RFC 36 revised, call for vote (or more comments)...

[Steve Lime]

>>> On 11/30/2007 at 3:10 PM, in message <47507C40.50909 at mapgears.com>, Daniel
Morissette <dmorissette at MAPGEARS.COM> wrote:
> Steve Lime wrote:
>> Fair enough. Perhaps these types of features should be implemented in a 
> sandbox first?
>> 
>> Steve
>> 
> 
> 
> I'm not sure if a sandbox could help in this case. My worries are more 
> long term, once that's released and the users start using the feature in 
> real life apps, or when we try to extend it, we may then discover that 
> we missed something important in our thinking.
> 
> I think it's the bit where we set the template inside the outputformat 
> object that feels a bit odd to me:
> 
> 1- OUTPUTFORMATs are for rendering maps, but templates operate only on 
> query resultsets... so we end up with some OUTPUTFORMATs that work only 
> for queries which feels a bit clunky... or did I miss something?

The way I think about it is that maps are really just queries, a bounding box
query and the rest is presentation. Converting to a raster or converting to
converting to a text representation is simply the next step after identifying
which features to process.
 
> 2- The OUTPUTFORMAT concept to me refers to an actual file format (e.g. 
> GIF, PNG, etc, or even KML, SVG, GeoJSON, etc.) ... however when we 
> assign a template we actually tie the OUTPUTFORMAT to a specific set of 
> layers... this is more like a profile of the format than the format itself.

The tie is indirect though. The OUTPUTFORMAT knows nothing about which 
layers might be referenced in the template, so it feels ok to me.  

> I don't want to be splitting hairs in 4 (do you have that expression in 
> English?), but there is something in this that bugs me and worries me 
> for the long term viability of this feature.

Lemme know when you come up with a problematic use case.

> Actually Assefa's RFC-39 gives me the same feeling...
> 
> Daniel
> 
> 
>>>>> Daniel Morissette <dmorissette at MAPGEARS.COM> 11/23/07 3:47 PM >>>
>> Steve Lime wrote:
>>> Hi all: Back at it. I've updated RFC 36 to incorporate comments from folks 
> and would like to throw it out there again for comments and/or a vote. Please 
> see it at:
>>>
>>>   http://mapserver.gis.umn.edu/development/rfc/ms-rfc-36/ 
>>>
>> 
>> Sorry for the late vote. I really like the proposal but I can't give 
>> this a full +1 because I am not convinced that I understand all the 
>> possible ramifications well enough, so I'll have to go with a +0.
>> 
>> Daniel
>