MS-RFC-35 - Solving the WMS STYLES parameter issue

[Daniel Morissette]

Kralidis,Tom [Burlington] wrote:
> 
> As a WMS client, not supporting STYLES= is just as severe as not passing
> LAYERS, or REQUEST, etc. from a specification implementation point of
> view.  Will the time come where we are lax on those as well?
> 

I disagree.

STYLES has a stated default of <empty value> in the spec itself, OTOH, 
LAYERS and REQUEST do not have any meaningful defaults so they are 
clearly required and I don't see any good reason to make them optional.

I am in favor of this change to make STYLES= optional by default since 
STYLES being mandatory was in my opinion a mistake in the spec. This has 
also been recognized on the WMS-dev list and a change proposal has been 
submitted to the WMS.RWG to make it optional in future revisions.

I am not in favor of being relax on lots of stuff related to specs, 
hence this note that I included twice in the RFC:

"Developers should keep in mind that adding more exceptions in the 
permissive mode should not be taken lightly as this encourages misuse of 
the specs and bloats the code with unnecessary exceptions."

However in this specific case I think its for the best. Making STYLES 
optional is not going to hurt interoperability in any way now or in the 
future, it's actually going to help with clients who didn't bother 
implementing a unused parameter, that's why I'm cool with this change. 
Yes, those clients are wrong, but they are out there and being used, and 
not dealing with this fact would just mean unnecessary frustration for 
users.

Daniel
-- 
Daniel Morissette
http://www.mapgears.com/