I don't see the stack overflow model as being a great fit for what facebook api developers need. The main problem I saw with facebook api was that the right way to do things kept changing, and the docs and forum didn't keep up well. You basically needed to begin facebook API work by reading the entire blog in chronological order and then keep on scanning the dev blog daily, which may work ok if you're a fulltime facebook API developer but it is not a good fit for contract or part-time or hobby development.

I think changing APIs is begging for a wiki much more than a Q&A site. Facebook used to have a wiki for their api, but they went and deleted it, which I considered basically unforgiveable and proceeded to do my best to avoid having to touch the API after that.

In short, bringing back http://wiki.developers.facebook.com/ and focusing on keeping it current would be a much better start than a Q&A site, even a really good Q&A site. You'd still have the need-to-keep-with-changes problem, but at least you'd have somewhere to start from, and wouldn't keep getting referred to out of date information when googling answers.

Yes I know Stack Overflow has community editable wiki question styles, but then you just compound the problem and now need to update 100 different 'how do I [x]' questions every time the API changes.

Many developers I know, myself included, avoid facebook integration like the plague for just this reason. Other services manage to version their APIs and depreciate them gracefully. Facebook apparently doesn't understand how to do this, or simply doesn't care.

> Facebook apparently doesn't understand how to do this, or simply doesn't care.

They didn't care. They didn't have to. Now they do, because they have to. (Hello, Google+)

I disagree - I find the stack overflow model much better than a wiki. In practice the search on SO is much better than a wiki because of how the questions are phrased. Using your how do I (x) example that is how the SO title would be phrased and that is how a developer searches for a solution.

Facebooks API docs are improving too - so this combination of quality docs and SO could be great.

The SO model works better is the API never changes - but since it does, a Wiki would probably be best.

I've both responded to SO questions with updates, as well as had one of my top answers be replaced by a newer, better solution. In browsing questions, I've also seen a number of updates to existing answers.

I guess the point is, if you've spent a lot of time figuring something out, and you've proved that a number of solutions on the internet are now wrong, you're tempted to make it easier for the next guy. I think this is true regardless of the medium, but SO seems to have done well at creating an environment where it happens.

That's supposed to be what the Change Log is for: http://developers.facebook.com/docs/changelog/ But yeah, they need to work on the relevancy of that content in regards to the constant changes going on.

Maybe developers should just start an unofficial Facebook Wiki.

Most people would rather ask and answer questions than maintain or update a wiki. That's what they have been doing, over at stackoverflow.

Since people are already discussing Facebook API over at Stackoverflow, they might as well make it official.

I disagree.

Why try and build your own community when there's already an existing one with a lot of useful questions, engaged moderators and smart people answering?

Especially when you've failed to support your existing community at all.

You do realize we already have 13,700 facebook related questions? http://facebook.stackoverflow.com/questions

Yeah but that takes work, why would they want to do that when they could just let the community-based Q&A site take care of it?

They are still keeping the API docs though, and hopefully they will be able to re-allocate resources from managing the forum to keeping the API docs up to date.

Maybe I've missed it but nobody has mentioned that SO allows posts to be in a community wiki format. Doesn't this reach a nice balance between the two?

Anybody can suggest edits ( http://blog.stackoverflow.com/2011/02/suggested-edits-and-ed... ) without even registering.

Community wiki posts can be outright edited by low rep users, and higher rep users can edit any post.

The idea that a wiki is better for rapidly changing content is... odd, to me; it seems it'd be a wash at best. Then again I'm an SE Inc. employee, so perhaps I just don't see the difference.

Agreed that the wiki would have been useful. I think the Stack overflow site would just be a replacement for their forums

I agree. Their platform changes more often than their underwear. A wiki would be very helpful. I hate to search for a problem and find 10 different solutions all of which have been deprecated. Curiously, even in the old documentation wiki developers were not allowed to edit pages. I think there must be some FB development wikis out there, lost in obscurity.

Should I feel better about Facebook trying to crowd-source their developer support? This doesn't address any of the real concerns with usable documentation or proper API deprecation. I get the impression that this is an attempt to spin past bad behavior into acceptable future bad behavior.

I don't love that facebook.stackoverflow.com looks exactly like stackoverflow.com — there’s no easy way to tell where you are, or to get back to the main stack overflow.

Even more confusing, you can get to un-Facebook-related questions on facebook.stackoverflow.com (e.g. from your inbox, or from a user page). Since there are no redirects between regular and Facebook SO, this means that every question on Stack Overflow now has two URLs, and you might see links to both.

At first I figured they just created a separate SE site specific to Facebook development but it appears to just be a filter on all questions which have a tag that contains "facebook" while hiding all others.

Yeah, that's what I thought.

Why not just create another stack exchange just for facebook? I mean, they've already got that system working for other topics.

Putting a filter on all SO questions and slapping a subdomain on it feels like it's a solution they created in 30 seconds.

Well, first off Facebook API questions are still on-topic on SO, so they will have more eyeballs and therefore a greater probability of being answered.

Second, programmers can keep their existing reputation, while a new SE site would put everyone back to 100. Since the topic is a subset, keeping it makes sense.

I don't really see what problems this solution brings that can't be solved with a custom logo.

I agree with your second point, but your first point goes against established precedent at Stackexchange.

I mean, if it was about getting more eyeballs then I think the best way to go would have been to use tagging, so that questions would be posted to 'stackexchange' and would then be tagged with 'programming', 'codereview', 'superuser' etc and as they were tagged they would show up on the programming.stackexchange.com (and other) sites. IIUC the current way is that if I post a programming question to stackoverflow, it will get 'migrated' to programming.stackexchange. That works, but it detracts from the total number of eyes because it is removed from stackoverflow.

That design decision is something that has always kind of confused me about the Stackexchange sites.

It's about the topic of the site. FB API questions are on-topic with SO, others aren't. The Code Review is debatable since it's arguably a subset of SO, but other SE sites (like Programmers) don't really intersect that much.

> Why not just create another stack exchange just for facebook?

If Facebook wanted to, it is not like they can't write a question-answer site or use one of the open source alternatives. This is an endorsement of Stackoverflow from facebook. Nothing more.

More and more this is just a patch for the problem, while what they need is a 180 degree turn and commit to actual developers.

Facebook treats its developers just like its users and throws out updates, new features and changes very very often. With so many users they can afford it for their users, but they need to see they can't just do that for their developers. As a user there are so many little bugs and quirks in Facebook, especially in the News Feed, that popup and a few days later disappear. Facebook needs to realise they need developers too, and that many advertisers run campaigns on their facebook pages, meaning that the API is not just a fun addon for users, but a money maker too.

I agree this is mega confusing, we are probably going to change this and introduce a meta.so style brand for the mini site

"We have been ignoring you developers a bit much on the existing forums, so, how about a do-over where you guys help yourselves?"

Accurate or no?

Very accurate. Facebook makes it clear that they don't give two shits about developers because they're top dog and they don't have to. This just solidifies that impression.

They should improve the clarity of their the documentation and add more examples, in addition to keeping them updated... or pay a couple people to answer questions. God knows they have the money.

See also the blog.stackoverflow.com side of the partnership here:

http://blog.stackoverflow.com/2011/08/facebook-stackoverflow...

Basically, it's a siloed version of stackoverflow.com containing only the tags relevant to Facebook. They're thinking about creating mini-sites for other major subdomains (see also http://meta.webapps.stackexchange.com/questions/913/is-this-...).

Hate the idea. SO's search is decent. This is just the kind of fragmentation that I already don't like about programmers., unix. etc.. You can easily filter your lists by tags, you can easily search by tag.

Where's the benefit?

Facebook will likely benefit a lot more.. They had a crappy forum, and they can let the rest of the folks actually do their support work I think :)

Well, come to think of it, I think google's support forums suck even more. You can post a problem on google support forums, and pray and hope for an answer from their developers.

Purportedly, this works better than filtering by tags.

I'm not sure that this level of filtering is quite necessary, but I think that the community fragmentation is useful: People go to each site (SO, SE, SU, Programmers, etc) with a specific purpose, and the topics on each site are broken up by tags. Trying to filter both purpose and topic on one site would be problematic.

(Also, the Unix/AskUbuntu/SuperUser debacle is a special case, IMO)

The benefit is SO gets more traffic and users. Facebook benefits by off loading the management of their messy forums and rely on free crowd sourced moderation.

I would imagine that facebook actually writing non-terrible documentation for their API would be a better strategy then integrating with stack overflow.

This can't hurt, but hopefully they don't see it as an replacement for real documentation.

I dont think there was anything missing from the existing FB developer forums except for the main missing ingredient - active participation from FB itself.

If they are going to be unresponsive on SO as well, migrating to the SO platform doesnt make sense at all.

Omg, this is pretty awesome. Facebook API is terrible, and having one, unified and usable channel to bounce questions would be great. Given that we are all used to Stack model, this should really help at speeding things up.

This is awesome. I develop quite a bit on facebook, but turn to SO for most of my needs. The forum is terrible, but it really does not make sense for them to build a new system from scratch.

It’s really just a view into Stack O. A set of filters and customizations under a subdomain. (It took quite a bit of work under the covers, natch.)

If you’re already a Stack O user, this will hopefully provide a focused experience if you’re also a Facebook dev. Or not. In which case it’s ignorable.

If you’re a Facebook dev who isn’t involved with Stack O, it’s a curated experience and hopefully better than what existed previous.

I would love to see companies with APIs and active developer communities provide rewards to high reputation developers.

e.g. Facebook offers devs with over X reputation access to alpha APIs, devs with over Y reputation direct email support, devs with over Z reputation are recommended on the Facebook.com site as developers to hire for freelance work, etc.

Yeah, right. But still, better to have help from ourselves than no help at all.

I like the documentation model of API pages with comments allowed, plus a semi-protected wiki model to edit the API pages. This allows the API maintainer to easily update, provides change history, plus allows the community to step in and provide corrections or code examples without the burden of rewriting the article.

damn..

Joel and friends, Google+? ;(

This is a huge improvement over their developer forums, which are a mess of unanswered duplicate questions, terrible search, and didn't even support Facebook login.

Did you RTFA?

This is just providing a view of Stack Overflow that's limited to Facebook questions, and Facebook staff providing more support there. Facebook has been a login option for Stack Overflow for a long time, but don't even require you to even make an account to ask questions.

Of course it doesn't mean that, but even if it did, what user data are you talking about? The only non-public info that Stack Overflow has from (some) users is their email address and real name, which Facebook already knows.

But facebook can't associate Real Names with questions asked on stackoverflow.

If they had data-sharing in place, they could.

Oh my god, Facebook will know that you asked a question on how to write Hello World in lisp!