yaml vs perl complexity ( was Re: [Catalyst] JSON instead of YAML? )

Brandon Black blblack at gmail.com
Wed Jan 25 20:42:38 CET 2006


On 1/25/06, Corey <corey_s at qwest.net> wrote:
> On Wednesday January 25 2006 11:48 am, A. Pagaltzis wrote:
> > * Bill Moseley <moseley at hank.org> [2006-01-25 17:40]:
> > >On Wed, Jan 25, 2006 at 10:41:10AM -0500, Daniel Westermann-Clark wrote:
> > >> [10:38:00 dwc at fortuna ~]$ cat refs.yml
> > >> ref: &r
> > >>   key1: value1
> > >>   key2: value2
> > >> deref: *r
> > >
> > >At least it isn't complex like writing perl. ;)
> >
> > Yeah, look at how simple it is. An end user will certainly have
> > fewer problems with that than with Perl.
> >
>
>
> Well I don't see any control/flow structures/logic in YAML, and you can't
> define subroutines or call external modules, etc. etc.
>
> If there were a dialect of perl that did nothing more than set variables, then
> sure.
>

We'd still need a parser/validator/importer/exporter/etc for this
subdialect, since we wouldn't want to eval it directly if we're trying
to restrict what people can do with it.  Which puts us back in the
boat of writing Yet Another YAML-Like Module (Catalyst::YAYAMLLM
anyone? :) ).

YAML was a great solution for what Cat needed, it's just the Spiffy thing...

I say branch a copy of a good version of YAML, apply any relevant
fixes that have occured during Spiffy-time that apply to old YAML
also, and call it Catalyst::YAML, for now - roll it into the Cat
distribution instead of publishing it seperately.  I know - the
comparisons to the CDBI frozen thing....

Or we go with JSON, or something like Config::Scoped.  Whatever :)

-- Brandon



More information about the Catalyst mailing list