Use the right tool for the job. This is one of the guiding principles of good craftsmanship, and it would be a fine slogan for an Open Source and Proprietary Geospatial Foundation.

GIS, as practiced, is more craft than science, and requires both a diverse tool kit and a pragmatic sensibility. Almost every open source-using practitioner, myself included, applies a combination of open source and proprietary tools to their problems. However, the pragmatism that gets us through our daily tasks is not an appropriate ideal for the entity that calls itself the geospatial open source leader. OSGeo should be single-mindedly promoting an end-to-end open source GIS software stack. Interoperability with proprietary software is a means to this end, but is not, by itself, a great enough goal for OSGeo.

The Wine Blog Atlas looks promising. I'd like to see KML as well.

I've been trying to ignore this nagging pain, but it just won't go away. The Open Source Geospatial Foundation will not assert technical superiority of open source software, and worse, makes statements like this in its FAQ:

The foundation respects the important role that proprietary software plays in our industry, and is not trying to get rid of it, or the companies that produce it.

Better to be like the ASF and simply make no statement at all about proprietary software. The one above is an embarrassment.

However, the foundation takes the position that free and open source software can and should play an important role in the geospatial industry. Furthermore, having quality open source alternatives to proprietary software can be good for the end user, the industry, and even the proprietary software vendors. In fact, most proprietary geospatial software is built on open source software to some extent.

Having alternatives can be good? Of course, but the thing that motivates me -- and I'm not alone, here -- are superior alternatives. The open source and free software movements are winning. We're producing the best web server, the best web browser, the best operating system. GDAL, PostGIS, and MapServer (to name a few) are displacing proprietary software in shops on their technical merits. Come on, OSGeo milquetoasts, stand up for open source. If you need help, I'll loan you my Troy McClure self-help video collection.

Comments

Re: Get Confident, Stupid!

Author: Allan

I think you're right. OSGeo needs to be confident. The recent email thread about a potential workshop at FOSS4G by Oracle is a case in point. There seemed to be a considerable breadth of opinion in the response.

Re: Get Confident, Stupid!

Author: Matt Perry

The blanket statement that all open-source is superior to proprietary is a bit far fetched. I think OSGEO's statement strikes a perfect tone. Open source and proprietary feed off of each other and the competition is healthy for all. The key is to highlight areas where open source is ahead of the pack. My totally off-the-cuff analysis of the current playing field: In some realms open source GIS is technically superior (web mapping), in some it is at least as capable but more difficult to learn and deploy (geostatistics, analysis) and in others proprietary software is leading the way (Desktop). For myself, its just about picking the right tool for the job. If a piece of software lets me get my job done better and faster, then I use it. And if the goal is to increase adoption of open source, playing the superiority card won't get you very far with the pointy-haired boss types :-)

Re: Get Confident, Stupid!

Author: Sean

Matt, you just pulled that blanket statement out of thin air.

Be Confident

Author: Matt Giger

Open source is great, and indeed superior in many areas. However commercial software is important and often have the best products out there since they can devote millions or billions of dollars into the making of a platform. I don't know where OSGeo is coming from, but I do think their statement is reasonable. Open Software provides a baseline of functionality that the commercial world must meet or beat in order to survive (theoretically).

Re: Get Confident, Stupid!

Author: Andy

I like Matt's analysis. He's being nice about geostatistics, analysis though. Web mapping hands down Open Source wins the day with MapServer. Nothing comes even remotely close to Mapserver. Desktop hands down ArcGIS wins by a land slide even over other proprietary solutions. ArcGIS may suck but it sucks less than anything else by a long ways. As far as superior OSs go from an end user stand point Open Source has a long ways to go before it even comes close to Windows. Just in Fonts alone Windows totally spanks the competition from and end user stand point. Every time I fire up an Open Source OS I just sigh and realize how far Open Source OSs still have to go before non-super users are ever going to take a look at them. It's getting there but even Novell's stuff isn't close yet, and I love my Suse but I am also realistic. I deal with end users when I deploy software I have written and 99% of them aren't even remotely ready for a non-Windows or non-Mac solution. PostGIS is a better database than Oracle Spatial. PostgreSQL is just plain the best database out there period in my opinion but it has no good management UI and the largest enterprise GIS system (SDE) doesn't work with it. That is fixing to change hopefully very soon and once you can use PostGIS from ArcCatalog I think we will see a big shift away from Oracle. All that said I agree they shouldn't open their mouths about proprietary software.We should be concentrating on writing better software and promoting it on it's merits and sticking up for solutions we believe in. Let their spin doctors do whatever they want but our spin doctors should at least be sticking up for the software we write so long as it is the technically superior alternative. If we aren't the technically superior alternative then we need to shut up and code better software until we are.

Re: Get Confident, Stupid!

Author: Matt Perry

Sean, Yeah I see your point. You didn't say ALL open source software was superior but you did claim: "The Open Source Geospatial Foundation will not assert technical superiority of open source software" and "The open source and free software movements are winning." .. both of which indicate you feel that open source software is superior to proprietary across the board. Sorry if I put words in your mouth.

Re: Get Confident, Stupid!

Author: Frank Warmerdam

Thanks for your analysis of my modest FAQ write-ups. They are an attempt to avoid unnecessary controversy and to be a statement that is broadly acceptable to those involved in OSGeo. I don't see any compelling reason to change them. Of course, I'm also a big fan of "good enough" instead of always aiming for technical superiority. No doubt a character flaw.

Re: Get Confident, Stupid!

Author: Sean

Frank, thanks for the comment. I'm not calling on you to give any of your friends and customers in the proprietary software business the finger. Just eliminate the proprietary bias in the FAQ, and show some pride and confidence in open source. Like GeoServer does.

OpenLayers 2.3 is out (since 21 February, in fact). For details, see the release notes. Christopher Schmidt told me that MetaCarta is bringing a community developer in for a vector sprint (previously). I'm eager to see the results. Pleiades has a milestone this spring which will produce a release of PleiadesOpenLayers compatible with OpenLayers 2.3 or 2.4. Currently, POL distributes a patched OL 2.1.

Update: more from Christopher Schmidt.

The MapServer 5.0 plan has some compelling items (feature-level transparency, AGG rendering in particular), but the one thing I'd really like to see is separation of MapServer's cartographic and web application aspects. These are entangled both in the code and in configuration, so it's no small task, but one that would reward the community.

Frankly, making effective maps and administering a web server are skills that rarely go together. Separating cartography from the web application allows changes to be made to the cartographic presentation with greatly reduced risk of breaking the application that serves up the map. The separation would also make it easier to test and deploy maps, even across platforms.

Similarly, separating these concerns would make development of MapServer a bit more sane. Cartographic enhancements would be much less likely to break the web application, and the upcoming changes to accommodate WFS 1.1 and WMS 1.3 (for example) would be much less likely to break the cartographic engine.

Last, but not least: I consider the built-in MapServer web application (mapserv, or mapserv.exe) to be cumbersome, and a bad investment today. The W*S protocols are better implemented in your favorite web programming language. PHP, Python, Ruby, .NET -- all surpass C for working with HTTP requests, strings, and XML. All have better templating. All have better memory management (except perhaps PHP). Back when I was playing ultimate frisbee, we used to heckle our teammates, yelling "Drop the piano!" at the tired guy who couldn't run down the disc. MapServer's built-in, written in C, web application is that piano.

Google Maps is being used to make the case for REST in the Web and Web Services world.

The Cascadia Users of Geospatial Open Source. I'm not sure if they are hewing to the official Cascadia boundaries or not.

The provocative post by Nat Torkington about pseudo-open source companies at OSCON makes me wonder how many of these will be showing up at FOSS4G 2007, and if anybody will even object as Nat does.

James, are you looking into Sam Ruby's new Planet branch, Venus? You can see its features, including a meme tracker, at Planet Intertwingly. Pretty sweet, eh? Mail me for Python help.

Meanwhile, my increasingly awesome feed (brand-new Atom categories, complete with schemes, and more liberal license starting today) is being massacred at the old Planet Geospatial.

Smells like the 3DConnexion SpaceNavigator is the Ferrari Vista laptop of geo-blogging. More of this to come, I'm sure.