Jeff Atwood Blog: Has Joel Spolsky Jumped the Shark?

[mouser]

Jeff Atwood's blog today demands return of the old Joel Spolsky:

Has Joel Spolsky Jumped the Shark?
..
two weeks ago we found out that Joel's company wrote their flagship product, FogBugz, in a proprietary language they created themselves.
...
You couldn't possibly have heard it, but that was the sound of fifty thousand programmers' heads simultaneously exploding. Writing your own language is absolutely beyond the pale. It's a toxic decision that is so completely at odds with Joel's previous excellent and sane advice on software development that people literally thought he was joking. He had to write an entire follow-up post to explain that, no, he wasn't kidding.
...
I reject this new, highly illogical Joel Spolsky. I demand the immediate return of the sage, sane, wise Joel Spolsky of years past. But maybe it's like wishing for a long-running television show to return to its previous glories.

http://www.codinghorror.com/blog/archives/000679.html

Personally I have always found Joel Spolsky alternatively insightful and a unbearable self-promoter.

[mouser]

It will be interesting to read how this continues; I suspect Joel knew quite well that his negative comments about Ruby would garner some real attention and require some follow up.

I'm also not sure I agree with all of Atwood's points.. I think one could make an argument on behalf of Joel that goes something like this:  Rather than choose to write the product in a fixed language that would hold us captive to that platform, we built a homebrew language well suited to the product so that we could change the target output when that becomes more advantageous.  I agree with Jeff that this is usually a very bad decision, but there are long-term projects where it might make sense.  It's also not clear to me how windows-centric Joel's product is; if the answer is "very" then it seems to be that the argument against using Ruby makes more sense.

[Carol Haynes]

Entertaining read - but I think I am beginning to agree that Joel rarely says anything interesting these days that isn't deliberatiely provocative or just plain daft

[rkarman]

Writing a compiler is not such a big deal, especially if you get back that you only have to maintain 1 codebase instead of 2 or more.

Then i have to agree that typeless/dynamic typed languages (like ruby) are really bad news for professional enterprise systems (not looking at speed at all here) but dynamically typed languages delay the moment you find bugs, and this delay can be deadly in enterprise software. i have to admit i frowned when i read he was actually saying PHP was a good choice.... or did they strong type it by now?

To me all this guy says makes perfect sense and i suspect everyone going nuts at him are all hobby programmers. which hobby programmer cares if their code fails (on some type error) after if has been tested and is running for a month? not many i guess. personally i always try to proof my code is correct instead of proofing that my code is not correct (that last one is called debugging) and dynamic typing can only proof your code is not correct in some obscure situation when it happens!

[mouser]

well i definitely disagree that writing a compiler isn't a big deal; it's getting easier very year but it is no small matter no matter how you do it.  I should note however that his "language" is not compiled to machine code or even byte code, but instead builds vbscript or another high level stuff, so it's not clear at all what kind of processing it really does and it may turn out to be a glorified preprocessor.

[Carol Haynes]

The main issue (as I see it) is that if he gets run down crossing the road no one will be able to maintain the software !

[rkarman]

yeah, i meant a compiler like Joel Spolsky wrote is not a big deal at all. like compiling some language to another (non machine) language.

i guess writing a genuine compiler could be done in a few months too. but you have to have a lot more knowhow to pull that off. if you have to first go figure out the PE format or the machine operations of the target platform then you are more looking toward a small year worth of work, lol.

[f0dder]

Even "just" compiling one language to another isn't necessarily trivial; C++ used to be a frontend for C, and iirc the comeau compiler (currently the most compliant compiler, and only one that supports some of the _very_ advanced C++ stuff wrt. templates) also depend on another C++ compiler...

Also, google's AJAX toolkit compiles from JAVA to JavaScript.

But anything that outputs to vbscript can hardly be that complicated

[rkarman]

But anything that outputs to vbscript can hardly be that complicated

LOLOL