[mapserver-dev] vector symbols

Tamas Szekeres szekerest at gmail.com
Thu Feb 10 06:38:24 EST 2011


2011/2/10 thomas bonfort <thomas.bonfort at gmail.com>

> I know Tamas extensively uses the GD way of defining
> symbols, by creating vector symbols that are basically a bitmap whose
> color can be changed by style color. It works and allows to create
> repeating symbols that fit nicely together, but fails whenever the
> user applies a SIZE to the symbol. I consider this "abusing" the
> vector symbol meaning by relying on GD quirks, but understand that
> this is a behavior that we might want to keep. Thoughts?
>
>
Thomas,

I have been using this way to implement the mapinfo brush patterns, but I'm
not too satisfied with the current results since this can only work with the
GD renderer properly. The other renderers provide significantly different
look because of the difference in the rendering approach you mentioned.
With regards to the brush patterns I would be in favour of creating a new
symbol type (something like a bitmap pattern mask) where the layout of the
pixels could be defined, but the actual color would be set to the style
color and style background color.


> I'm not sure what we can do to have backwards compatibility with GD
> vector symbols and at the same time have a consistent cross-renderer
> output. In the current state, this seems impossible to me. I see two
> ways of moving forwards:
>  * remove backwards compatibility with GD vector symbols, and treat
> vector points as paths
>

I'm not sure what we would like to get out of the GD renderer in the current
state. It only supports PC256 images which may rarely be appropriate for an
average user. In this regard removing the compatibility would not be a big
problem in MS 6.0 in my opinion.


>  * define a new symbol who's points are treated as a path, and
> interpret vector symbol points as pixels.For this new symbol, I would
> definitely recommend that the points be generated on a grid where
> (0,0) represents the anchor point of the symbol.
>
>
That sounds interesting to me. What did you mean by "interpret vector symbol
points as pixels". Would this be a renderer independent way of preparing a
bitmap from the symbol and have the renderers to use this bitmap in the
rendering operations?

Best regards,

Tamas
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.osgeo.org/pipermail/mapserver-dev/attachments/20110210/5544111d/attachment.html


More information about the mapserver-dev mailing list