It's fascinating, because writing your own custom language is one of those things that everyone knows is a bad idea.
It's one of those things that people do because it seems like the path of least resistance (and in the short run, is), but it inevitably snowballs into a pit of technical debt. Spolsky knew this quite well (he'd written eloquently on the subject).
...and yet he still did it. His defence was that it was the easiest option in the short term, and he was probably right, but it doesn't matter. People only do stupid stuff that seems smart; saying "this stupid thing seems smart!" is only a defence if you have no idea that it's actually fundamentally stupid. Of all the people in the world, Spolsky is one of the least able to mount this defence.
Contemporaneously with his decision to go all in on Wasabi, he wrote a scathing condemnation of Ruby for being slow, unserious, obscure; he suggested that a serious company shouldn't opt for Ruby because it was risky, and that choosing it would put you at risk of getting fired.
Was he right? In 2006, maybe? I mean, he turned out to be wrong, but I don't think it was entirely obvious that Ruby was a serious choice 15 years ago. Of course, he wasn't writing 15 years ago, but even nine years ago, a very conservative, safe approach to choosing a technical stack very possibly did mitigate against selecting Ruby, for all the reasons he outlined. But those arguments applied twice as hard to Wasabi. You don't get to argue that there "just isn't a lot of experience in the world building big mission critical web systems in Ruby" (and hence you shouldn't use Ruby), and then turn around and use Wasabi for your big mission critical web system.
Of all the people in the world, Spolsky probably had the best understanding of why Wasabi was a stupid, short sighted decision. He did it anyway. And it was stupid and short sighted. Rarely is someone so right and so wrong about the same thing at once.
(And yes, Fogcreek is still around, and so is FogBugz. But I don't buy for a moment that Wasabi was actually a good choice. They survived it, but they didn't benefit from it.)
Edit: Spolsky has written too much about why things like writing something like Wasabi is terrible idea to link it all. Besides, a lot of it has been linked in other comments. But I don't think I can express strongly enough that my anti-Wasabi position is simply repeating the things the guy that signed off on developing it and using in production wrote. ...then he decided to write a new language because apparently Ruby was too slow to possibly use to generate a graph, and there was literally no alternative to using Ruby for graph generating than writing your own compile-to-VBScript/PHP language. Words fail.
It's one of those things that people do because it seems like the path of least resistance (and in the short run, is), but it inevitably snowballs into a pit of technical debt. Spolsky knew this quite well (he'd written eloquently on the subject).
...and yet he still did it. His defence was that it was the easiest option in the short term, and he was probably right, but it doesn't matter. People only do stupid stuff that seems smart; saying "this stupid thing seems smart!" is only a defence if you have no idea that it's actually fundamentally stupid. Of all the people in the world, Spolsky is one of the least able to mount this defence.
Contemporaneously with his decision to go all in on Wasabi, he wrote a scathing condemnation of Ruby for being slow, unserious, obscure; he suggested that a serious company shouldn't opt for Ruby because it was risky, and that choosing it would put you at risk of getting fired.
Was he right? In 2006, maybe? I mean, he turned out to be wrong, but I don't think it was entirely obvious that Ruby was a serious choice 15 years ago. Of course, he wasn't writing 15 years ago, but even nine years ago, a very conservative, safe approach to choosing a technical stack very possibly did mitigate against selecting Ruby, for all the reasons he outlined. But those arguments applied twice as hard to Wasabi. You don't get to argue that there "just isn't a lot of experience in the world building big mission critical web systems in Ruby" (and hence you shouldn't use Ruby), and then turn around and use Wasabi for your big mission critical web system.
Of all the people in the world, Spolsky probably had the best understanding of why Wasabi was a stupid, short sighted decision. He did it anyway. And it was stupid and short sighted. Rarely is someone so right and so wrong about the same thing at once.
(And yes, Fogcreek is still around, and so is FogBugz. But I don't buy for a moment that Wasabi was actually a good choice. They survived it, but they didn't benefit from it.)
Edit: Spolsky has written too much about why things like writing something like Wasabi is terrible idea to link it all. Besides, a lot of it has been linked in other comments. But I don't think I can express strongly enough that my anti-Wasabi position is simply repeating the things the guy that signed off on developing it and using in production wrote. ...then he decided to write a new language because apparently Ruby was too slow to possibly use to generate a graph, and there was literally no alternative to using Ruby for graph generating than writing your own compile-to-VBScript/PHP language. Words fail.