The enterprisey one managed to linkbait me this morning, in his "why Smalltalk doesn't get used" post. First, I'll note that Cincom Smalltalk's user base is up, and our profits are up accordingly - so there must be some happy users out there :) But look at his rationales:

• Microsoft didn't invent Smalltalk, IBM doesn't push it and Oracle doesn't support it and therefore Enterprise Architects can't incorporate it into their power vendor strategy

Hmm. So that would explain the rise in interest in Ruby how, exactly? Never mind that Smalltalk connects quite well to Oracle - the "database of record for our internal version control system is Oracle, and it's the most widely used DB by our customers.

• Smalltalk is simply too hold. [sic] Enterprises sometimes choose the latest and greatest because it helps motivate their staff and not because it is the best tool for the given problem. This is why I believe Ruby on Rails will succeed.

Yeah, C and C++ have just withered out on the vine, haven't they? You know, I suspect that he wouldn't like this argument if someone used it as an explanation for outsourcing, or failing to hire older workers.

• Smalltalk is slow compared to other first-class languages such as C++ and Java. Developer productivity is great in Smalltalk but developer productivity should always be a second-class concern over runtime performance

If you measure trivial things, sure. If you write a big app in C++, it often ends up slower, because the developers end up writing a GC system (this is one of the big wins Java gave C++ developers - same syntax, but with modern facilities). Performance hasn't been an issue for anyone using it in a long while - and if it is, then Ruby sure isn't an answer (and yet somehow, even though Ruby is very slow, that productivity thing seems to matter.). Go figure.

• Smalltalk is not open. Have you ever heard of any of the Cincom bloggers championing Smalltalk implementations should all be open source?

We don't live under an OSS license, but all the code is available (the Smalltalk level code is all available, for both commercial and non-commercial, and the VM source is delivered to all customers). How's that source access to Oracle working out? Have you stopped using that, or SQL Server? We aren't under OSS for a simple reason: there's no business model I've seen that would allow us to profitably deliver the product that way.

• The vendors in the Smalltalk community are greedy and for the most part filled with employees that are idiots. Have you ever witnessed coopetition in the Smalltalk community amongst its major vendors to build marketshare? Maybe they could learn something from the Java community in this regard

The Java community? There's a vendor there who jealously guards the code to prevent any possibility of code forking. Way back in the day, ParcPlace freely licensed Smalltalk for $1. That's actual open-ness, as opposed to what passes for it now. Oh - and the link I didn't reproduce when he called someone an idiot above? That guy isn't a Cincomer, and has never worked for Cincom.

I'd go on, but it gets tiresome. If McGovern doesn't want to use Smalltalk, that's fine - plenty of other people do, and it's a growing number all the time. Based on my reading of his blog, he's stuck in an unproductive atmosphere, in a shop that can barely deliver working code. I'll make this point: once upon a time, back in the early 90's, that shop was productive - they were using Smalltalk. That stopped, and it looks like the productivity stopped with it.