[gdal-dev] TMS driver -- Google SoC Proposal

[Tamas Szekeres]

Hi Keo,

I think the idea behind this proposal is fairly compelling. I have the
following questions/comments according to the implementation details.


1. I wonder if we could represent the tilesets as separate overviews
by the driver. I would make the behaviour of the driver quite similar
to the others from this aspect. The reader should finally provide the
images for arbitrary sub-areas but it should use the most appropriate
"overview" to produce the result. But the GDAL API user could also
extract the data of a certain overview if he want that.

2. We should not rely on any particular drivers when creating or
reading the tiles. It would be reasonable to support any of the
current GDAL drivers. The main benefit of this was to be able to store
and use potentially large static images efficiently even if the data
representation of the bands are not restricted necessarily to the
"displayable" types.

3. I would also happy to see at least a ready to use kml structure
that can optionally be generated by the driver to make the results
ready to consume by GE for example.

4. A separate HTTP read driver would also be valuable for a large
number of reasons, it would be great to implement that in this scope.
But I think it shouldn't be bound strictly to this driver. It could be
used in the same way as the png and the jpeg drivers are used.


Best regards,

Tamas




2008/3/27, keo at keo.cz <keo at keo.cz>:
> Hi,
>
>  I am a student interested in working on TMS tiled maps driver for GDAL as
>  a part of the Google SoC. As suggested by Frank, I created a new page on
>  the wiki (http://trac.osgeo.org/gdal/wiki/SoCTMSDriver) with the
>  proposal and extended the existing SoCProjects page with a link.
>  Comments are welcomed.
>
>  Vaclav "Keo" Klusak
>
>
>
>
>  _______________________________________________
>  gdal-dev mailing list
>  gdal-dev at lists.osgeo.org
>  http://lists.osgeo.org/mailman/listinfo/gdal-dev
>