Curt Sampson wrote:
> On Tue, 25 Sep 2007, Ruby Maniac wrote:
> 
>> Once upon a time, Smalltalk was a commercially successful language....
> 
> Well, that's debatable. It all depends on where you set your bar;
> Smalltalk was never all that widely deployed.

Once upon a time, there were *two* -- or was it three? -- Lisp machines.
They existed ... they had revenues and users and all that. Once upon a
time, Paul Graham wrote an e-commerce application in Lisp that he sold
to Yahoo, who then re-wrote it in C/C++.

Was Lisp a "commercial success?" Hardly.
> 
> If measured by the number of dollars paid to programmers to write code
> over any given period of time, I would not be suprised if at this point
> Ruby is more "commerically successful" than Smalltalk ever was. It's
> important not to forget that a vast amount of programming is done not to
> make products which are sold to others but for internal consumption. In
> fact, this appeared to me to be the majority of Smalltalk use.

I personally think that it is almost always a mistake to do "a vast
amount of programming ... not to make products which are sold to others
but for internal consumption". Programmers should be paid to produce
software that fills profitable markets, not to do infrastructure
projects that could be done using off-the-shelf software.

I personally think if you need software infrastructure, you should
document your requirements and select a product from one of three (or
more) vendors that can meet them. If there *aren't* three vendors that
can meet them, your requirements need to be re-done.