Hacker News new | comments | show | ask | jobs | submit login
All-Star Team Backs StackOverflow to Go Beyond Programming Questions (readwriteweb.com)
36 points by aditya on May 4, 2010 | hide | past | web | favorite | 38 comments



The closing paragraph:

Spolsky says that the easiest way he's heard people explain the difference between Stack Overflow and old fashioned forums is that when you go to Stack Overflow, the right answer is at the top of the page. That's a charming way to put it and it's sure to be interesting to see the team that's assembled take a shot at building that kind of experience around other kinds of topics.

Now THAT is an elevator pitch: Stack Overflow is just like old fashioned forums, only when you go to Stack Overflow, the right answer is at the top of the page.


That's correct. And the iPad is just a tablet notebook like the one from Microsoft in 2000.

http://www.microsoft.com/presspass/features/2000/nov00/11-13...

And the iPhone is really just a cleaned up smart phone.

StackOverFlow took an old idea, cleaned it up and made a better product. Seems pretty simple after the fact.

Getting the small things right in a product really makes a big difference. StackOverFlow isn't perfect (and neither was the iPhone or iPad) but they got a good minimal viable product and now they're off to the races.


I'm all for Ries, Blanks, lean, mvp, etc. etc., but all the lean jargon is getting overused and abused.


It's not always at the top, though. Call it the reddit effect: sometimes the top-voted item is not the one that is the most informative, relevant or correct, but the one that appears so by virtue of being well written, backed by someone with sufficient karma, and clever enough to get upvotes. The asker may then mistakenly pick this as the accepted answer and the better but less hive-mind-friendly answers get pushed away.

When skimming the site I've occasionally seen better answers at the bottom of a page getting ignored while people shout about the top few answers. If I didn't know better, I would have ignored them too.


This happens a lot less on Stack Overflow than on reddit. The asker isn't supposed to accept the most popular answer, either. They're asking a question because they have a problem to solve. They should be testing solutions and accepting the one that works. I often see accepted answers with fewer upvotes than the most popular ones. The fact that both the most popular and the accepted answer are at the top of the screen is a nice feature of the site.


The asker may be testing out some of the answers, but in reality most of the problems to be solved during software development are about balancing tradeoffs, e.g. security vs. performance. The reason the best solution is the best may not become apparent for years, when the project needs to scale, changes hands, needs to upgrade, gets hacked, etc.

In such a case, how is the asker, lacking such experience, able to pick out the right answer except by number of upvotes? Same goes for anybody reading the page. I'm just saying a popularity-driven Q&A site has its drawbacks.


You're assuming that both the asker and the crowd are incompetent. No kind of Q&A site is going to produce good results under those conditions.


I'm not assuming that. I'm suggesting that certain answers may become more popular and rise to the top for reasons that have nothing to do with correctness. These include but are not limited to: the style in which the answer is written, the assertiveness/aggressiveness of the answer, the decision to include or not include sample code, the karma or reputation of the answerer, and the current vote count and visibility of the answer.

I think it is safe to assume that there is a gradient in experience among SO users and that for those on the lower end, votes may be more influenced by the issues I just listed than any standard of correctness. It's simple crowd psychology, and already well established at other voting sites like digg and reddit.


Digg and reddit aggregate links across a wide variety of topics and have a very diverse user base. Stack Overflow's focus is more narrow (much like that of Hacker News). The crowd is a lot more focused as a result. People are a lot less easily influenced by writing style than by correctness. (Try posting code that won't compile and see what happens.)

Yes, people are influenced by reputation and good writing on subjective questions, but those are discouraged on SO, and they don't really have one right answer anyway.


Stack Overflow is optimized for Q & A, while you can do more than just asking questions in a forum (i.e. actually making friends and building communities).

They are both great. Really, every forum should just have a Q & A section that implements Stack Overflow or use a similar widget.


Stack Overflow is about instant gratification. This isn't a bad thing. It's what it is. Another way to look at it is this way:

Stack Overflow gives a man a fish. Forums teach a man to fish.

Now, when you need a fish, you need a fish. You don't need to learn to fish. But in the long run, forums, and discussion, usually provide better value. This doesn't mean SO can't teach people. However, it doesn't encourage teaching. It's setup for questions and answers, not discussion.


Books aren't set up for discussion. I'm guessing that you think they aren't good for "fishing" either?

The basic idea with StackOverFlow is to ask well-defined questions and get well-defined answers, without the long messy threaded response found in most forums. The worst is when the thread ends in a dead end.

The wiki feature and the ability to include formated code are very useful. The reputation economy is essential, so more questions get answered, including older ones. What happens in a forum when no one answers a question? Ask again a few days later?

Also, they do have short discussions through comments, which I have found to be effective.

Finally, I think you, and many others, are completely missing the point of StackOverFlow, and hopefully other future sites like it. If say, for example, 100,000 people are learning [Lisp, Scala, Java, C++, Go, jQuery, C#, etc], they are going to have the a lot of similar questions that really only need to be asked once, with one or two solutions provided. Now the next 99,999 people don't need to ask the same question. So, now we have taken the FAQ and improved upon it by creating it dynamically through voting, and question matching.

There are various improvements that could be done to StackOverFlow. They are essentially creating a huge database of programming knowledge. Ideally it, or a similar project, could be organized to enable someone to start learning any new topic and become proficient by browsing and searching the site. Questions could be further categorized to enable better filtering (eg. Mac OS bug, compiler warning message, compiler error message), include specific version solutions (eg. Java 1.5, Java 1.6). A Rosetta Stone feature could be added linking solutions. For example, the question: "How do I reverse a string in [Java]?" could provide links to the Perl, C#, Java, Lisp, Scheme, Javascript solutions.


"I'm guessing that you think they aren't good for "fishing" either?"

I'm guessing you think being a smart-ass is effective?

"Books aren't set up for discussion."

No. Books are set for learning as well. Discussion isn't the only means for learning. IRC/Chat is the same way. You can learn with this as well.

"StackOverFlow is to ask well-defined questions and get well-defined answers"

Which is another way of saying what I said.

You then go one to promote several things that SO appears to do.

Reputation, for example, promotes answering questions. My experience is that the answer is succinct, and simple. It provides the answer, and that is good.

It provides the fish.

SO doesn't, however, encourage teaching. Even the setup is simple: ask a question, get an answer. The answer is most likely the answer that provides the quickest fish. Oh, you do find occasionally someone will provide a very thoughtful response, but this is the exception, not the norm for SO.

"Finally, I think you, and many others, are completely missing the point of StackOverFlow"

No, I haven't. My post that you replied to even said that what SO does it does well. You're just getting defensive, whipping insults, and for some reason, seem upset that someone wouldn't kneel down before the all might SO. I wouldn't be surprised if you are an Apple zealot.

My main contention is that SO is setup for providing answers to questions. It's not setup for discussion, or learning. SO's success is a result of human nature. People needing an answer go there, get the answer, and leave. They copy/paste and move on. That's the nature of SO. You can disagree, but you'd be wrong.

SO isn't bad. It provides fish. I love fish (tilapia is a good fish). It's a fish monger. It doesn't encourage providing bait, or tackle. Just look through the majority of answers that are rated highest. They provide code. Code someone can copy and paste. Sometimes they'll provide a link to the place someone can get code, or get a direct answer. Rarely is their discussion as to why someone is asking a question. This isn't encouraged. Answering a question with a question does't get you reputation. The reputation economy encourages answers.

I do want to address this:

"What happens in a forum when no one answers a question?"

I assume you actually mean when no one responds to a question. Answering a question is different from responding to one.

No response usually means no one is capable of answering the question. Rather than provide a half-assed answer for reputation, it simple goes unanswered.

However, no answer usually means the question is a poor one. If SO's great power is in answering questions like "Reversing strings", then it's really not tackling anything difficult. These questions get answered all the time. The questions that are more difficult, however, require discussion. When their is no direct answer. Indeed, suggesting that their is an answer (as SO does) hurts the discussion process. Suddenly, instead of learning, you read the Question, and the Answer, and assume that the Answer is the end of the discussion. It's THE Answer. But it's not. It's merely the one with the most votes.

So no, I'm right. SO gives fish. It doesn't teach fishing.


I expected Google to buy them up. StackOverflow really has become a destination for answers, especially after all the negative goodwill that Experts Exchange and to a smaller degree IT Toolbox have built up by creating barriers to viewing answers.


I'm still looking for words for describing how StackOverflow is so much better than it's competitors. I think it's more than just ill-will for Experts Exchange (though that's undoubtedly a factor). StackOverflow is built right. It runs smooth and the barrier to joining is damn near as low as possible. But what else gives it that extra something?


Reputation, aka karma. People like being recognized.

(That, and Jon Skeet.)



I don't know why more sites don't use Google's Ajax API and instead waste resources on making really bad search engines. Sure, users get more features if you roll your own engine (sort-by-hotness, hide self-posts etc.) but what users want are good results, not features.

I wish reddit, HN, and SO used Google's Ajax. I made a very simple JS/CSS tool to add a Google search box to any site: http://chir.ag/projects/drop-search/ - I use it on my blog and it works pretty well.


This is true. Luckily Google aggressively crawls and indexes SO, so that really is the preferred way to search the site. I've seen results in Google for questions that were posted 10 minutes earlier.


Stack(Overflow|Exchange) is a textbook example of why you might want to take VC rather than bootstrap. Whether they succeed or not, they perceive that their chances of success are higher if they try to get big fast—and there's nothing wrong with that.

Oftentimes the bootstrap vs. venture capital debate gets rather heated and adversarial. But the dichotomy is a false one.


Software developers are the perfect audience for a Q&A site because they do their job online. I think StackOverflow's biggest challenge is extending their audience to people who aren't already on the Internet.


Listening to the plan for the proposal-heavy StackExchange site creation process on the podcast, it seemed to me that it relies on people being super motivated in desperately wanting a StackExchange created for their topic.

I wonder where this drive is supposed to come from for people who are indifferent to the StackExchange software because they haven't participated in sister sites much (or at all).

I mean, I'm even very familiar with and enjoy using the software (StackOverflow), yet I can't be bothered using SuperUser for my general tech needs, let alone creating a site of my own.


It isn't supposed to come from people who haven't participated in sister sites. They're taking ideas from within the existing communities for a reason. They're going to take a handful (or several) from those ideas. Those sites will attract more people in those new niches. In the next round of proposals there will be a bigger pool of users to propose what the next sites should be.


"We're not going to have to worry about the cost of servers, hiring a few people, getting office space. Before I would have walked around New York for weeks looking at all the office space to save $2k/month. We don't have to be careful anymore."

Hmm.


Not a red flag, given Joel's history and experience. Now he can optimize for growth and opportunity cost as opposed to cash flow.


Made me raise an eyebrow too but I think Joel's previous business experience with Fog Creek will help them steer things in the right direction. I don't think a bubble-like spending spree is in their future.


Their single biggest problem is OpenID. Normal people have zero tolerance for such shenanigans. It's confusing to create an account and login. On top of that asking for such information seems like a scam to non-techies. I know, I know there is significant support for OpenID. Yet it's still a distraction to the site. And it still scares people away. This may be acceptable for tech to tech communities but not for the rest of the world.


Has Joel always been listed as the CEO of StackOverflow?


I don't understand why SO doesn't generate enough money for the founders to find an office space without having to worry about cost.

"Before I would have walked around New York for weeks looking at all the office space to save $2k/month."

They receive 25M page views a month. So even at $5CPM that's $100k/month in revenue. I'd think that would be enough income to get the ball rolling in an office downtown.

Maybe they have a hard time selling ads?


"Now Spolsky says the plan is for the company to launch a handful of very targeted sites running the same software but focused on offering objective answers to technical questions in other verticals."

So is this just a pickier StackExchange?

How do they plan to offer objective answers (or is this just restating what StackOverflow does today)?


This article makes it seem like Spolsky had way more to do with the founding/creation of StackOverflow than he actually did. Jeff Atwood was the visionary behind Stackoverflow / StackExchange, and (as a far bigger "voice" in computing than Joel ever was) the guy who built initial consensus/mindshare. Spolsky was (at best) just along for the ride, at least initially.


I listened to their podcasts right from the start which doesn't qualify me to comment. But I will anyway.

I'm just not sure this is the case. Joel seemed to have a much better grasp of the big picture. You could tell that he'd thought everything through really well, in the way that you'd expect from a successful business owner.

There's no doubt that Jeff did all the 'work' between the pair - Joel was mainly advising. But I would be very careful about writing off his contribution. I would personally bet a lot of money that Atwood going it alone would have been far less successful.

I'm not saying that Joel contributed more than Jeff either. When you look at two partners, and claim that one is more important than the other... that's a pretty big call to make. And unless you personally know them, I'm not sure you have the evidence to back it up.


Fair enough. But don't you think the coverage in this particular article was a little Spolsky-centric? I agree they're a team (equals), and don't mean to imply Spolsky was a Johnny-come-lately, but the RWW article really just mentioned Atwood in passing. I thought it was a little strange that his name wasn't mentioned til below the fold but, no big deal.


Have you listened to the StackOverflow podcasts? I give Jeff a lot of credit for leading a the SO dev team, but he has shown several times on the podcast that he can be very clueless (for example podcast #48 where they discuss localization). There was another one where he asserted that taking code you wrote in a job to an interview was fair game. He deserves credit for much of the design of StackOverflow and for his team leadership, but I think you are understimating Joel's "strategic" role in this venture. There is a reason why Joel is announcing the SO financing, and not Jeff. It seems pretty clear to me that Jeff is less capable when it comes to "big picture" and "long term" stuff - and I mean no disrespect by that.

Also, Jeff was never a bigger voice in computing than Joel.


Sure; every last podcast. And Joel does have the expertise on the financing/startups side. And maybe I am underestimating his role on the strategic side.

But in terms of who has the bigger "voice", there's no comparison. Atwood is a great popularizer of tech-related topics; he writes short, sweet, lozenge-like posts that appeal to programmers as well as technical non-programmers; whereas Spolsky's focus has always been more narrow. And going by any metric you care to throw out there (Compete, RSS, Alexa, whatever) there's no comparison in terms of traffic.


joel was the one who had the idea to build Q&A side. without joel, there would be no stackoverflow. jeff was the one who researched the whole concept and built the final product. i think both of them were equally important to stackoverflow.


{{cn}}


Sorry, I guess the above commment wasn't liked much. What I was trying to asy is the above comment sounds a bit short on anything but personal opinion, which I guess is fine... I was in a rush and will write this out next time. Please site your sources when you make such a strong comment about someone's contribution to a successful project because when you don't it sounds like you are just jealous.




Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: