inlcude Matt Perry's demtools

[maphew]

adopt Matt Perry's utilities into GDAL to streamline use and maintenance.

svn source: ​http://perrygeo.googlecode.com/svn/trunk/demtools/

See ​http://lists.osgeo.org/pipermail/gdal-dev/2008-October/018773.html for related thread.

[warmerdam]

I have added Matt to the cc: list.

A few things we need to establish:

1) Is Matt OK with this?

2) If so, is it true that the code is all under the Apache license? Would Matt be willing to relicense it under the GDAL MIT/X license which I believe is similar to the Apache license? This isn't critical but is desirable to keep the GDAL licensing story simple.

3) how will we handle stringtok.h? Would it be acceptable to just incorporate it into each of the programs using it to avoid complexity?

4) Should the names of the utilities be changed? We normally try to avoid very generic names for GDAL utilities that will go into the path. I'd prefer to rename them gdal_aspect, etc.

PS. Should we go ahead with this, I'd also like to propose Matt as a GDAL committer if he is interested so he can maintain the utilities directly in GDAL SVN.

[cbalint]

Replying to warmerdam:

3) how will we handle stringtok.h? Would it be acceptable to just incorporate it into each of the programs using it to avoid complexity?

See patch for #2389, it should get rid of stringtok.h and use internal GDAL API.

[maphew]

I would love to see it happen. Unfortunately I won't have any time to devote it, otherwise I would have done this a long time ago!

You have my full permission to fork and relicense the tools under an X/MIT license - no problem at all.

• Matt Perry

-- ·´¯·.¸. , . .·´¯·.. ><((((º>·.¸¸.·´¯·.¸.·´¯`·...¸><((((º> "The best way to predict the future is to invent it." -- Alan Kay Matthew T. Perry ​http://www.perrygeo.net

On Thu, Apr 23, 2009 at 3:12 PM, Matt Wilkie <maphew@…> wrote:

Hello Matt,

I'd love to see your demtools included in the core GDAL package. Frank Warmerdam is open to the idea but has a couple of questions, namely: are you okay with the idea? and are you amenable to changing the license to match GDAL (X/MIT)? The latter is not a showstopper. (see http://trac.osgeo.org/gdal/ticket/2640).

thanks for considering it!

-- another matt

[hobu]

ok, I have reworked hillshade to use the C API and I have updated it a little bit. If someone could test trunk to make sure we're doing the right thing, it would be greatly appreciated.

Matt Perry, please check the licensing and copyright ​http://svn.osgeo.org/gdal/trunk/gdal/apps/gdalhillshade.cpp

Thanks!

r16829

[maphew]

thanks Hobu; Might want to change "illshade" to "hillshade" :)

[hobu]

Upon reflection, I think we should congregate these utilities into a single utility -- gdaldem (I'm totally flexible on the name -- and have each be a separate operation within that rather than proliferating five more GDAL utilities.

Their basic operation, open a dataset, fetch a band, create a new dataset/band, process is all the same. I think this approach would ease maintenance in such a way that someone unfamiliar with the algorithms would be able to maintain the housekeeping code easily and wouldn't have to do it in five places. Additionally, it would allow compartmentalizing of the algorithms such that they could be put in the library proper if we had the wherewithal to do so.

I also think Matt Perry should be a GDAL committer. I will propose so on -dev now.

[maphew]

I like the idea of gdaldem with commands rather than seperate utilities.

[Even Rouault]

Various fixes and improvements in r16833 and r16834

[Even Rouault]

r16834 : Rename gdalhillshade.cpp as gdaldem.cpp to enable more sub-modes to be added

[Even Rouault]

Grrr. I meant r16837 in my previous comment.

[Even Rouault]

'slope' mode added in r16839

[Even Rouault]

Matt,

I see that the author and copyright holder of color-relief.cpp is Paul Surgeon. I think we also need his explicit authorization to relicence his utility under X/MIT. Would you mind asking him if he agrees, and if so, could he post in that ticket a note similar to yours ? Thanks.

[Even Rouault]

'color-relief' mode added in r16913. This is a clean room implementation, not based on color-relief.cpp.

[warmerdam]

Note, at least parts of Matt's code are apparently directly derived from GRASS GPL code and cannot be included in GDAL due to licensing differences. See #2975 for details.

[Markus Neteler]

For the record, the license issues have been clarified in #2975.