exporting layers to OGC interfaces

[Stephen Woodbridge]

Well, likewise we would expect you to upgrade and not test that things 
worked as you expected :)

5.0 is a compatibility breaking release I thought. So this is a release 
that we should fix things that ARE, but shouldn't be. I think Daniel's 
point was that if one were building a mapfile and one did not INTEND to 
support OGC services that one might not think to see if it is disabled 
by default. I also think commercial services need to worry about people 
figuring out that a site is mapserver driven and the smart people might 
figure out how to use mapserver in ways the owners did not intend and 
steal services that might impact services to paying clients.

I do not think we should break compatibility just for the sake of 
breaking it, but I do think we should do whatever is needed to clean 
things up, rationalize APIs and interfaces and position mapserver for 
the next 5 years.

I have been on the don't break things bandwagon for sometime, but there 
are a lot of things that I would like to see cleaned up in mapscript and 
mapfiles and if a some of these make me rework my code to get to a 
better, faster, cleaner mapserver then there is an incentive for me 
rework things.

-Steve

Paul Ramsey wrote:
> But it is what we are used to. Changing behavior is worse. I mean, I 
> don't write a WMS map file and then not test it to see what capabilities 
> it is advertising. So the fact that it is advertising all my layers is 
> hardly a surprise.  On the other hand, "upgrading" my mapserver and 
> finding that all my layers mysteriously "disappear" from my service -- 
> *that* would be a surprise.
> 
> P
> 
> Daniel Morissette wrote:
>> Paul Ramsey wrote:
>>> In order to retain current behavior, the absence of the keyword 
>>> should be equivalent to "all", and there should be a "none".
>>>
>>
>> I really meant to suggest that we break compatibility because the 
>> current behavior is opening up everyone's data by default, which I 
>> find bad.
>>
>> Daniel
> 
>