Are there any stats for Documentation about the activity in a specific tag? I have observed that a niche tag like Matlab is pretty much dead (last action on Dec 2nd). I personally lost interest in contributing to that niche a while back.

Is this a common trend for the whole docs?

UPDATE

Stats that would be interesting to look at (as suggested in the comments):

  • number of links to Documentation in SO Q&A
  • number of links within Documentation (proxy for repetition)
  • contributions, i.e. edits, size of edits, new topics, ...
  • who is contributing
  • views per topic

It would be nice to have them over time.

share
106  
Docs were a failed attempt from the beginning IMO. – πάντα ῥεῖ yesterday
13  
What kind of stats would you like? The stats I have handy show a huge uptick lately, but may not reflect your area of interest; I can probably dig something up if you let me know what sort of metrics you're interested in though. – Shog9 yesterday
38  
The official MATLAB documentation is awesome. So, in my opinion, there is really no need for an SO documentation on MATLAB. And that is probably why it is more or less dead now. Still, there are lots of badly (if at all) documented languages, frameworks, and tools - so don't judge based on one tag. – hbaderts yesterday
11  
So far SO Documentation doesn't come up in Google search results. I think if it did it would be used a lot more. – Suragch yesterday
13  
I think it could be really great to be able to close SO questions as duplicates of Docs entries. – TigerhawkT3 yesterday
24  
Throwing away free reputation, people copy pasting from random websites, robo reviewers are ready to accept minor, unhelpful, incorrect changes, what else can go wrong? – Mr. Alien yesterday
21  
Well, I don't know if its dying or not, but the fact that garbage like this got approved by a C++ silver badge holder does not in any way make me confident that Docs.SO is producing quality work. – Nicol Bolas yesterday
44  
@Shog9: One of the prime motivators for Docs.SO when it was first suggested was to "mitigate a source of repetitious (often exact duplicate) Questions". Has that actually happened? Since Docs.SO went live, have we seen any change in the number of questions being marked as a duplicate? Similarly, what are the numbers on Docs.SO citations, particularly relative to citing other SO posts? Have they changed over the last few months? – Nicol Bolas yesterday
29  
Documentation honestly feels like it fell victim of what it was trying to prevent. Its initial goal was to make documentation for languages that lack it, but - in my personal, very subjective opinion - the reputation gains ruined everything. As soon as you start giving rewards for something that might already exist somewhere else you can bet someone's going to milk that instead of making something new and original. I have a strong belief that while there's any reputation to be gained from Docs it will fail to serve the purpose it was built for. – SeinopSys yesterday
6  
@Trilarion Q&A's purpose is to answer questions and be rewarded for it. It does not dismiss the possibility of someone taking information of another place and get reputation for it here, quite the opposite. Documentation's purpose is to fill the gap that some language developers failed to. Taking information from somewhere else in this case should not be rewarding, as it just adds to the already existing fragmentation. If there was no reputation incentive, we could be more certain that editors and reviewers are in it help others, not to get imaginary points in return for their actions. – SeinopSys yesterday
6  
I love the concept of Documentation by example. And when I was contributing I tried to focus on creating a code snipped that would work as soon as possible after pasting it into an editor. I think this is what Documentation project should focus on, rather than syntax description. – lonelyelk yesterday
10  
@Nicol Bolas: That reviewer has given "557 non-wiki answers with a total score of 692." I don't know C++, but based on my experience with the [c++] tag, that sounds like a really abysmal ratio. I'm not particularly inspired by them holding a silver tag badge. (And that's why I don't think awarding documentation privileges based on tag badges works, or scales, well.) – BoltClock yesterday
10  
I stopped reviewing Documentation when I found that I couldn't flag anything. People can copy-paste official sources and Wikipedia to their heart's content, it gets robo-approved, and the only way to stop it is custom mod flags. It needs much better quality control. – S.L. Barth 21 hours ago
8  
@S.L.Barth I guess the review process as always been an issue on the main site, I don't know why we thought it would magically be fixed in the documentation reviews. – Ashley Medway 20 hours ago
4  
@Evk: 557 answers with a total score of 692 amounts to a mean of about 1.2 score per answer. That's potentially at least a few hundred answers with a score of 1, 0 or negative, even (especially) if we consider a handful of answers that lucked out and have really high scores (Good Answer, Great Answer). For such a high-traffic tag as C++ that hands out votes like candy, your answers would have to be consistently mediocre or not great for so many of them to not have that many upvotes, if at all. – BoltClock 18 hours ago

Is this a common trend for the whole docs?

It definitely seems like most experts have given up on it. Most posters that are still active seem to have poor to average knowledge about the topic and therefore write sloppy or incorrect articles.

Instead of becoming the canonical documentation over a certain tag, Documentation has now turned into a random dump of "this is cool", "this is a fun trick", "teh codez". Not to mention countless duplicate posts.

(Have a look at for example the C++ Documentation - I don't know if I should laugh or cry. For example, go look for info about the const keyword and const-correctness and you shall find at least 5 different categories about the same thing. None of them named type qualifiers, which would be the correct and formal category name. I would imagine this is because Average Joe has no idea what a type qualifier or a qualified type is, so he makes up his own category "stuff about constant variable things".)

Personally I stopped contributing pretty much immediately, because of the low quality, the lack of organisation and the lack of policies. So much crap was posted that the whole thing quickly went far beyond repair during the very first week.

At this point there is nothing that can be done but to either nuke all Documentation contents and start over with much stricter rules, or to ditch the whole project and move on.

share
13  
Very clear diagnostic, 100% agree with all of this. – ken2k 21 hours ago
8  
This. Documentation should have never existed. It seemed a bad idea when it was proposed, it sucked when it launched, and still is the same crap. Burninate it! – Oriol 21 hours ago
27  
As the [css] top user and [html] number 2, my comparison of their Documentation tags to W3Schools ("W3Schools is better") holds true. Half of them are mediocre attempts at examples, the other half are attempts at copying official or other third-party docs while still somehow managing to get things wrong. How can people - in plural, collaboratively - make something that's worse than the already notoriously bad W3Schools?! I honestly don't remember the last time I approved an improvement request that added content rather than deleting it. – BoltClock 21 hours ago
3  
This is exactly how I felt after a week or so of contribution, and hence stopped. – Oleg 20 hours ago
1  
@BoltClock To be fair(ish) that metric might be metered a bit by the fact that silver-badge holders don't need a review process to make edits. So you don't necessarily review (m)any high-quality edits/additions, by design. – TylerH 19 hours ago
2  
"None of them named type qualifiers, which would be the correct and formal category name." However correct and formal that name would be, you still have to know that's where you would look up info on const. And quite a few C++ programmers don't. So such an organization scheme is utterly useless to them. Docs.SO should not look like cppreference.com. And that's fine. – Nicol Bolas 19 hours ago
12  
So can we now declare documentation to be dead officially? As an expert, I waited for exactly that, so that we can now start over and focus on quality this time… – Bergi 19 hours ago
    
"At this point there is nothing that can be done but to either nuke all Documentation contents and start over with much stricter rules..." Doesn't sound like a really bad idea. Would be a waste of the work that went into the framework if not, I guess. However maybe you would need to nuke only some tags. For example the Matlab tag mentioned in the question suffers much less, maybe because the language doesn't know so much fancy stuff like type qualifiers. It mostly suffers from an infinitely large number of trivial and uninteresting (but otherwise absolutely correct) examples. – Trilarion 16 hours ago

Is Documentation failing?

I think it is. Take for instance one of the top 3 tags, the C# one.

The number of topics didn't change for weeks. The activity tab shows no update at all for the past 2 days, barely some additions for the past week, the majority being done by a single or two people.

I don't even talk about the quality of the last additions.

Once the initial easy-to-win reputation has been earned by rep-whores, almost no one else appears to be contributing anymore.

share
4  
They had a lot of attention in July/August when the idea was fresh. If only the system had been more mature then.... Maybe they just went public too early but then how do you make the necessary experience to correct your errors without spoiling the fun for the audience? Also: Not sure what activity would be expected? Surely less than in Q&A. In Q&A there is a lot of duplicate questions, while a wiki-like approach will saturate at some point. So we should not compare activity with Q&A but rather with other documentation systems. How much daily activity is there for MDN or cppreference? – Trilarion yesterday
    
"..almost no one else appears to be contributing anymore." And since contributing content is the one thing StackOverflow cannot provide by themselves, the question is, why. My speculation would be (in case it truly fails) that the framework (rep, the structure, the organization) just wasn't what people expected. I would still kind of believe in the example-centric documentation idea. But it also might be that just not enough people believe in the existence of good enough examples. – Trilarion yesterday
1  
@Trilarion I don't think one should compare some community driven content such as Documentations, with other documentation systems such as MSDN. On MSDN, once a new framework version is released, the huge pile of new doc that is associated with new features is added. You wouldn't expect constant additions over time, more punctual additions when new products are released. On the other hand, for something designed to be built by the community, if you don't see any new addition in a while, then just like for any other community-driven project, it's kinda dead. – ken2k 21 hours ago
5  
@Trilarion As a side note, as a gold C# badge owner, I can tell the C# documentation is (really) far from perfect as of today, so no activity on it is definitely not a good sign... – ken2k 21 hours ago
    
I totally agree that it is not a good sign. Maybe the experts are just waiting to see what the StackOverflow team will make out of it regarding the framework. All that would be necessary are a bunch of experts feed up with answering duplicate questions on Q&A all day long and willing to build up a canonical catelogue of good examples and there you go. They could coordinate with a thread on meta and wipe everything they don't like (deem useful) and create what they want using all the tools. Maybe this is by itself not an interesting task, or it is too complicated or... – Trilarion 16 hours ago
    
... or the framework is just not the right one for the task, then Documentation will fail. But my guess is that every moment someone could go out and change something in Documentation. It's kind of usable. StackOverflow failed to deliver a really good system at the beginning of the public beta and therefore wasted a lot of good will, but they also modified their course. If you would see it now for the first time, would you use it? – Trilarion 16 hours ago
2  
@Trilarion you are a glass-half-full person. SO.docs has already failed. There is nothing that can save it. There were fundamental flaws built into an inherently poor design of a product that's ill intentioned, cynical and conniving. Sometimes this recipe of traits and exploitations is successful, like Facebook. Most times it's not. The originators of the idea went a little too far in their desperate attempts to extract time and coin from their audiences, this time. It has had a negative effect on all of SO, too. Kind of like Google Glass. – Confused 14 hours ago
    
@Confused Maybe, maybe not. I don't think it's really evil. Other things are but this? It's mostly the design that's not fitting. – Trilarion 5 hours ago
    
Designs are always a result of the people behind them, and their intents, dreams, desires, views, limitations, etc.... As to the design not fitting, that assumes the targets and objectives they had are the same you're considering. I highly doubt that. I think you're looking at their objectives through very naive and rose tinted glasses. – Confused 5 hours ago
    
@Confused I doubt that. I'm not stupid and when contributing to a StackExchange in any way I typically know who profits most from it. If you don't like their business model you could just not contribute to it. So, are they incapable or do they have an evil master plan? I don't know but I would bet my money on them being just not capable enough. – Trilarion 4 hours ago
    
"This is where we are at right now, as a whole. No one is left out of the loop. We are experiencing a reality based on a thin veneer of lies and illusions. A world where greed is our God and wisdom is sin, where division is key and unity is fantasy, where the ego-driven cleverness of the mind is praised, rather than the intelligence of the heart." ~ Bill Hicks – Confused 4 hours ago
    
@Confused Okay, but what do you propose to do? Celebrate the failure of Documentation? Create a better Documentation somewhere else? Contribute further on StackExchanges. Boycotting StackExchanges? If the Q&A part dies, I would be quite unhappy, if SO Documentation dies... I wouldn't care much. – Trilarion 1 hour ago

Is Documentation failing?

Maybe. It's probably a bit too early to tell.

(Warlords of) Documentation was meant to be an example centric approach to Documentation including unicorn points. It features tags, topics, examples, votes on examples, requests, reviews and now also discussions.

Activity was very high at the beginning of the public beta with a strong decline which is quite typical for many new things. Now we are in consolidation phase with rather low activity. Interest could pick up or the idea could die slowly. Time will tell.

So far activity is not zero. Let's take another tag, let's take the Android tag and you see multiple modifications per day. Is this too low? I don't know. You would have to compare with other documentation systems.

The Stack Overflow team showed commitment to improve Documentation. There are regular updates and new features. For example they greatly lowered the rep gains. So at least there is ongoing technical support.

Documentation does not seem to be extremely popular with search engines but that may be kind of expected given the age and the quality of the content (I experimented a bit and google seems to currently prefer other more established sources of documentation).

The true questions are:

  • How high is the quality of the content currently?
  • Is the system capable of delivering high quality content at least in the future?
  • Is example centric the right idea?

I'm not convinced of that, but currently I'm also not convinced of the opposite. I'm just waiting and watching it. I'm only sure that if high quality content is there, search engines will eventually list it prominently.

Summary: A fair estimation is that probably no-one knows if Documentation is failing. It might be too early to tell. Stack Overflow team members should have more information/statistics, which maybe they want to share? Anyway, the only thing that is sure in my eyes, is that Documentation so far is not an overwhelming success. There were strong critical points always mentioned and the stance of Stack Overflow was as far as I remember that some things in life you have to try out.

So far, Documentation is not a success.

share
4  
I think that example centric idea is the main thing that majority of SO visitors are looking for. – lonelyelk yesterday
4  
@lonelyelk But still even with examples at the focus you can still fail by for example making examples hard to find, or making examples about topics no-one needs, or making trivial examples. – Trilarion yesterday
2  
true that. But almost all examples I find with google in SO Q&A are useful. I guess it would be great if Documentation examples could do that. Hard to say if it is achievable though. – lonelyelk yesterday

I asked for metrics in the comments; the suggestions I got were links, activity, and views.

Linking and activity:

Month      InterLinks ExtraLinks MinorEdits MajorEdits Topics Examples ProposedChanges ImprovementRequests TopicRequests Votes 
---------- ---------- ---------- ---------- ---------- ------ -------- --------------- ------------------- ------------- ----- 
2015-12-01 1          null       40         145        13     71       177             19                  3             112   
2016-01-01 3          null       36         205        34     86       183             32                  null          102   
2016-02-01 null       null       50         233        24     101      234             35                  1             163   
2016-03-01 6          null       99         689        126    351      640             194                 8             327   
2016-04-01 33         null       254        1197       216    617      1102            210                 11            610   
2016-05-01 6          null       44         131        18     59       151             25                  1             84    
2016-06-01 10         null       144        528        65     301      502             88                  8             433   
2016-07-01 560        363        7870       24390      3248   11653    29605           6243                1063          29228 
2016-08-01 349        586        2538       6035       802    2926     10966           2066                531           12402 
2016-09-01 291        344        1577       6971       1244   3313     4937            932                 347           6430  
2016-10-01 137        329        741        2469       444    1128     3263            620                 174           4924  
2016-11-01 109        286        673        1525       242    652      2631            374                 137           3640  

graph of the numbers above

Views:

As with Q&A, I don't have a particularly good way of tracking views over time... So here's a graph from Google Analytics:

google analytics page views - take with grain of salt

As with the activity graph above, this suggests an initial spike in activity when Docs went public, followed by a gradual fall-off over the next couple months.

Folks with the site analytics privilege can compare the shapes here to activity on Stack Overflow and other sites in their early days.

share
    
It's not clear how much the total number of Links, Topics, and Examples is developing because possible deletions of links, topics, examples are not listed. Do they play a significant role? – Trilarion 16 hours ago
    
I just ignored deleted things for the purpose of this post, @Trilarion; at this point I think they'd be more of a distraction than anything. A year on, might be more useful to consider how deletion plays into activity. – Shog9 16 hours ago
    
Since you cannot count views over time, perhaps you could just give a snapshot of total views for the top 10 (or more) tags? – Travis J 12 hours ago
1  
select sum(ViewCount) from DocTopics -> 563091 @TravisJ – Shog9 12 hours ago

It was a fun beta, we now know what to expect of the system and its users. Now tweak some variables (reputation needed to participate, reputation needed to approve, vetoing and rollback rights) and implement what has been missing from the start:

Discussion / Talk

It was about the first thing I asked for in the closed beta, and I haven't seen it implemented in the year following it.

If you can't coordinate with your fellow users, if you can't explain to current and potential contributors what the documentation for a given tag is supposed to look like, the only thing you're going to get out of it is people throwing mud against the wall and seeing what sticks.

So implement that, archive all existing documentation and start over.

See also:

share

Failed, is the answer to your question.

You don't need stats to see this. Some things are obvious.

TL;DR When folks with a puerile love of crafting, implementing and enforcing rules apply themselves to building a creative platform (as opposed to a responsive framework built on and of inherent imbalance), the shit hits the fan.

Google Glass was fundamentally flawed in a similar manner. It failed to understand the needs, desires and concerns of its audiences.

Potential users weren't happy about the pricing, nor that it was being rolled out as an endlessly uncertain beta in an obvious and failed attempt to obscure the uncertainty of its direction and purpose, and the compromises inherent to its conception, execution, design and implementation.

Further, these first class users, expected to pay the price of their time, cash, reverence and reputation to support and promote these devices were also expected to develop content for the platform. The expectations and desires of the platform are simply too high and stink of hubris.

Similarly, those best equipped with the wisdom and insights required to give great explanations of suitable demonstrations, samples and examples were flummoxed by the equivalency given to people with near no ability beyond press-record, copy/paste and exploitive desires.

Those potentially in front of someone using Google Glass were uncomfortable with the uncertainty of a camera being pointed at them, all the time, and not being able to ascertain its state or activity of the user. This is somewhat akin to not knowing the experience, wisdom, sources, insights and motivations of those "creating" SO.docs examples, nor any meaningful criticims of the content existing.

Despite considerable astro-turfing, the observations and criticisms of legitimate commentators and thinkers that Google Glass and SO.docs were fundamentally flawed in design, intention and execution rang truest. That nobody in a decision making capacity took these genuine concerns and criticisms seriously speaks volumes for the decision making process.

You can't regulate good behaviour: a culture, community and belief system must be created, cultivated and curated that promotes it. Most people have an innate understanding of this, except for one very special group of people: misanthropes. Their peculiar view of the world sees them thinking there can never be enough fascism and there's nothing wrong with any form of legal exploitation. This odd contradiction fits their peculiarly irrational logic perfectly, as they side with themselves against their own species, with the rules against humanity's best aspects.

Google Glass comes from this perspective, as does SO.docs

share
4  
On the face of it, similar equivalency would obviously seem to apply to Q&A, which was built by essentially the same group and has similar values of equality. This post completely fails to differentiate these, then compounds this by making unsubstantiated accusations of astroturfing, misanthropy, and fascism. (In particular, suggesting that any time some organization fails to consider criticisms adequately, it's because misanthropes are in power, is at once staggeringly misguided and remarkably unhelpful in avoiding such mistakes in the future.) – Nathan Tuggy 12 hours ago
    
@NathanTuggy I'd like to draw your attention to the part you missed, about SO.Q&A: "(as opposed to a responsive framework built on and of inherent imbalance)" Your other comments seem to fall inline with the degree of obliviousness required to miss this obvious reference. – Confused 11 hours ago
1  
Suffering from @NathanTuggy's myopia? You ask questions of Google, it responds. You ask questions of SO's community, it responds. These are responsive environments, inherently unbalanced because those asking don't know what they don't know, nor what the system/others know, or even how it'll respond. Within this environment there's sufficient imbalance to architect and orchestrate rules evolvement and enforcement. Google Glass, VR & documentation platforms require creatively motivated contributions for success. Google, Facebook and SO are good at rules in imbalance, not spurring creativity. – Confused 11 hours ago

Answering the question as a learning user of Python, Tkinter and Crypto: the documentation helps to narrow down the long searches in SO looking for the answer to simple questions you have while trying to develop an application. In SO the questions are almost always answered after difficult plowing to separate the many 'chaff' answers from the real one. I found in the Documentation many fast answers with good examples. What I think the Documentations is missing is the combination of some subjects examples into a single one; if not is falling into the danger of becoming a difficult search like SO. But Documentation helps! Keep it up!

share
8  
Searching in SO is difficult because the team prioritizes things like the unikong game, rather than improving the search. – Oriol 21 hours ago
11  
@Oriol It's so true, we took 43 of our engineers off their projects for about 3 months and they did nothing but Unikong. We think it turned out pretty great though. – Nick Craver 20 hours ago
3  
@NickCraver You should have put some of your marketing team on it too, I didn't know about the game yet and missed out on tons of fun :( – Stijn 20 hours ago
12  
@Stijn Thanks, that's an excellent suggestion! We'll allocate at least 30% more resources next year. – Nick Craver 20 hours ago
11  
If you're using the Stack Overflow search, you are categorically doing it wrong, except in certain special cases where you really need features that only the SO search can provide. (Hint: you don't.) Use Google. They mastered the search business a long time ago. Constrain it to site:stackoverflow.com if you want, but it often isn't necessary. You don't care if the answer you find is on SO or not; you just want the answer. – Cody Gray 18 hours ago
4  
To add to @CodyGray's comment, if you want StackOverflow answers to show up on the top of your google search, just type "StackOverflow" at the end of your question. I do this all the time and Google automatically moves the SO answers to the top. This is way better than just searching SO since, believe it or not, once in a while SO might not have an answer to your question, yet a different site does. Sometimes this is even another StackExchange network site. – Tot Zam 17 hours ago
    
@tot I can't see why adding "stackoverflow" to the search query would be preferable than using Google's site: operator to constrain the search. You say that "once in a while SO might not have an answer to your question, yet a different site does. Sometimes this is even another StackExchange network site.", and while all of that is true, appending "stackoverflow" to your query seems like it would disrupt that the same as site:stackoverflow.com. – Cody Gray 2 hours ago

You must log in to answer this question.

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