact copy elsewhere

It is possible to have a virtually identical format, widely and productively used, under control of a non-SE entity that is invested in high-quality treatment of the subject area. I'm thinking of this site's bioinformatics equivalent, BioStars.



Recent efforts along this path have been made by several other contributors from StackExchange/StackOverflow and discussiona about it are happening on https://forum.codidact.org/

  • $\begingroup$ It's common practice in science to identify effects by changing one parameter at a time while holding others constant. This would allow us to change ownership without changing format. $\endgroup$ Oct 30 '19 at 14:50
  • 2
    $\begingroup$ I find the biostars more closely resembling the SE format than physicsoverflow. It might be the easiest solution to copy everything. 72k questions is a good number and a good example that it can be done. $\endgroup$ Oct 30 '19 at 15:23
  • $\begingroup$ The physicsoverflow FAQ might be the best place to start reading, rather than the homepage: physicsoverflow.org/faq $\endgroup$ Nov 6 '19 at 17:00

Solution: create an independent WIKI portal

The Q&A format of StackExchange has generated a sort of yellow pages for all sort of problems with several advantages:

  • It generates content/information that is relevant to practical cases.
  • The problem-based question and answers introduce concepts in a way that is easy to understand.
  • The generation of the content has been completely free based on volunteering work.

Stack is changing

However, we may wonder how long this system is gonna last (besides the company that is running the system is turning it's back towards the volunteers). The graph below indicates how StackOverflow/StackExchange is changing by looking at the time evolution of the number of votes, where different curves are shown based on the score of the questions after 30 days (from this Query)

time development

To compare the trend more easily, the curves have been scaled such that the value is equal to 100 at the time point 2012 year.

The problem is: The number of monthly questions with a high vote count is decreasing (especially on StackOverflow on CrossValidated it is still somewhat stable). The number of monthly questions with a low vote count is growing.

StackExchange as a comic:


Most of the pretty fruits have already been picked. The sort of activity we see now are monkeys picking the low hanging fruit (I am one of them) and occasionally there might pass by a giraffe that is able to pick some fruits from the top of the tree. There is an increasingly number of fruit that ends up rotten on the floor and we do not have enough ants to clean that up.

Solution: move over the (majority of the) content to a wiki

We could say that the question activity is currently mostly in lower quality questions (an increasing number is not getting answered and receiving low score). We may ask ourselves whether the Q&A format is still useful.

We could argue that the database is already very full and now needs to be organized and maintained. This is a different ballgame than the continuous posting of questions and answers.

To do this. We could create a portal to the current content (and possible future content) summarizing annotating and clarifying the content on CrossValidated. On this portal, we could use the format of wiki pages (since the Q&A phase is over or will be soon). The additional advantage of a wiki is that there is lots of software available that may help to generate the website (only the conversion of the content from StackExchange to Wiki may be a problem but I imagine it should be possible).

And also, we gain independence.

  • $\begingroup$ Stack Exchange uses some flavour of markdown for posts and there are wikis that use markdown as syntax, so the technical part of the conversions shouldn't be too hard. Rewording Q&A content to wiki pages is a different story. $\endgroup$ Oct 30 '19 at 8:38
  • 1
    $\begingroup$ @Josef . Indeed, in terms of keeping the formatting intact the content is easy to copy-paste. One of the problems is changing framework from Q&A to WIKI. For this I am thinking about some format like a FAQ page where the fruits (questions) are brought together in bunches based on similar topic, style, keywords, etc. The 150k fruits on stack-exchange can be brought down to much less bunches. StackExchange has single fruits, the wiki will have bunches of fruit (not ordinary grouping but with added content). $\endgroup$ Oct 30 '19 at 9:39
  • 1
    $\begingroup$ The bigger problem is how we are gonna deal with credits and licensing. Ideally an already existing organization might come to help with these issues, like R-foundation or Wikimediafoundation. Ideally StackOverflow embraces the project and will themselves start up this foundation that can become an independent partner which is run by the community rather than the company. Let StackOverflow deal with questions and answers and related software (it seems like they want to go that way), let the community deal with the IP and collection of (open/free cc by-sa) information that comes out of this. $\endgroup$ Oct 30 '19 at 9:45
  • $\begingroup$ I guess that some sort of version control system which keeps track of all changes in the wiki and also incorporates the changes that had been done before in the copied content from CrossValidated will be a way to give credit to all people that have added their contributions to the wiki articles. (I wonder how this would go when the CrossValidated site ever seizes to exist, how would one credit the previous contributors when they only have a nickname like user234412 and their profile page with possible additional information does not exist anymore). $\endgroup$ Oct 30 '19 at 9:54

Solution: The my blog model

Something like a system/platform that integrates blogrings related to Q&A by using some shared standard and being a distributed system rather than being centralized. A sort of inverse process how Usenet (distributed) turned into Google Groups (centralized).

With the blog model 'we' (the contributors) will have more control over the platform(s). The CrossValidated website will become more like a portal on which we can place short introductions and mix the information and content from different blogs.

  • $\begingroup$ This answer is very vague. It is an idea in development. I have made it community wiki such that others can add to it. $\endgroup$ Oct 24 '19 at 14:38

Solution: Wait.

More time is needed.

Eventually, more and more users will be fed up with declining SE and look for something fresh, new and better.

If the site does not innovate, if the site does not keep in touch with the online community, then eventually it will be overtaken by some other site as the major platform for Q&A exchange of (expert) knowledge.

Now might not (yet) be the time. But it does seem to have become clear that the question is not 'if the community will move somewhere else', but instead the question is 'when the community will move somewhere else'.


Solution: Do nothing and accept the ways how Stack Exchange works.

For the moment there seems to be not enough discontent with StackExchange and StackOverflow that makes a large share of the community think about change. It makes a move to another platform difficult.

Without a community-wide incentive to move, it is too much effort to re-start the community elsewhere and build another library with statistics knowledge. Or at least, the largest part of the current community is not much thinking about leaving and starting alternatives to create a more democratic internet. So this will cripple any effort to start something new elsewhere.


You must log in to answer this question.

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