[Etoys] First use of OLPC feedback (bugs, suggestions,
yoshiki at vpri.org
Mon Nov 12 18:41:02 EST 2007
> I would like to comment on that without the intention to be nasty.
> Just sharing up my thoughts
> on that. We spent a lot of time on 3.9 integrating a lot of changes
> (besides traits which could be removed and
> apparently scared a lot of people for no reason).
> Taking 3.8 as a
> basis for OLPC was a clear message that
> all these efforts to improve Squeak were not interesting. So after
> that why would I spend time
> on helping while I already did that and I most of the time got
> criticism and sarcasms from great squeakers.
It wasn't like that at all, Stef. One of the biggest reasons why we
ended up with a 3.8-based image was because the old Squeakland image
was based on 3.8. In the other words, we didn't pick mainstream 3.8
image over 3.9. We picked the Squeakland image as the basis.
And why? Because .pr files created in the Squeakland image are not
compatible with the 3.9 image. The class shape change introduced by
Traits; you cannot load them into 3.9-based image. That was a big
We, the whole Squeak community, weren't too careful about the
implication of Traits and existing contents when Traits was being
integrated to the mainstream image. This was really unfortunate, but
we couldn't simply ignore the existing contents.
And, to be honest, having a real control on the entire system was
important as development has to be done quickly. Just sticking with
an image we understand was another motivation. (At the risk of losing
the "improvements after 3.9, that is true.)
> Of course may be I misunderstood the message behind choosing 3.8 but
> now we get a squeak system that does not
> have all the improvements of 3.9 and a nice fork. But this is ok.
> I'm sad and I'm hesitating to do something to bring all the energy
> together by looking at the changes made in
> OLPC and bringing them to squeak. But I have the impression that
> nobody cares, so may be I will simply do something
> of my free time and energy.
> The underlying message is that: with a bit more communication you and
> the community could have got a much better
> system but it seems that communication costed too much so everybody
> was playing in his corner. This is ok too and I
> can understand that. But then we get what we have.
And, like Bert wrote (quoted below), the primary concern is always
to provide a better Etoys experience.
> > The core team is busy enough
> > with making Etoys work, we can't spend too much time on making the
> > Smalltalk experience better, yet. We try nevertheless - in the latest
> > version we support the view-source key, for example.
More information about the etoys-dev