[OSGeo-Standards] Re: Use of logo for clickable icon

[Jody Garnett]

Hi Ian - That is not too much of a problem for WMS layers; WFS with
TextSymbolizers could be granted some slack. Render the image once
without symbolizers - and once again with and just check that there is
a difference.

I also have been thinking hard on the "phase 2" automation side of
things ... how can we check the output of a client if we are just
provided a full screen snap. And I think I got something tricky.

The initial WMS can generate a grid of squares where we change the R
component for latitude and G component for longitude. Then as we make
the initial WMS Context request we can look in the image for a large
block of the correct color; and based on the color changes we will get
some idea of the resulting area being displayed (regardless of how the
client is constructing their screen - we should be able to respond to
3D clients in this way - although can sometimes be hard to screen
capture). This would at least let us identify the area of the screen
to crop for the rest of the conformance tests.

Jody


On Sat, Jul 18, 2009 at 12:05 AM, Ian Turton<ijturton at gmail.com> wrote:
> On Thu, Jul 16, 2009 at 8:36 PM, Jody Garnett<jody.garnett at gmail.com> wrote:
>> For reference we do have some some rendering tests boiled into our
>> JUnit test cases (comparing rendering against expected output etc). It
>> is not that useful; but it does point to the image comparison
>> techniques being useful.
>
> I wrote some of those tests for geotools (way back) - the problem we
> ran into was that the same server on different machines produced
> different results due to font availability (this was WMS). I can see
> the same issues arising for WFS if two client implementations have
> different fonts etc.
>
> Ian
> --
>
> Ian Turton
> http://www.geotools.org
> http://pennspace.blogspot.com/
>