A roadmap for 3.9

Tim Rowledge tim at sumeru.stanford.edu
Mon Dec 13 01:23:57 UTC 2004


Michael Latta <lattam at mac.com> wrote:

> 
> On Dec 12, 2004, at 2:08 PM, Stéphane Rollandin wrote:
> 
> >
> >> - my often repeated design philosophy
> >> Perfection is achieved not when there is nothing more to be added
> >> but when there is nothing left to take away
> >> Antoine de Saint-Exupery
> >
> > come on, this is NOT a design philosophy :)
> 
> Sure it is.  If you add the implied "While meeting the requirements of 
> the system".
Exactly - though you can get tied into a bit of a tautology without
careful thought.

Making things as simple as possible and no simpler is a very powerful
approach to making powerful systems. Lisp & Smalltalk share that basic
approach at their core. Squeak has drifted a bit into bloatland.

A large part of the problem here is lack of quality time. Pretty much
everyone is doing Squeak as a part time/fun activity and thus the less
fascinating tasks get less attention. Wonderland gets a bit adrift
because people are busy using it instead of paying rapt attention to
every little message on this list. People offer fixes to bugs and
don't have enough time to have become expert enough to realise that it
will alter the behaviour of some other object they've never even heard
of. Someone works hard on a capability that others have requested
urgently and then never hears a thing in response. It's a bugger.

Another part of the problem is that Squeak has become a very big system
with poorly defined boundaries, negligable documentation, lots of left
over unfinished experiments, and tools that haven't evolved to work with
many new facilities (example1 - exceptions are hard to track down and
relate to places where they are raised. example2 - named primitives
can't be caught by 'senders of').   


tim
--
Tim Rowledge, tim at sumeru.stanford.edu, http://sumeru.stanford.edu/tim
Strange OpCodes: SHUTDOWN: (See EFB)



More information about the Squeak-dev mailing list