begin  quoting "Rodolfo Garc??a Pe??as (kix)" as of Thu, Mar 21, 2013 at 
01:52:50AM +0100:
> On 20/03/2013 22:59, Carlos R. Mafra wrote:
[snip]
> > Massive cleanups in core objects smells like trouble and should have
> > a very compelling reason to be accepted at this (very stable) point.

Stability is good!

> I can continue with my other things! :-)
> 
> Ok, no problem. The current code seems to be stable, but we now that
> some things "could be better". We can stop here for a while and continue
> "testing" this stable code and only solve bugs.

If the only problem with the code now is the underlying design, then
rewriting the code won't necessarily solve the problem; it might be
better to describe the current design, and how it is deficient, and
discuss that... and work out a path from here to somewhere better.

> But the problem is that (probably) we are building over code with some
> problems. I am not talking about code with bugs, is about bad design.
> This design make difficult to do some things (like xrandr, yes) and
> other things. Yes the things works, but that doesn't mean that they are
> fine.

I like the metaphor of "technical debt". Deficient design borrows
against the future, and it must eventually be paid off. It's easier
to make small payments than large payments, as large payments are
painful, and might contain hidden debt all their own.

This is also where automated unit tests and refactoring techniques can
help.

[snip]
> Finally, if the code is "frozen for a while" with changes about design,
> could be a good idea check the documentation, the web and create a BTS
> in windowmaker.org. I cc John for the BTS.

Mmmm.... documentation.

-S.


-- 
To unsubscribe, send mail to [email protected].

Reply via email to