"Gustavo Pistoia" <grpistoia at sinectis.com.ar> wrote:
>   Suppose you have to implement a commercial application. Should you use
> a commercial smalltalk or squeak?
>   I have tried with Cincom Smalltalk ...
 ...
>   What do you think or recommend?

There are a lot of factors and so the decision is not necessarily easy
or obvious.  On the other hand, getting the slightly wrong answer is not
necessarily fatal.  I have been near the verge of doing a deal with
Cincom for perhaps 3 years now.  I think of the advantages but then I
think of the disadvantages and irritations.  I have swung like a
pendulum, but I'm far on the Squeak side at the moment.

I'll propose a "model" for trying to understand the Cincom deal.  I
myself have had difficulty understanding the Cincom deal, partially due
to their lack of spelling it out clearly on their web site, but with
occasional attempts at clarifications in comp.lang.smalltalk, to the
point that I admit I may be incorrectly (but perhaps not unfairly)
characterizing their deal as

      $7,000 per server per year

Yes, they offer a VAR (value added reseller) deal at 6% per year on
something (presumably your gross related to using VW) with a minimum of
$500 per year per developer.  However, I believe you must get permission
from Cincom for *each* project you want to do.  So, if you plan to sell
a "shrink wrapped" product, this might work out well.  What bothers me
is that, if I spend most of my time in VW, I'll wish to use it for
*everything*, but I may well find that a contract programming gig, where
I would like to use it, is one where Cincom requires the $7,000 per
server per year rather than 6% of what I bill my client.  Thus, I may
wind up *having* to use another product for that client.  Well, why not
just use the other product for everything (such as Squeak or IBM
Smalltalk or Lisp) from the beginning?  Then I remove the uncertainty as
to whether I can *use* the product.  Just the thought of having to
consult Cincom before I can discuss a deal with a prospect is enough to
turn me off of VW.  

So, a VW solution might be to say that, ok, I'll accept the 6%/$500 per
year deal on the shrink-wrap product and tell prospective clients that
if they can't both pay me my hourly rate *and* pay Cincom $7,000 per
year per server, that their project simply isn't valuable enough for me
to work on.  There would be advantages and disadvantages to that policy
and I do not know where the balance point is.

Would Squeak be suitable for your project?  My guess is that it would
be.  Let me offer this thought for our amusement:

     The only reason Squeak would be too slow for a product is that it 
     would let you finish the product before the hardware had caught up.


-- Frank