On 19 avr. 05, at 20:06, Dean_Swan at Mitel.COM wrote:

>
> Doug, Andreas, Stephane,
>
> Thanks for the explanation.  I certainly agree that the string 
> refactoring is an important thing to do, and sooner is better.  I have 
> yet to see an existing codebase get migrated from 8 bit characters to 
> wide characters that didn't invovle some (or a lot of) pain.   Squeak 
> seems to be making the transition "relatively" smoothly.
>
> I also agree that fewer releases less often is probably appropriate.  

I think that 2 per year is the max we can afford.

> What I'm used to seeing is "the date is set, so we'll just pull 
> features and keep the date if things aren't ready", which often leads 
> to the next release starting before the current version is released, 
> multiple parallel streams, merging hell, etc.  ...and then an 
> important customer need some feature we don't have, so we delay 
> everything and squeeze it in to get the sale....  (I am speaking 
> hypothetically here, of course.  I don't mean to imply that this 
> actually occurs in any real software development organization ;-)  ) 
>  I would hate to see Squeak get sucked into that vortex, and it seems 
> that you are all aware of the dangers.

yes this is why we "voted" to delay 3.8 to get it more stable first 
since we will carry it around for at least two years....

Stef

>
>                 Thanks.
>                 -Dean