A roadmap for 3.9
stéphane ducasse
ducasse at iam.unibe.ch
Sun Dec 12 13:16:50 UTC 2004
> It is important that you understand this issue. When I looked at 3.8
> for porting Tweak to it I was seriously appalled by many of changes
> that have crept into the system. Proliferation of base classes has
> become a significant problem over the last versions and it won't get
> any better with more extensions to the base image.
You think about ProjectRefactoring I guess or another of such a beast.
But nobody gets impacted by this one....except when
one want to understand project. Same for BookMorph and related
SqueakPage.
Or just SystemNavigation, Beeper (I know this one was bad), but
seriously SystemNavigation is needed, Namespace and SmalltalkImage =
SystemDictionary are needed too.
> Something that -in lack of a better word- I will call "random
> reclassification" of methods (quite honestly I do not see any
> difference in putting methods into SystemDictionary vs. SmalltalkImage
> except that a) the name SmalltalkImage sucks and b) you now have to
> guess which place to look at) does not help to improve the situation.
Just another feedback on this one: I think that you are quite insulting
but this is not even the problem
because I do not really care. Now what we proposed is a fix to the
situation ie merging SmallltalkImage and SystemDictionary but
decoupling namespace from SystemDictionary. But you do not have any
constructivism on that
so we can let the system like that (this will not really get seen if
you consider the amount of half done or half broken code in the system
nobody will notice it), and work on our solution for our own brand.
Now again if our cool community or its main actors could agree on
important steps to be performed:
- we could do that with a better result
- less frustration
- more feedback
- less time spend
But we prefer to talk over emails only when it hurts. Because this is
what I see, you got hurts by our changes
and I'm **really** sorry about that. We are humans too and do make
mistake but if we do not have discussions
and feedback and we wait the last moment then of course we will get
damaged and frustrated.
I think that more important about this exchange is the syndrome that
there is no communication, concertation,
in this community. Of course you said in the past that it was a not so
good idea to split SmalltalkImage further so we listened and stopped
there, but how can we clean this part of the system? Do we all want to
live with that? What are the pros and cons? We thought a lot about that
(so of course may be we minored the impact even if we have all the
deprecated methods so normally your code should work and if you jump
from 3.6 to 3.8 then you have to load the deprecated methods of 3.7.)
and for us having a good and slick system is important because it will
enable new creation.
When we are doing our clean we always have in mind the sentence of Dan
Smalltalk a system that a single person can understand...note that
understandibility is not linear with the number of classes because 10
clean classes are easier to understand than one giant.
But the solution is not in shouting after (ok this is the first level
and we will learn from that) but be present and participate at the
beginning and this is what we are trying to do with 3.9. So let us face
reality how can we make sure that:
- we have a plan for 3.9
- build a process for the future
- we repair the problems of SystemDictionary once for all (because
this is not fun nor bold to be a human garbage collector). We can
build a task force but if you do not even provide arguments so what?
- still clean this part of the system (note that the ReleaseBuilder
class is based on Mike advice and class like that make a lot of sense)
And the real solution is in a better communication and process. Look
at what you are saying to us after all the efforts we spent and all the
emails sent around because the worse is that everything was public. I
think that I'm much more worried about this absence of decisional
process than just your words (I hope that you feel better now). But
this email discussion is clearly showing that there is a problem in the
community.
Stef
More information about the Squeak-dev
mailing list
|