Why Google Really Ban the AGPL

Chris DeBona, Google’s Open Source guru, goes a little limp when confronted with the AGPL. The AGPL license is designed so that ASP’s using AGPL licensed code for web services are required to deliver their code back to the open source community. It was used in relation to e.g. the Google database search ranking algorithms, they would all have to be laid bare. Google disallowing it, DeBona claims, is a procedural matter, because dealing with it would tie up too much time. This is almost certainly true, but omits the most obvious and more important precedent truth. Dealing with it would tie up so much time because Google would have to be so damned sure it gets us nowhere close to being used on anything important within the business and fear of the wrong thing occurring would tie every software project up in so much red-tape the CEO’s sign off would be required for an employee to save a variable to memory on a Casio scientific calculator. Chris has picked out what we might call a “happy co-incident truth” as to why the AGPL is not used. He has picked an aspect of the truth that sounds more open source friendly and avoids highlighting Google are only interested in open source in so far as it helps them strategically rather than as a matter of principle. E.g. their motive is first and foremost to run a business and make profit - big news there.

The sad fact is many have been allowing themselves to be duped by Google’s spin on open source. Yes open sourcing Android et. al. is better than staying closed. However you can deliver as many open source projects as you like when your business and revenues are all dependent on a closed centralised database and the products of your open source projects are deeply integrated with the same. DeBona’s use of language and Google’s attitude to AGPL is a very public reflection of this reality. Yes they are committed to Open Source, while it supports their preferred centralised database money making machine. Of course there is nothing wrong with Google making money or keeping their Db closed. Only don't argue they are serving the good of the greater global community by doing so. The reality is, open source provides a great way co-opt worker bees to help build services and an ecosystem that bolster that core business, which will remain closed and earning money for Google for as long as they can manage.

Which leads us on to Honeycomb, Google’s forthcoming tablet specific OS. The most likely reason for the delay to publishing Honeycomb source is due to the fact the tablet market isn't tied-in to operators who have the same incentive to stick with Google's momentum as they currently do for Android on phones. For tablets there is a greater danger a large manufacturer, will branch Android to create a tightly integrated killer device. So why should Google open themselves to that risk? Of course the official line about wanting to perfect the user experience before publishing the source is another one of those happy co-incident truths thrown out to placate and distract the worker bees in the open source community from commercial reality. It deflects but cannot change the simple fact that Google have the keys and they are keeping the box firmly closed until a time of their choosing - thus demonstrating they hold all the cards. Honeycomb will probably end up being open source software, but it isn’t Free Software (free as in “speech” not free as in “beer” etc.). And it is free (as in beer) only while it channels efforts into bolstering their centralised database and services.

I suspect Google are looking at Apple and realising the tablet market has such large potential, they are missing a trick. Google’s position is not so secure as was Microsoft’s during the 90’s and early noughties. The core business is search and unlike when you own your own OS you can be dislodged from it. There is a lot of noise about cloud based services, but encouraging adoption is proving harder than they anticipated. Now they are facing Occam’s razor; stick with Open Source in a hope it will bolster centralised database services, or recognise the tablet market offers huge potential revenues in it’s own right and the opportunity to get an asset people pay for like Microsoft had and still have. They are right there, in tablet number two spot, as the market is exploding but still not making money from it because they are pretending they don't own Android OS (they only don't wholly own the last publicly published version, which will always be 6-9 months behind HEAD - an aeon in IT). Is Apple revenue envy bringing about an identity crisis ? Don’t do evil / yes but make money / drop open source / no ! don’t do evil / yes but make money / drop open source / no ! don’t do evil / yes but make money.

[Edit: Of course I would suggest Google examine adopting an Open Interface license - then the Open Source community and Google could work without friction and Google could fairly deliver a commercial User Experience and potentially allow access to HEAD.]

Paul Lancefield on Twitter