stackexchange for osdev?

[proxy]

I just noticed this posting on stackexchange (the make your own stackoverflow site) http://blog.stackexchange.com/post/5184 ... change-2-0. It seems that the next iteration of it will be free. I've noticed that already many people here visit stack overflow and was wondering if the people here are interested in having a stackoverflow site dedicated to os development.

Personally, I think it's a great idea and could effectivly deprecate the forums in the long run. It isn't available for free yet, but I was curious what the people at osdev.org think.

[Brynet-Inc]

No.

Please note that I'm in no way an authority figure on these forums, I just consider the idea ridiculous and unnecessary.

[proxy]

heh, that's fine I figured not everyone would think it's a great idea. But if you don't mind me asking... why is it "ridiculous and unnecessary" (ok, i can see it might be unnecessary, but ridiculous...).

My point of view is that (sorry that this turned into a long list):

a) stackoverflow like sites have proven to be a very effective method for exchanging information between the experts and the newbies.
b) the "tags" system it has is also very effective for both finding information and watching certain topics. (ex: x86-64 seems to be a relative hot topic here lately, would be nice to easily find all posts tagged x86-64).
c) these forums have a certain degree of redundancy (granted if people just searched first that wouldn't happen, but non-the-less it happens). stackoverflow like sites have a nice machanism for dealing with redundancy
d) the search feature for most forums (regrettably this one too) suck. Far too often i just end up googling: "my set of terms site:forum.osdev.org"
e) most questions are not "discussion" oriented, they are "here's what i want" and the replies of "here's how it could be done" (though it's often "you don't really want that" ).
f) let's face it, this hypothetical site would largely be used by the same people as here, so the "flood of newbie" factor is somewhat minimal IMO.

I'm sure there are plenty of cons as well, but it's not like anyone would be forced to use the site .

To be honest, I'd consider doing it myself. But i think it would be a terrible idea to try to "compete" with this site, it already has a good functional community.

Anyway, I have no intention of creating an argument, just curious.

[quok]

Colonel Kernel has in the past suggested stack overflow many times, but even in his latest post in the Newcomer Shield thread he agrees that StackExchange wouldn't work -- it's doubtful we have the critical mass needed for such a thing.

He does propose that we point what we would consider "off topic" posts (those that typically get a RTFM or STFW reply) to Stack Overflow. Perhaps we could give that a try, but lately I've seen a few newbie type posts that have gotten decent replies and haven't just been RTFM'd.

I'm typically against moving things away from this site over to Stack Overflow or related sites, but perhaps we could give Colonel Kernel's suggestion a try. Alternatively, we need to start accepting newbie posts more.

Either way, it's a change from past behavior on these forums, and both would result in extra work, and of course no matter what happens, not everyone is going to be pleased.

Bottom line, I don't think a StackExchange site would work, but perhaps Colonel Kernel's suggestion may be worth a shot.

[Colonel Kernel]

Colonel Kernel has in the past suggested stack overflow many times, but even in his latest post in the Newcomer Shield thread he agrees that StackExchange wouldn't work -- it's doubtful we have the critical mass needed for such a thing.

Yes, there are too few really knowledgeable OS dev experts to create a separate StackExchange just for OS dev. Plus, we'd miss out on cross-pollination with the wider developer community. This is why putting our objective, factual questions on StackOverflow would be better than using StackExchange (and better than the status quo).

He does propose that we point what we would consider "off topic" posts (those that typically get a RTFM or STFW reply) to Stack Overflow.

Not exactly. The criteria shouldn't be whether a post is "off topic" or not, but rather whether it is a straightforward, objective question with a factual answer (i.e. -- not subjective or based on opinion, or otherwise "discussion-y" for lack of a better term).

Examples of questions that would get redirected:

• How do I enable protected mode on an x86 CPU?
• Why do I get a triple-fault when I do XYZ?
• Why do I need to use the --freestanding switch when compiling my kernel with gcc?

Example of questions that would stay here:

• What's a better OS architecture, microkernel or monolithic?
• Linus Torvalds said XYZ -- is he right?
• What's the best way to initialize my kernel's memory manager?
• Can exokernels really work?

You get the idea.

[Brendan]

Hi,

I just noticed this posting on stackexchange (the make your own stackoverflow site) http://blog.stackexchange.com/post/5184 ... change-2-0. It seems that the next iteration of it will be free. I've noticed that already many people here visit stack overflow and was wondering if the people here are interested in having a stackoverflow site dedicated to os development.

[tongue-in-cheek]

Unfortunately, my OS isn't large enough yet, and therefore it'd be a bad time for me to employ a "divide and conquor" strategy. I'd prefer to wait until I've got at least 1% market share before fragmenting OS developer resources in an attempt to reduce the chance of anyone creating competition for my project.

[/tongue-in-cheek]




Cheers,

Brendan

[Solar]

Quoting myself, edited a bit:

I'm a big fan of StackOverflow, but if you look at the difference between stackoverflow.com and superuser.com, you'll see how much of the "drive" behind SO depends on the sheer number of users. (Which they actually state on the SE site: "A Stack Exchange Q&A site only works when it has critical mass: enough people have to go there every hour so that questions get answered.")

I also don't believe in hosting our knowledge on a third-party system, and one in beta phase to boot:

• * One, there would be much work involved in transferring our existing knowledge over. Either that, or we'd be facing a knowledge split.
  * Two, I don't think SE caters to the kind of tutorial / descriptive / technical documentation we have here.
  * Three, if we got a weak point, it's the OSDev / General Programming forum - the Q&A thing that spoon-feeds people custom solutions. Which is exactly what SE caters for: Not going for FAQ links, but to give individual answers in hopes that they get upvoted. (You remember the opposition against voting systems in this forum?)
  * Four, something like SE doesn't cater for information presentation (as in the "Expanded View" of the Wiki) or interlinking of information. Yes, you can link, but you don't get something like BackLinks or WantedPages, the strong points of a Wiki.
  * Five, I think we need the forum, as JackScott said, for the socializing and peer review of ideas. SE doesn't cater for that either.
  * Six, hosting our knowledge on a third-party service makes us dependent on that service. We could back up the current Wiki / Forum and rebuild elsewhere with comparatively little effort, something I don't see with SE.

[Colonel Kernel]

Solar, I responded to all your concerns in this thread:

http://forum.osdev.org/viewtopic.php?f= ... 45#p175722

Unfortunately, it got locked so I can't easily copy my reply and paste it here, but please read it if you haven't already.

[Solar]

{read it}

I don't think offloading OSDEV questions (as opposed to basic questions e.g. about compiler options etc.) to SO is a good idea.

• The OS related questions, and the discussion about them, is what is at the very core of this community. Putting it away somewhere will leave this board pretty much deserted.
• Also, the questions, and the discussion about them, is one of two things that's feeding the Wiki. Having those things happening on SO means there'll be little feedback into the Wiki from that front (because no-one will browse the "osdev" category of SO to look at what's a FAQ or what might be lacking from the Wiki).
• The general SO community isn't about kernel-space coding, and I fear there might be much repetitive noise on that kind of questions ("why bother", "use Linux" and their ilk).

Get me right: It might actually work, and make SO into the prime resource for OSDev related information. But effectively it could mean we close doors on this community here. Do we want that?

[Colonel Kernel]

The OS related questions, and the discussion about them, is what is at the very core of this community. Putting it away somewhere will leave this board pretty much deserted.

It would change the forums dramatically, but as long as there is a desire for community and subjective discussion, it won't turn into a ghost town.

Speaking for myself, I rarely participate in forum discussions anymore, outside of these existential meta-discussions. This is because of the signal-to-noise ratio, not just from beginners asking repetitive questions, but from all the flames that seem to ensue. It's just a complete waste of my time to go digging through all this. I want to go to a forum where I can actually learn something new and have a design discussion with other experts.

I've always thought that Lambda the Ultimate is a good model to follow in terms of the level of maturity and expertise of its users. They achieve this through very strict and impersonal moderation (they tend to just remove posts rather than lecture people), and by having very complex subject matter. If you want to know how to do XYZ in programming language ABC, you go to SO. If you want to know how to design a better programming language or the theory behind programming languages, you go to Lambda the Ultimate. There is room on the Internet for both. I believe it's the same with OSDev.

Also, the questions, and the discussion about them, is one of two things that's feeding the Wiki. Having those things happening on SO means there'll be little feedback into the Wiki from that front (because no-one will browse the "osdev" category of SO to look at what's a FAQ or what might be lacking from the Wiki).

Wikis tend to be maintained by a core group of dedicated editors. Not every passerby gets involved. The people contributing to the wiki today will not forget about it just because a lot of OSDev questions start showing up on SO.

As for FAQs, it's easier to see what the most popular questions are on SO because of the voting system. Those questions that have complicated answers that would benefit from more explanation and deeper cross-linking with other topics could be explored further on the wiki.

The general SO community isn't about kernel-space coding, and I fear there might be much repetitive noise on that kind of questions ("why bother", "use Linux" and their ilk).

If anyone gave answers like "why bother" and "use Linux" on SO, it would get voted down really fast. That's the beauty of their system.

Get me right: It might actually work, and make SO into the prime resource for OSDev related information. But effectively it could mean we close doors on this community here. Do we want that?

While I don't believe that would happen, if I had to choose, I would rather have:

• The entire worldwide community of programmers able to participate in answering OSDev questions
• A fair, impartial, impersonal, and non-judgemental system for moderation (instead of imperfect human mods that sometimes lose their objectivity and sense of diplomacy), and
• An effective way to filter out the noise so that we can see the very best questions and answers quickly and easily.

However, I think it's actually a false choice: We can continue to have our community here, as well as Q&A on StackOverflow.

[Nathan]

I have a account at StackOverflow and other StackExchange sites, and I saw that as a much more powerful working tool than our message boards(vBulletin, phpBB, SMF...). I aprove your idea, because it's much more smart to use it, than phpBB.

The only problem of it is that it uses .Net technology, then our server needs to be changed(and the value of it will be more expensive), to support all the requirements of it.


Also, remember that StackOverflow is just a site.

[Brynet-Inc]

That horrible picture of BillyG is likely under copyright, which probably doesn't allow for alterations.

[Nathan]

That horrible picture of BillyG is likely under copyright, which probably doesn't allow for alterations.

It wasn't altered.

[Combuster]

Your previous banner was cropped, your avatar is rescaled. And you're still wasting bandwith/screenspace.

[earlz]

I would say that Stack Overflow is on topic enough to make an OSDev SE(stack exchange) be redundant. The only thing we can't talk about on Stack Overflow really is how to create drivers. (which we can, but the device has to be very popular to see any amount of views)

Also a ton of the conversations that happen here are *discussions* not simple Q and A. This would be ill suited for a SE.