[ANSI-Smalltalk] Consense driven standard...
dr at littlesmalltalk.org
dr at littlesmalltalk.org
Thu Dec 6 10:33:30 GMT 2007
Hi all,
First let me say that I am really happy to see that the ANSI process has
good chances to become an ongoing effort now.
As far as I understand from the recent discussions you plan to have a
community (users and vendors) driven standardization process.
As a result the group will produce one document or a set of documents that
reflect the consenses that have been found on all detail issues for both
the Smalltalk core and extension modules.
Of course in general I like this method.
But please have Michi Henning's article about the rise and fall of CORBA
in mind where he says that such processes also are dangerous.
(http://acmqueue.com/modules.php?name=Content&pa=showpage&pid=396)
So although the group should really try to find consense we should not
repeat too many of the OMG's failures that caused CORBA to be a complex
beast.
I think it would be good if there wouldn't be too many voting members in
the project commitee to reduce problems there...
So maybe five or six people would be good.
I would also suggest that for all proposals of 'extension modules
standards' a reference implementation (as a proof of concept) and example
of their usage should be required before the commitee finally votes on the
proposal.
What do you think about standardization of Middleware modules?
(Ok, the CORBA standard specifies a Smalltalk language binding. But how
about general XML or webservices modules or so?)
And finally (ok, nobody besides me wants a standard for this :)) how about
a standardized GUI toolkit API?
Just some thoughts...
Best Regards,
Danny Reinhold
More information about the ANSI-Smalltalk
mailing list