Class comments!?

goran.krampe at bluefish.se goran.krampe at bluefish.se
Mon Oct 18 19:26:53 UTC 2004


Hi all!

Just can't stop replying. It is like an itch that simply don't want to
go away. Aargh! :)
Ok, final posting. If you really, really want to discuss this further
with me and thus removing my short Squeak time from implementing
dependencies in SM :) - then email me in private. :)

lex at cc.gatech.edu wrote:
> First a few observations:
> 
> 1. Class comments are good in general.  80 classes with not one single
> class comment is rather extremely uncommented.
> 
> 2. On the other hand, I would be happy for Squeak to have the
> multinationization code in it even without comments.  In fact, I'd feel
> that way about just about functionality you can name.

Well, I understand the philosophy and even sympathize to some extent -
but it can easily be translated into "We don't care at all about the
quality of the code that we put into the image, as long as it does
something cool.". And I don't agree with that.

We all know where such a philosophy will take us.

> 3. It is a shame to think about *anything* that will slow harvesting
> down even further.  Please let us think carefully before adding any more
> barriers; instead, file it as a wishlist bug on mantis and we can
> revisit it if and when harvesting becomes so fast that we have time to
> impose a higher burden on that process.

Hmmm. Ok, so why don't we just stop reviewing, stop testing and stop
urging people to write unit tests and stop all those procedures? Just
throw it all in! Yiha! :)

(irony dripping here, but I am not trying to increase the tension of the
thread - in fact, I have detached emotionally a while back)

> 4. As well, it's a shame to put pressure on people doing good work, even
> without comments.  (And I do think there is such a thing.  :)) 

Yeah, well - you have all been so "kind" to point this out even though I
was pretty clear in my posting that I have the utmost respect for
Yoshiki's work (and anyone else involved).

But, sure. I will never do it again. Never. I DO NOT CARE ANYMORE. Ok?
End of story.

> Let us
> also keep in mind that raising the bar means that there are more coders
> who just won't bother to contribute any longer.  This gets back to #2:
> if it came down to a take-it or leave-it situation, how often would we
> really want to leave it?
> 
> Given all of the above, how about the following strategy: as long as a
> change moves Squeak forward, let's put it in, but add a bug report that
> it needs comments?  Uncomented but useful code is a step forward (and in
> the case of m18n, a big step!).  Adding comments is another step.  Why
> should we insist that both steps be taken at once?  The Squeak project
> has many thousands of steps in front of it.

Well, obviously we have a different view in all this. Let's imagine that
Squeak is a glass of muddy water.
The muddiness is the quality level. Morphic is muddy. Other parts of
Squeak are quite clear, but a lot of other parts are also less so. Now,
how do we move to a clear glass of water?

My idea is that when we pour more water into the glass we make sure the
new water is clearer than the current clearness of the glass - so that
the net effect is that the glass turns clearer. :)

Your idea seems to be that we don't need to care about that - just pour
it in, it doesn't matter if it is pitch black of dirt. We can get around
trying to filter it or something "tomorrow". Now, the problem is -
tomorrow never comes.

> Lex

End of thread for me.

regards, Göran



More information about the Squeak-dev mailing list