Thanks, t0m. The discussion on irc helped me clarify some things and
figured out which way I wanted to pursue things. Comments inline.
On 06/18/2010 05:15 AM, Tomas Doran wrote:
On 17 Jun 2010, at 20:22, Sir Robert Burbridge wrote:
My questions are:
* Would this kind of functionality be better done in
Catalyst::Helper::(Model|View|Contoller):: or somewhere else?
No, you want to put it in about the right place.
Ok, great.
Currently the skeleton code generation only generates app templates
and basic Model/View/Controller classes, as nobody has made the code
generation flexible enough to generate other things.
Patches to do so would be welcome, however it generally really is just
a couple of lines of code, so the effort to work on it hasn't
scratched anyone's itch really.
I'm going to mull over the skeleton code generation matter and see if it
makes sense for me to work on that. If so, I'm happy to pass back some
patches.
Your talk of 'help enforce the practice of' makes me think that you're
trying to solve a political / development organisation problem with
code - which is rarely the right solution. I personally think you'd be
better writing documentation / providing training to your team about
why to do things a certain way, and/or instituting a code quality
assurance process (such as code review).
One of the interesting things about some of the stuff I'm working on is
that there's an influence from both the Ruby and PHP cultures. I have
had to work on several php projects and it is the ugliest stuff I've
ever had the displeasure to work with (maybe it was just those
projects...). Ruby culture is pretty weird; they act like new money,
with lots of ferraris and beach houses, but not a lot of substance (in
fairness, that might be the rails culture, I'm not sure). Either way,
RoR feels like a california investment banker in a convertable.
In my experience, the PHP culture just plugs stuff in randomly until
something works, then seals it up and builds on top of it. No real
consultation of documentation, etc. RoR culture seems to need vods and
helpers like crazy. Getting them to adopt Catalyst probably requires
some "meeting them halfway" (in my projects, anyway).
Both of these approaches are more likely to succeed in stopping people
from generating shitty code than mandating a recipe with which you
build applications by rote.
I'm actually OK with them generating bad code as long as it's well
contained. I'm not trying to optimize for great code now, I'm trying to
aim for the best use of time *over the next couple of years*. By
creating generating stubs, I can help us create clean bundles of messy
code that are relatively easy to straighten out as time goes on and we
all get better at the craft (refactoring a terrible sub that has
standardized input/output is way better than refactoring a module that
has no discernable API!).
Cheers
t0m
Thanks,
-Sir
_______________________________________________
List: [email protected]
Listinfo: http://lists.scsys.co.uk/cgi-bin/mailman/listinfo/catalyst
Searchable archive: http://www.mail-archive.com/[email protected]/
Dev site: http://dev.catalyst.perl.org/
