Sunday, May 3, 2009

StackOverflow and ServerFault Are The Future





Going out on a limb here: the StackOverflow and ServerFault websites are the future for finding answers to your questions about SQL Anywhere. The first site is for programming questions, including SQL, and the second site will be for "system administrator" questions so that will cover system setup, peformance and tuning and so on.

At least, I hope that last part's right... ServerFault is currently in beta, and no, I'm not in the beta; I can't even get access to the FAQ. But it's from the same folks who gave us StackOverflow so it's probably not vaporware.

Folks who read my earlier rant about StackOverflow may be surprised at my new position. The truth is, even as I was writing that article, I knew that I'd be back. I knew that StackOverflow was so much better than all the alternatives that my negative opinions meant very little... ok, they meant nothing.

I knew that eventually I'd suck it up and return to StackOverflow.

So here I am, having just watched Joel Spolsky's Google video Learning from StackOverflow.com, all excited again...



You don't need to know anything about StackOverflow to watch the video, but you will know all about it by the time it's over.

If you're already familiar with StackOverflow, it's still worth watching; you'll learn all sortsa stuff that isn't in the FAQ:

  • In its first 8 months of operation StackOverflow has reached about 30% of the world's English-speaking professional programmers.

  • Joel is most proud of the fact that answers to new questions appear right away.

  • He is least proud of the common practice of closing off questions that are off-topic without having a place to send people. The user community is too focussed on the purity of the home page, and off-topic questions are jumped on in ways that are not friendly to newbies. Also, edit wars are recognized as being an ongoing problem.

  • The observation that "people are really good at tagging their questions" surprised me... and makes me much more optimistic about StackOverflow's future. It might have something to do with the fact that while StackOverflow's user interface works well for professionals it probably wouldn't work well for the general public; i.e., a StackOverflow for "gardeners" might work, but not one for "gardening".

  • StackOverflow is place of choice for questions about new technology; e.g., it's the number one resource for iPhone programming. New technology users tend to organize around a StackOverflow tag. SQL Anywhere may have been around for years but it's "new technology" too: web services and clients, built-in HTTP server, XML, JSON, PHP and C# stored procedures, and so on.

  • The badges you earn in StackOverflow are Napoleonic in nature: "A soldier will fight long and hard for a bit of colored ribbon."

  • The search facility isn't very good; it uses SQL Server. No surprise there, but it's kinda moot: the majority of StackOverflow hits come from Google search.

  • The StackOverflow organization: four people, two servers. No mention of backup, sure hope it's not another Ma.gnolia waiting to happen.

  • Future possibilities include lots and lots of vertical StackOverflows; e.g., StackOverflow for tax accountants. Also: StackOverflow for recruiting.

  • People have asked about licensing the software to large corporations, and although "in the long run it completely makes sense", they don't have the resources to do anything about it right now.

  • ServerFault will be up "in a couple of weeks" ...which might mean "any day now", now. If StackOverflow is any indication, ServerFault will be useful from Day One.

  • On the subject of program maintenance: There is nothing you can do to a large, working software product that is worse than deciding to simply start over and rewrite it from scratch. Examples include Netscape and Perl 6.
Why am I so interested in StackOverflow and ServerFault? Because for all sorts of reasons NNTP is doomed, and if the public SQL Anywhere Q&A process is going to survive it has to move to the web. Today. Not a year from now, not six months from now.

It has to move to a modern website, one that works well, one that's popular, one that's fast and Google-searchable.

Today.

4 comments:

Ron Hiner said...

I love the concept... but the signal to noise ratio is way too low. As near as I can tell, 'ignoring' tags simply dims the question -- I don't want to see it at all.

I see the future of this in very tight verticals... 'server management' (serverfault.com) I think is perfect. 'software development' (stackoverflow.com) is to broad and has way too many subjects -- unless they can find a way to filter the noise, and respect that one person's noise is another's answer from heaven.

And I've often wondered where SQLAnywhere would be in the market right now if the user community was headquartered in HTTP world instead of NNTP.

Anonymous said...

You can get to beta easily: see http://blog.stackoverflow.com/2009/04/server-fault-private-beta-begins/

Jeff Atwood said...

Hi Breck -- fantastic, glad to hear it!

Also, you might be interested in our full-blown meta discussion site we just launched:

http://meta.stackoverflow.com

I agree that your previous experiences were less than optimal.. we're hoping to have a saner "move this meta discussion over here" migration strategy in place.

Justin Willey said...

Now that the Stack Overflow / Server Fault etc platform is being licenced, http://stackexchange.com, perhaps it's time for the SQLAnywhere community to start thinking about the future of the newsgroups.

I feel that the newsgroups, which have been a fabulous resource, are increasingly in-effective. Having the full history is invaluable, but hard to search well. The lack of tagging makes searching inefficent and posts are frequently missed. You have pointed out the unreliability of Google group search, and searching the full 121k messages in the main group alone with MS Desktop Search (never mind Outlook Express) is messy.