Call for comments -RFC 39

Yewondwossen Assefa yassefa at DMSOLUTIONS.CA
Thu Nov 15 12:11:44 EST 2007


Stephen Woodbridge wrote:
> Yewondwossen Assefa wrote:
>> Hi all,
>>  I have put together an RFC outlining changes that could be done to be 
>> able to support Named styles in WMS/SLD. I would appreciate your 
>> comments on it:
>> Thanks
> Assefa,
> Thank you for a well written RFC. I have a couple of questions/comments.
> When I think of "NamedStyles", I think of a section of the mapfile that 
> just has a bunch of STYLE blocks, that can be reused anywhere by 
> referencing there names. This make the STYLE object reusable via the 
> name reference. Also redefining a STYLE only needs to be done in one 
> place not N places, which greatly simplifies the mapfile maintenance.
> The advantage of this is that it modularizes the mapfile and compresses 
> it greatly. It would be great to define a 13/11 red/white casement road 
> STYLE and then reference it multiple times in my mapfile via its name 
> for example.
> Your proposal instead moves us in the opposite direction of adding more 
> text tags to the layers and making them more verbose and less 
> comprehensible.
> Was thought given to anything similar to my description about? What 
> other implementations were considered and rejected? Why? if any.

> Would the above idea be compatible with WMS/SLD issues you are trying to 
> solve or could it be adapted to this problem?

  I agree with you (and Bart who made the same remarks), that have a 
representation table (RST) at some higher level and being able to access 
the "styles" by name in the layer would be a great addition. We could 
have something similar to what we do with symbol files and attach the 
RST to a map file.  I used the term NamedStyles in the RFC since that is 
what is used in the SLD specifications and I understand that outside the 
SLD world it's interpretation could lead to a confusion.

  But I think, I am not addressing the RST support in this RFC, at least 
not intentionally :)
  I also think that even if we had this RST concept in MapServer, I 
would still need to find a way to define/publish  several mutually 
exclusive styles on a layer. The mechanism proposed I think could well 
be applied even if  Mapserver implements an RST

> What are the levels of effort of these two implementations?

  I did not estimate really what It would take to implement and RST in 
Mapserver although this idea has been floating around for a long time. 
It would be interesting to see in details how this would be implemented. 
I am convinced that it could be a good addition to Mapserver.

> Thanks,
>   -Steve W

Assefa Yewondwossen
Software Analyst

Email: assefa at

Phone: (613) 565-5056 (ext 14)
Fax:   (613) 565-0925

More information about the mapserver-dev mailing list