[mapserver-dev] Request for review/comments on RFC 85 (Contour Layer Rendering)

Tue Mar 26 10:46:47 PDT 2013

Hi Brent,

We discussed this important question here with the rest of the team and 
came to the conclusion that the current approach (CONNECTIONTYPE CONTOUR 
+ CONNECTION pointing to raster data file) could be extended some day to 
support a reference to another layer of type RASTER as the CONNECTION 
source (similar to the way a TILEINDEX can refer to a separate layer as 
the tileindex source). Doing this would provide more flexibility but 
would also require lots of work under the hood so that is not likely to 
happen soon, but at least we know that we have a future solution to this 
problem that would be consistent with the current approach in the layer 
definition.

And for the short term, if one has multiple DEM files to use as input 
then they could be indexed using a GDAL VRT. Providing a GDAL VRT as 
CONNECTION input to a contour layer is already supported by the current 
implementation.

Hopefully my explanation makes sense

Daniel

On 13-03-26 10:33 AM, Brent Fraser wrote:
> Steve,
>
>    I'd rather not overload (pollute? erode? highjack?) the meaning of
> existing object keywords. For example, "TYPE CONTOUR" would imply
> contours could only be rendered as lines; the future we might want to
> render them as polygons.
>
>    While not ideal, using the PROCESSING directive to specify the type
> of processing doesn't step on the other keywords.
>
> Best Regards,
> Brent Fraser
>
> On 3/26/2013 8:10 AM, Lime, Steve D (MNIT) wrote:
>> It would be nice to avoid using the processing directive in this way.
>> Why not simply use the layer TYPE or "TYPE CONTOUR".
>>
>> Steve
>>
>> ________________________________________
>> From: mapserver-dev-bounces at lists.osgeo.org
>> [mapserver-dev-bounces at lists.osgeo.org] on behalf of Brent Fraser
>> [bfraser at geoanalytic.com]
>> Sent: Tuesday, March 26, 2013 8:47 AM
>> To: Alan Boudreault
>> Cc: mapserver-dev at lists.osgeo.org
>> Subject: Re: [mapserver-dev] Request for review/comments on RFC 85
>> (Contour Layer Rendering)
>>
>> Instead of
>>       CONNECTIONTYPE CONTOUR
>>
>> how about using
>>       PROCESSING "TYPE=CONTOUR"
>> or
>>       PROCESSING "TYPE=DEM2CONTOUR"
>>
>> to allow the CONNECTIONTYPE to specify a connection to a database, etc
>>
>> Best Regards,
>> Brent Fraser
>>
>> On 3/26/2013 6:21 AM, Alan Boudreault wrote:
>>> Brent, yes, that's what I mean in my first point.. but typo. * the
>>> connectiontype use will be kept*... until we define another way for
>>> hybrid layers.
>>>
>>> Thanks
>>> Alan
>>>
>>> On 13-03-25 10:07 PM, Brent Fraser wrote:
>>>> So we're sticking with the "CONNECTIONTYPE CONTOUR"?
>>>>
>>>> Best Regards,
>>>> Brent Fraser
>>>>
>>>> On 3/25/2013 1:22 PM, Alan Boudreault wrote:
>>>>> Dev,
>>>>>
>>>>> I would like to do the inclusion this week. A few comments about the
>>>>> current implementation:
>>>>>
>>>>> - For now, I think the contourlayer use will be kept, we will need to
>>>>> think more about hybrid layers for mapserver 6.4.
>>>>> - Tileindex are not supported (yet)
>>>>> - Postgis raster are not supported (yet)
>>>>>
>>>>> http://mapserver.org/trunk/development/rfc/ms-rfc-85.html
>>>>>
>>>>> I'll do the call for vote tomorrow morning if there's no objection.
>>>>>
>>>>> Alan
>>>>>
>>>>> On 12-09-21 03:22 AM, Havard Tveite wrote:
>>>>>> I am not sure if contour generation should be a job for Mapserver,
>>>>>> but if we decide to implement it, I think there is a need for
>>>>>> processing directives for "smoothing" / generalisation.
>>>>>> It should be possible to specify the level of "smoothing" for the
>>>>>> DEM as well as for the resulting contours.
>>>>>> Smoothing levels could be specified as the maximum deviation
>>>>>> allowed from the "original".
>>>>>> Contour "smoothing" / generalisation needs to pay attention to
>>>>>> neighbouring contours / topology, so that contours don't cross
>>>>>> after smoothing.
>>>>>>
>>>>>> Possible processing directives for a DEM tolerance of 25 units
>>>>>> (xyz) and contour tolerance of 10 units (xy):
>>>>>> PROCESSING "DEM_TOLERANCE"=25"
>>>>>> PROCESSING "CONTOUR_TOLERANCE=10"
>>>>>> They could be made scale dependent in the way that is suggested
>>>>>> in the RFC.
>>>>>> "TOLERANCE" might not be the best word - "GENERALIZATION"
>>>>>> is also be a possibility.
>>>>>>
>>>>>> It would also be good to choose an approach that will not
>>>>>> make it impossible to use a point layer (xyz or xy+height
>>>>>> attribute) as a data source.
>>>>>>
>>>>>> Håvard
>>>>>>
>>>>>> On 9/20/2012 5:40 PM, Alan Boudreault wrote:
>>>>>>> Hi devs,
>>>>>>>
>>>>>>> I'd like to request some review on a pending RFC regarding support
>>>>>>> for
>>>>>>> contour layer rendering.
>>>>>>>
>>>>>>> Please see
>>>>>>> https://raw.github.com/mapserver/docs/master/en/development/rfc/ms-rfc-85.txt
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> The official rfc page should be updated in about a hour.
>>>>>>>
>>>>>>> http://mapserver.org/trunk/development/rfc/ms-rfc-85.html.
>>>>>>>
>>>>>>> Thanks,
>>>>>>> Alan
>>>>>> _______________________________________________
>>>>>> mapserver-dev mailing list
>>>>>> mapserver-dev at lists.osgeo.org
>>>>>> http://lists.osgeo.org/mailman/listinfo/mapserver-dev
>>>>>
>>>>
>>>
>>
>> _______________________________________________
>> mapserver-dev mailing list
>> mapserver-dev at lists.osgeo.org
>> http://lists.osgeo.org/mailman/listinfo/mapserver-dev
>>
>>
>>
>
>
> _______________________________________________
> mapserver-dev mailing list
> mapserver-dev at lists.osgeo.org
> http://lists.osgeo.org/mailman/listinfo/mapserver-dev

-- 
Daniel Morissette
http://www.mapgears.com/
Provider of Professional MapServer Support since 2000