[E-Lang] Re: E questions

[Mark S. Miller]

At 07:27 PM Friday 7/6/01, Ken Kahn wrote:
>Visual Basic has literally millions of programmers. There are probably
>plenty of others: PL/I, Postscript, Zeta Lisp,  Qunitus and Sicstus Prolog,
>Cobol, Smalltalk, Eiffel, and probably lots more that spread while being
>non-open. And maybe someday even ToonTalk ;-)
>
>Maybe they should open source their language but it is hard to make that
>argument based upon history.

Zeta Lisp, Quintus & Sicstus Prolog, non-Squeak-Smalltalks, and Eiffel are 
examples of the kind of non-success that would argue against your point.

The nature of ToonTalk's appeal is so different than the kind of appeal all 
these other languages have, that its success or non-success, whether you 
open source it or not, should probably not be taken as corroborating either 
side of this question.  You've created a beast of a truly unique flavor -- 
an exceedingly rare event.

Visual Basic, PL/1, Cobol, and Postscript support your point.  Of these, 
only VB is reasonably recent.  Since VB became prominent, several new 
languages with open source implementations have become prominent.  Besides 
Java, what other big successes have their been for proprietary languages 
since, let's say, 1995?

I think the climate is different now than it was in the Cobol to Postscript 
era.

Further, VB and PL/1 were backed by the industrial giants of their day.  
(I'd guess the same for Cobol, but it was before my time.)  Postscript's 
success is peculiar -- it was promoted as a protocol for talking to printers 
and not as a programming language.  Despite universal adoption in this 
narrow niche, it has never spread beyond that, despite brief proprietary 
attempts by Sun and NeXT to promote it as a display/ui extension language.

Looking back an Jonathan Rees' message, I see that I an unwarranted 
assumption.  I assumed that the company he's advising isn't one of these 
industrial giants, though he doesn't say this anywhere in the message.  If 
it is such a giant, then, judging by your examples, it may have an ability 
to make a proprietary success.  Otherwise, I still think *recent* history 
supports the open source argument.


        Cheers,
        --MarkM