From the guy who founded an entire development ecosystem out of the web framework he built on what was, at the time, an obscure Japanese programming language, that he originally built for a project management application, it seems kind of strange that he'd be surprised that someone might build a complex tool in similar circumstances.
In a parallel world where the cool developers were all attending Wasabiconf, Joel Spolsky Prime wrote a sneering blogpost assuming that DHH Prime was joking about the crazy 'scaffolding' and 'proprietary database mapping' framework he was building BaseCamp on
I would like to remind you that we are talking about Ruby vs essentially VBScript. I mean, I pretty much hated the Ruby hype and especially the attempts to make it seem better than it really was, but we are indeed talking here about web software written in VBScript that ran only on Windows. For many people working with web tech today, that is inconcievable. Not that it didn't have its fair share of detractors back then...
Hats off to FogCreek for building a successful product, but that doesn't mean that people have to admire or even accept their technical choices without any criticism. I for one consider DHH's criticism to be on mark.
Well, 2006 was a -long- time ago, and Ruby and Ruby on Rails have come a long way.
I don't think Joel's conservative approach was all that bad.
He made a judgment call for --his-- company's products at the time. He could afford to, and it appeared to have worked well for his product line. And believe it or not, people wanted to work at Fog Creek.
In 2006, I wouldn't have imagined the explosion of language acceptability that we have today. We see everything from Ruby, Erlang, Scala and other languages widely used -- a great thing that you can use the best language for software development instead of being limited to the "safe enterprise" choices of Java or Microsoft .NET.
But, as the linked post points out, all those arguments fall by the wayside when the alternative is an in-house language used by no other production project on Earth. Developing your own language is not really "conservative" at all.
But the in-house language was basically a variant of Visual Basic. They basically created their own DSL because they didn't feel like any other tool satisfied their needs. Joel's not an idiot -- I'm sure he weighed the cost of doing his own language vs. the alternatives and found that the direction he took was the better way for the time.
It would have been more of a mistake if Fogbugz turned out to be a load of crap, but it's pretty good. I prefer it to JIRA.
>Well, 2006 was a -long- time ago, and Ruby and Ruby on Rails have come a long way.
Right. I recently listened to a lot of the archive of the StackOverflow podcast. Joel & Jeff periodically discuss Rails, and Joel eventually recognizes Rails as a viable option once it improves. This blog post must have been before that.
Even at the time, Atwood (who obviously can hardly be described as a toady for Rails -- at least not until his more recent projects) suggested his rationale for rejecting Rails was pretty questionable. And his rejection of .NET was also pretty zany.
One of the best parts is that he pretty obviously has Spolsky dead to rights here, but typically DHH is thought of as an asshole and Spolsky as the software nice guy. But it's possible to spread FUD with nice-sounding words.
http://david.heinemeierhansson.com/arc/2006_09.html
Scroll down to the part "Fear, Uncertain, and Doubt by Joel Spolsky" from September 01 (permalink 404s)