On the road to Damascus... GPL to BSD

In response to a Twitter comment I made about my change in attitude towards the BSD and GPL open source licenses, Martin Daly asks:

Blog post on the Damascene (or otherwise) conversion from GPL-ish to BSD-ish please. In your own time.

The reason I was even mentioning the license issue is because I was watching a panel discussion at Where 2.0, in which Steve Coast, in defending the GPL-ish license used for OpenStreetMap, said (paraphrase) that the rationale was keeping a third party from taking the open data, closing it, and working from there. And that rationale pretty closely mirrors how we were thinking when choosing the GPL license for PostGIS, back in 2001.

I have experienced no spectral presence or flashing lights.

However, over the years, I have spent far too much time talking to various corporate folks about how the PostGIS GPL license wouldn’t affect their plans to use PostGIS as a database component in their systems. For everyone who came and asked, I am sure many didn’t. But in general the license was an impediment to some organizations engaging with the project. So there was a downside to the GPL. And was there an upside? Did the license protect us from privatized forks?

Well, yes, insofar is as it legally prevented them from happening. But it is worth questioning the implicit assumption that such forks are actually harmful to the project.

And here my experience watching the MapServer community (working with the BSD MIT license) was useful. Because over those same years, I watched the MapServer project chug healthily along, even as some third parties did in fact take the code and work on closed forks from time to time.

The lesson I have taken from my observation is that the strength of open source projects resides not in the licenses or the code, but in the communities arrayed about them. Copying the code of MapServer and doing your own thing with it does not stop MapServer developers from working, and in the long run you’ll probably be better off working within the community than outside it, to gain from the efforts of others. Trying to maintain a private parallel branch and patch in the changes from the open development will quickly become more effort than it is worth. At the same time, because the BSD license is so permissive, there are no legal impediments for companies to engage with the community.

Look at any healthy open source project and ask yourself: what is more valuable, the code or the community? You could take all the code away from a healthy project, and it would start right up again from scratch and probably do a better job the second time. The value is in the human relationships and the aggregation of cooperating talent.

The GPL tends to keep corporate actors out of the community (for good reasons or bad, that’s a separate discussion, but it does). That slows down the development of the project by reducing size of the development pool. Which, counterintuitively, makes forking or ignoring the open project more attractive. Because a slow moving project is easy to beat. A fast moving project is risky to fork, because it is likely that your (relatively understaffed) fork will be left behind by the open project.

So, back to what originally made me delve into license wankery, the question of OpenStreetMap license: the value of the OSM community and philosophy and infrastructure is way higher than the data at this point. And bringing corporate actors into the OSM community would only increase the relative advantage of the open project over any equivalent closed effort.

But changing licenses is a brutally hard thing to do, and it gets harder the longer a license is in place. PostGIS will never change licenses, for example. There are too many developers who contributed in the past under the GPL, and changing license would require the assent of all of them.

But if I ever start a new project, it will definitely be under the BSD.

FOSS4G 2011 Decision Process

The 2011 FOSS4G siting process has begun. After the 2010 process, where there were too many good proposals and not enough slots, we decided to do a couple things to lower the effort level required by bidders in aggregate.

First, we wanted to be more definitive about our regional siting plan, which is to rotate between Europe (2010), North America (2011) and elsewhere (with elsewhere being Asia (2012) in the current rotation plan). For 2011 the regional arrow points to North America, since 2010 is in Europe.

Second, we wanted to lower the bar to submitting a proposal. In 2010 we got four full proposals, and could choose only one. So 75% of the bidders did a large amount of preliminary research and spade-work for no purpose. For the 2011 process, we will start with a “letter of intent” phase, a short document outlining expressions of interest. Hopefully having all the interested parties publicly declared will promote locations coming together and forming joint bids.

The OSGeo conference committee will vote on the letters and only the top two will be asked to prepare full bid documents. That would still leave one potentially disappointed party at the end of the process, but is a big improvement over leaving three, as happened in the 2008 and 2010 processes. And hopefully the letters phase will encourage groups to consolidate bids so we won’t have to even run the final round.

The winning site will still have to submit a proposal and budget, so the Board and Conference Committee have confidence that the local team has thought things through, but the kind of research necessary to prepare a full proposal is a necessary precursor to putting on a conference, so the effort will be put to good purpose.

I’m looking forward to seeing the letters!

How to get Your bug fixed

Mike Leahy is providing a textbook demonstration of bug-dogging on the PostGIS users list this week, and anyone interested in learning how to interact with a development community to get something done would do well to study it.

Some key points:

He does as much of the work in diagnosing the problem as possible, including combing through Google for references, cutting down the test data as small as possible, trying to find smaller cases that exercise the issue.

He responds very quickly (you’ll have to read the timestamps) to questions and suggestions for gathering more information. Since the problem appears initially to manifest only on his machine, any delay on his part risks disengaging the folks helping him.

He prepares a sample database and query to allow the development team to easily replicate the situation on their machines.

And he also gets lucky. The problem is replicable, and the discussion catches the attention of Greg Stark, who recalls and digs up some changes Tom Lane made to PostgreSQL which in turn leads me to find the one-argument-change that can remove the problem. Very lucky, really, it’s unlikely I would have been able to debug it by brute force.

NoNoSQL

I’m really looking forward to seeing the “NoSQL” buzzword head over the top of the hype-cycle and start heading downwards, since I think it’s really doing damage in what should be a fairly straightforward discussion of matching customer use cases to appropriate technology.

NoNoSQLBecause the term is framed (“No!”) in opposition to the almost the entire family of existing data persistence technology, anyone who comes to the discussion fresh assumes there’s a replacement process going on, wherein NoSQL stands in opposition to SQL.

That’s a shame, because the term is only one letter away from a (slightly) less polarizing buzzword: NotSQL. Even then, though, the core discussion would be lost, because the big difference isn’t programmatic API versus 4GL. The big difference is use case matching, in particular the high-volume, high-availability use case which has emerged in the age of consumer web services.

People with public-facing web applications face a potentially unconstrained read/write load (in their happiest dreams) and the techniques necessary to scale a traditional RDBMS to match that load proceed from the straightforward at the low end to the increasingly byzantine at the high end.

The scaling story for traditional RDBMS technology just is not great: start by adding servers and extra technology to hook them together; get increasingly smart people to handle your increasingly complex infrastructure; finally, start hacking at your data model to allow even further partitioning and duplication.

The new breed of databases have a great scaling story: once you get set up, scaling requires plugging in new nodes and turning them on. That’s it. No model changes, no extra replication and high availability technology.

There’s no free lunch though. In exchange for the high-throughput/high-availability you lose the expressiveness and power of SQL. Henceforth you will write your joins and summaries yourself, at the application level. Henceforth performing an ad hoc query may require you to build and populate a whole new “table” (the terminology is highly variable at this point) in your model. And of course this technology is all pretty fresh meat, so just learning enough to get started can be a bit of a slog – kids aren’t exactly coming out of school with a course in this stuff under their belt.

So, it’s a whole new world, and if you are planning on serving an application where the numbers (hits, pages, requests, whatever) are heading into the 7-digits, it might be good idea to start with this technology (can you tell I can’t stand to use the “NoSQL” term? It is just too awful, there really needs to be a non-pejorative term for this application category).

Of course, there’s nothing new under the sun. Getting the best performance for a specialized use case requires (a) modifying your data model and (b) using technology that can leverage your specialized data model. This is exactly what OLAP databases have been doing for a generation to provide data analysis on multi-billion record historical databases (special data model, special technology).

Database guru Michael Stonebraker wrote a nice article about the brave new world of databases, called “One Size Fits All: An Idea Whose Time has Come and Gone” in 2005, and the conclusion is that we are going to see increasing fragmentation of database technology based on use case. “NoSQL” (shudder) is just the latest iteration in this process.

Meanwhile, I’ll put my oar in for the general purpose database: it’s easy to run OLAP queries on a general purpose database, you just can’t do it once your table size gets over a billion; it’s easy to run a public web site on a general purpose database, you just can’t do it once your load gets over a million. On the other hand it’s well nigh impossible to run even a small web site on an OLAP database and pretty darn hard to build even a small OLAP system on a NoSQL foundation.

Horses for courses folks, horses for courses.

For more of this kind of geekery, see the several articles linked off of “The Case for the Bit Bucket” at the Oracle Nerd blog.

Where 2.0 Drinking Game

It’s that time of year again, and I’ll be sitting in the audience with my flask (I hope you will too!) playing the Where 2.0 drinking game. Here’s some of my phrases, what are yours?

  • … find a Starbucks…
  • … we’re releasing an API …
  • … friends list …

Also, take a big slug if someone talks about working with geospatial data more complex than a lat/lon point!