Best practices for API wrapper development?

[Dave Hodgkinson]

Wow. That has WAY more than I want! But in a way, it's quite similar.

One of the things I'm trying to avoid is building a class for each kind of thingy
we get back, which is what Net::Amazon does horribly and unmaintainably. Yahoo::Search
appears to sidestep this with a single ::Result. I like this.

On 8 May 2012, at 17:03, Alex Balhatchet wrote:

> Way back in '06 when I first wrote WebService::Nestoria::Search I was
> encouraged to copy the style of Yahoo::Search - given that it was
> written by Jeffrey Friedl and is now maintained by Tim Bunce I think
> that counts as "on the shoulders of giants" too :-)
> 
> The style is very Do It All Yourself but it's simple enough to
> maintain given small/incremental changes in the API.
> 
> - Alex
> 
> On Tue, May 8, 2012 at 4:52 PM, Dave Hodgkinson <davehodg at gmail.com> wrote:
>> 
>> I hated the Amazon one, it was overkill. Probabl my usual approach of
>> flinging hashes around is suboptimal. Schwern's position on the shoulders
>> of giants by using DBIC for a relatively simple interface might be too heavy.
>> 
>> Examples you like?
>> 
>>