# from Guy Hulbert
# on Tuesday 18 March 2008 19:21:
>> shorewall-perl might be a place to start; shorewall comes up zero on
>> cpan search, yet is
>> a big perl program; so converting it to an O-O module should be a
>
>Actually it seems already broken into modules:
>http://www.shorewall.net/Shorewall-perl.html
>
>and it would be more polite to check with the author first.
Sure, but don't let unresponsiveness stand in the way. What does the
author lose if you wrote some code before they knew about it?
>> simple matter of
>> time and coffee :)
Well, such a task maybe isn't really suited for SoC. Refactoring and
adding tests is useful, but is a totally different project than
creating something new. Both have merit and both are potentially good
SoC projects, but refactoring something which isn't already on CPAN
where the author isn't a mentor is probably not going to be very
successful (even if it gets done and works great, there's this elephant
in the room about "what should we call it" and "how to ship it" if the
author hasn't been involved.)
When I say "check for prior art", I mean only that we want to see that
the student did some legwork and is aware of potential starting-points
(or design flaws in the prior art.) There is certainly no rule against
reinventing the wheel if you can point to one or two square or
triangular wheels and describe your plan to do it better.
And of course, extensions/plugins to an existing codebase are great if
that is workable because you start closer to your problem domain and
get to spend more time refining your idea instead of debugging your
support code.
--Eric
--
---------------------------------------------------
http://scratchcomputing.com
---------------------------------------------------
