[ODE] code quality again
mark at image-engine.com
Wed Jul 11 13:08:13 MST 2007
Jon Watte (ODE) wrote:
> There are two problems facing ODE, or any fundamentally hard open source
> 1) Solid physics simulation is actually a hard problem that requires
> significant rigor. Such rigor is actually not commonly found among open
> source developers or hobbyists.
I'd agree, but numerical rigor for physical simulation is a different
problem to that which is ultimately quite straightforward - such as
using a style similar to the surrounding code when applying patches as
opposed to just mixing and matching with your own style, not leaving
single statements and/or huge blocks of code commented out without any
explanation as to why, being careful to free dynamically allocated
memory through all code paths, etc. I think some relatively minor
improvements could make a huge difference to ODE's codebase, and may
encourage others to get stuck in.
> 2) Everyone working with ODE want to "make a game" or "finish their
> thesis" -- they don't want to "make ODE better." Thus, there's a real
> shortage of people's time to actually improve it substantially in the
> ways that you suggest.
> Yourself, you have the same attitude: it's not what I want, and I'm not
> going to spend the time making it so.
To be fair, the author of the original post did say:
"A couple of times i sat down and wanted to fix things, but resigned quickly because the code was too hard to read and
Which sounds to me like their attitude is correct, but the realities of
the situation are a serious obstacle. I also remember somebody else a
few months ago volunteering to make a start on some code quality
>> I think it is a pitty because of this i actually consider using another
>> library. Which is not worth mentioning as my project is too small, but
>> it MAY reflect other possibile users/contributers opinion as well. So
>> PLEASE increase code quality before patches are merged.
What if patches were reviewed in some way before inclusion in the
repository? If going back and 'restyling' much of ODE is too
inconvenient for reasons already discussed on the previous thread might
it be worthwhile just to make sure things don't get any worse?
Also it seems like this mailing list is the primary means of
communication regarding all aspects of ODE - what if greater use were
made of the bug/feature/patch trackers in SourceForge?
More information about the ODE