On Wed, 2010-12-01 at 14:25 -0500, Kohei Yoshida wrote:
> But I've already become immune to this hard-to-figure-out code base, so
> I stopped thinking that way years ago. I've even trained myself to
> ignore those German comments. When I see functions, I see chunks of
> code, not functions. ;-)
The ability to quickly read, and work effectively on a vast, gnarly,
un-documented code-base is an incredibly valuable one - it turns out
most commercial code-bases are a total mess internally; LibreOffice is
nothing special. Furthermore, the code is -always- correct, and comments
bit-rot at an amazing rate.
However; I'm fairly convinced that there is a happy medium somewhere
closer to Lubos than Thorsten but in-between - I don't think that
investing tons of time into blindly documenting the behaviour of every
method is at all necessary; yet clearly having good, descriptive
per-method (and often more usefully per-class) documentation is really
important. IMHO we cannot and should not put a break on writing that
just to improve merge-ability, so I applaud another of Miklos' great
hacks :-)
Finally - wrt. ultimate code quality - I tend to think that growing our
test suite can also enable the re-factoring that can make our APIs both
pretty and intuitive over time, by providing good, clean, and wide
ranging tests that will catch inadvertent changes in side-effects
as/when they are introduced.
Regards,
Michael.
--
michael.me...@novell.com <><, Pseudo Engineer, itinerant idiot
_______________________________________________
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice
