Hi Marcus and all!

Marcus Denker <denker at iam.unibe.ch> wrote:
> Am 17.10.2004 um 22:41 schrieb goran.krampe at bluefish.se:
> > Hi all!
> >
> > "Andreas Raab" <andreas.raab at gmx.de> wrote:
> >>> If they can't bother doing that then I don't want their code in 
> >>> Squeak,
> >>> because it means they don't care about other people reading their 
> >>> code.
> >>> And I don't want Squeak to depend on developers with such a mindset.
> >>
> >> Strong words. What about your own code?
> >
> > Yes, strong words I agree. And perhaps these postings were not in the
> > best constructive sense - and I REALLY did not want to offend Yoshiki 
> > or
> > anyone else. In fact - I was more aiming at the harvesters with this,
> > not the original authors. In this case I think Yoshiki and/or the other
> > people working with m17n are simply so swamped with work that they
> > forgot to add these comments.
> >
> 
> I don't think that it would have been practical to look for 
> classcomment when
> we did the m17n update: We had to add around 160 changesets, and we were
> happy after that was done. It took around 4 hours (or even longer, 
> don't remember),
> and we had only that day. I don't think we could have done any better.

No, I agree - it sounds like a rough day. :) As I have written - my
posting was probably an overreaction.
But fact remains - if we let in new classes without class comments, then
I think we are moving in the wrong direction.

> All in all I have, for some people, gotten in a 'trusted mode": If 
> there is a changeset
> from someone I trust (e.g. Ned, Diego, Andreas and so on...) , it gets 
> added if there
> are no conflicts. Just like of they had CVS access.

I sympathize with that. Since I also think the harvesting model is
too... "slow".

But given this fact - that we now are running in a different model -
then IFF we want to make sure new classes do have class comments, we
will have to try to make that happen by some other means.

> I don't think that we should be too restricting when harvesting: We 
> tried that, and it failed
> completely. It did not improve the quality of the submissions, it just 
> did make sure that nothing
> happened.

I agree, but I also think that demanding class comments for new classes
is not really "too restricting". It is easy for the author to add some
basic class comments, it doesn't take much time.

But anyway, I am tired of arguing for this and have now officially given
up on my crusade for getting Squeak to have class comments and will
hereby never again bring up the subject.

>      Marcus

regards, Göran