[Catalyst] YAML vs. *

[Matt S Trout]

Andrew Ford wrote:
> Matt S Trout wrote:
>> Dean Troyer wrote:
>>   
>>> On 6/9/06, Matt S Trout <dbix-class at trout.me.uk> wrote:
>>>     
>>>> A few (IMO) better options -
>>>>
>>>> Config::General -
>>>>
>>>> <View TT>
>>>>    INCLUDE_PATH __path_to('root')__
>>>> </View>
>>>>       
>>> Too Apache-like...I've never cared for it because it can't make up
>>> it's mind what kind of format it really is.
>>>     
>> Yes, it can. It's a kind of format designed to be reasonable for people to 
>> write and machines to read.
>>
>> Sure, it loses points for architectural purity. But architectural purity gets 
>> you YAML.
>>   
> I think a major problems is that often configuration file contents are 
> not designed in anyway; they just evolve as a serialization of the 
> information that the programmer happens to need to initialize the 
> application (and often this will not be documented) and with no thought 
> for the people who will have to deal with them.  For example there will 
> probably be a dsn lurking in there somewhere, which will mean nothing to 
> a lot of people, rather than separate database name, username and 
> password that might make some sense to them (database type and such like 
> can probably be assumed or defaulted and the application could build up 
> the dsn).  So some poor support guy ends up getting told to edit a file 
> that is undocumented, obtuse and has a fragile format, and then things 
> break.

Absolutely. Forcing developers to think about this sort of thing early on is 
something we end up doing a lot as part of consultancy gigs.

-- 
      Matt S Trout       Offering custom development, consultancy and support
   Technical Director    contracts for Catalyst, DBIx::Class and BAST. Contact
Shadowcat Systems Ltd.  mst (at) shadowcatsystems.co.uk for more information

+ Help us build a better perl ORM: http://dbix-class.shadowcatsystems.co.uk/ +