Beneficial Proprietary Extensions of Open Source Software?

One of the arguments for choosing Apache/BSD/MIT software licenses over Free Software licenses like the GPL is that the former harness the profit motive of individuals and companies for the benefit of the open source users. There is hypothetical positive feedback: Apache/BSD/MIT licenses allow proprietary extensions, which in turn lift up the open source source software by giving back bug fixes and feature enhancements.

Does this actually happen? Is the give-back significant, or does value mostly work its way up from open source community contributions? How many fixes and features come from work on proprietary, pay sites? How many come from proprietary, commercial software products?

I've made proprietary MapServer-based sites -- sites available to paying users, but no downloadable code or configuration -- and when they required enhancements or fixes of MapServer, I made the improvements and then gave them back (with consent of my customers) to the MapServer community. However, most of my MapServer contributions were made through my work on community, for-the-public software. The same goes for MapServer in general: most recent work on MapServer was (i'm digging up the stats on lines of code) done to implement OGC standards (WMS, WFS, WCS, SLD) for public-facing Canadian government web sites.

Do any of my readers have examples -- I'm specifically curious about any involving MapServer and GDAL/OGR -- of proprietary sites or software that have extended and improved open source GIS software?

Comments

Re: Beneficial Proprietary Extensions of Open Source Software?

Author: hobu

http://www.gdal.org/credits.html provides a significant list of direct contribution in the form of in-kind development, contract funding, and general sponsorship. http://www.gdal.org/index.html has a good list at the bottom of examples where GDAL/OGR has been extended in both proprietary and free ways.

Re: Beneficial Proprietary Extensions of Open Source Software?

Author: Sean

Thanks, hobu. GDAL's history supports the hypothesis. Is there any more?

Re: Beneficial Proprietary Extensions of Open Source Software?

Author: Paul Ramsey

The last couple of revisions of GEOS have been partly funded by Safe Software and Autodesk, who both use it in their proprietary products. http://svn.refractions.net/geos/trunk/AUTHORS

Re: Beneficial Proprietary Extensions of Open Source Software?

Author: Chris 'Xenon' Hanson

3D Nature's software utilizes IJG's libjpeg, libpng, zlib, libtiff, libgeotiff, fips123, the USGS's GCTPC2 projection library and the OpenSceneGraph toolkit (which itself uses msot of the above and even more, like FreeType, ungif, shapelib, etc). Whenever we find a bug that we can solve, we contribute fixes back to the project. I think commercial use of open source libraries definitely has a beneficial effect on their quality. Commercial users often have a more compelling "itch" to "scratch" and are may be more motivated to get things fixed than someone who isn't working under the gun of a client or manager.

Re: Beneficial Proprietary Extensions of Open Source Software?

Author: Sean

Chris, I'm looking for examples that are more specific, like the ones from GDAL and GEOS.

Re: Beneficial Proprietary Extensions of Open Source Software?

Author: FrankW

As mentioned earlier, at least half of the improvements to GDAL were supported by proprietary software vendors. For MapServer the one item I did that comes to mind was OUTPUTFORMAT capacity which was done for i-cubed, a for-pay services company. Also, a few improvements (eg. raster resampling kernels) were for Tydac AG in switzerland who sells a proprietary product layered on MapServer. I will conceed when I dig through many of my other improvements to MapServer, they were related to contracts that were ultimately government or research oriented. I suppose the work done by Jani from SRC would definately qualify as a proprietary software company folding improvements back into open source. Another non-mapserver, non-gdal example is that the addition of BigTIFF support to libtiff (a substantial project) is being funded by four proprietary software vendors. I think the impact of propietary software vendors is in large part related to the nature of the product. If it is a component that is useful to proprietary software vendors, and the license is not a barrier, it is not uncommon to get their support.

Re: Beneficial Proprietary Extensions of Open Source Software?

Author: Robert Osfield

As the project lead of the OpensSceneGraph and proprietor of OpenSceneGraph Professional Services that provides services on top the project, I can first hand qualify the proprietary use of open source is major driver for success of open source projects. Without proprietary software companies using the OpenSceneGraph I would have no customers, with no customer you have no income, with no income you can't commit fully to a project. A number of small consultancies have sprung up around the OpenSceneGraph Project so making a living from open source development is perfectly viable, albeit low key. Proprietary software companies also contribute in the form of testing, bug fixing, developing new modules and contributing them back to the project. These contributions are kept open source and further benefit the project and the users. Extensions of open source that is kept proprietary, such as plugins and NodeKits will typically have little impact on the open source projects themselves. Lots of users of the OpenSceneGraph extend it for there own applications, and this is a perfectly fine way to work. There needs to be distinction made between licenses so, the LGPL allows such proprietary extensions, GPL doesn't. Robert Osfield http://www.openscenegraph.com

Simple Answers to Simple Questions

There is a question in All Points Blog this morning:

Slashdot reports that in preparation for the Google flyover of Sydney, a number of dotcoms spent hours making huge signs that would be visible from the air.

Is this spam?

No. Google's stunt was a clear invitation for exactly these kinds of smaller publicity stunts.

Comments

Re: Simple Answers to Simple Questions

Author: James Fee

True and KML is prefect for sharing these ads with others. I can't wait.

Re: Simple Answers to Simple Questions

Author: Sean

Agreed. It's not exactly the killer Geo-Web I've been looking forward to.

Maps for Google Book Search

My boss just pointed out to me that Google's book search has new Gutenkarte-like maps at the end of the page. Oops, sorry about page 326. Good thing this book wasn't destructively digitized as in Vernor Vinge's Rainbow's End.

Camp 5 and BBQ Sprint

Chris Calloway and TZPUG are ready to accept registration for Camp 5 (March 10-17, 2007 in Chapel Hill, NC). Philipp von Weitershausen is a great instructor; don't miss this chance to learn about Zope3 and Five from one of the community's most knowledgeable developers.

FOSS4G 2007

Big website update:

The annual Free and Open Source Software for Geospatial (FOSS4G) conference brings together the people who create, use, and support open spatial software. No other event brings together members of the open source development, open data creation, and open standards promotion communities like FOSS4G -- it is truly a "meeting of the tribes".

I am usually critical of OSGeo, but cannot deny that every statement on that welcome page is true.

http://sgillies.net/images/kirok.jpg

[Rumors that Kirok will deliver the keynote to the tribes are untrue.]

The line about meeting of tribes is groan-inducing -- and you won't actually get to witness the GeoServer developers and GRASS developers ceremonially brew coffee around a hooka -- but there really is a pervasive, fun community spirit at these conferences. If you're looking to learn more about the universe of open source GIS and open data initiatives, or have something that you're pushing forward, you should go to Victoria this Fall for the total immersion experience.

Comments

Re: FOSS4G 2007

Author: Chris Holmes

Just for you Sean I'm going to bring a hooka and coordinate a coffee brewing ceremony with Markus.

Re: FOSS4G 2007

Author: ROb McCulley

I'm hoping to attend. We generally have a 'no-out-of-province-travel' rule, but Victoria is probably as close it's going to get anytime soon (I'm in Alberta). Hopefully ...

New Bathymetry of the Great Salt Lake

http://sgillies.net/images/bonneville_sm.png

GeoCarta's Roger Hart just reminded me of something I miss: the Great Salt Lake. Follow the links through his blog to the USGS's new bathymetric maps of my favorite little inland sea. Back in the day, I fished for perch and sailed in Gunnison Bay. Evaporation from the lake contributes to the annual 500" of snow dumped on Alta. Sea Monkies (Artemia franciscana) frolic in the smelly, briny compartment south of the railroad causeway. The modern lake is a remnant of ancient Lake Bonneville, which suddenly poured into the Snake River about 14,000 years ago. Estimates of the flow during the flood event are as high as 33 million cubic feet per second.

The image to the left of ancient and modern lake levels is from an online class at the University of Utah.