On Tue, Oct 8, 2013 at 8:44 AM, Andrea Aime <andrea.a...@geo-solutions.it>wrote:
> Hi,
> discussing the CSS integration of the randomized fills in the geoscript
> community
> some feedback came up about the names and meaning of the properties.
>
> Since the current property names were originally discussed here, I guess
> it's just
> fair to share the feedback and give people in the GeoTools community a
> chance
> to comment.
>
> The full discussion is here:
> https://groups.google.com/forum/#!topic/geoscript/QphAa53CYfY
>
> While I cannot receive the feedback fully, some parts would require a
> partial rewrite
> of the current code, there are some cosmetic bits that I can still feed
> into the current
> code without having to rewrite much code.
>
> One is about having the random property be a boolean flag, and then having
> a separate
> property to control whether the randomization is gridded or free.
> I believe we could receive this feedback by having the random property not
> being a boolean,
> but have the following values:
> - none (default): no random distribution, the graphic is repeated as usual
> in the fill
> - grid: randomizing in the tile around a fixed grid
> - free: freeform randomization, truly random distribution inside the tile
>
> This would play better also with other possible randomization algorithms,
> like the poisson
> one that was cited in geoserver-devel
>
> David was also questioning the usage of the "random" keyword, suggesting
> fill-jitter might
> be a better option.
> Personally I don't like it, because I cannot associate jitter with the
> random distribution of
> symbols out of the box, but maybe it's because I'm not a native enligsh
> speaker.
>
It is more important to me to have "fill" in the name than to avoid calling
it "random." This is a design consideration specific to CSS because the
option will not be nested inside a <PolygonSymbolizer> element but jumbled
in with all other properties. If we keep the name 'random' then I can just
use the '-gt-fill' prefix in CSS as I do '-gt-label' for many
text-symbolizer-related vendor options.
Also both David and Tim did not like having rotation being a separate
> boolean parameter.
> A possibility here could be to have it controlled as second parameter to
> random:
>
> random: grid false (gridded, no random rotation)
> random: free true (freeform, random rotation)
>
> Personally I don't like this approach because it's not clear what "true"
> or "false" mean there,
> whilst random-rotation: true is verbose, but clear.
>
I guess it sounded kind of confusing that I proposed a boolean parameter in
the middle of my spiel against boolean parameters. My intent was that the
presence or absence of the keyword 'rotation' would control whether the
rotation was also randomized. Your examples then become:
random: grid rotation;
random: grid;
> Please comment.
> Any improvement that does not require a significant rewrite of the current
> code is welcomed
> (for the others.. it would have been better to comment on the random fill
> when I asked
> for feedback before writing the code... but let's hear it anyways, if
> there is a lot of consesus
> I guess I can do the rewrite in my spare time instead of reviewing pull
> requests
> or fixing bugs, although it would set a bad precedent for whoever is
> trying to do paid
> work in this community)
>
It's not my intent to prevent folks from getting paid to do work on
GeoTools - I am all in favor of it. As I said on the GeoScript list, if we
are past the design phase of this GeoTools feature then I will work with
that.
------------------------------------------------------------------------------
October Webinars: Code for Performance
Free Intel webinars can help you accelerate application performance.
Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from
the latest Intel processors and coprocessors. See abstracts and register >
http://pubads.g.doubleclick.net/gampad/clk?id=60134071&iu=/4140/ostg.clktrk
_______________________________________________
GeoTools-Devel mailing list
GeoTools-Devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geotools-devel