Tim O'Reilly vs. Eben Moglen

Something Kevin Dean linked me to yesterday that is related: "GPL is the new BSD in Web 2.0, and why this matters"

I'm beginning to see the whole problem here that Tim may be pointing to although apparently even Web 2.0 people themselves aren't yet fully involved in being a solution. If that remains so FSF may in fact end up being the one doing more about it than anyone else. AGPL is a start I suppose.

From Tims last paragraph:

Yes it's freedom that matters, but looking at the GPLv3 epic drafting exercise I think it's safe to say it would be very hard to extend it to protect wider freedoms.

I think the FSF should focus on software freedom, that's the core. People with other issues can wrap their conditions around the four freedoms all they like in their own separate code worlds like AGPL.

My big fear is that if I spend years writing a web app at great economic cost (earnings and advancement foregone), and release it under the GPL, it can be harvested and some key innovation - my unique selling point - bolted into Google Apps for example. What chance has my shoestring hosting startup site got to compete against that? Any chance they'll share those indexing and search algorithms?

At a lower end the hosting companies are fed up in the commodity game and are chomping at the bit for value-added development work, competing with their own resellers. They can take and not give back too. Freedom should be a two-way street or we've got a market even more stacked against innovators.

So ok the web app startup might have a better chance if a community forms and the code gets better faster, but a group of developers in a co-operative can do that pretty well and still not take the risk of losing their one edge through a GPL release.

Startups face much tougher competition now that the market is maturing. You see that in the fact that there just aren't that many new success stories, it's more and more about the incumbents adding features, and pac-man buyouts in a great shakedown. That said craigslist shows another way, for 29 staff anyway.

I've been considering a wrapper for the (A)GPL licence that only gives the four freedoms to users but not abusers, IE no corporate predators. I don't mind competing, that's todays reality, so long as it's a fair fight for my livelihood. Maybe I'm too adversarial in this anti-predator attitude. Elsewhere there is a similar wrapping exercise aimed at preventing code being used for military purposes. But soon you could have every conceivable issue wrapped around the four freedoms and end up with a creative commons roll your own type facility creating huge confusion. So I'm still looking at the AGPL.

If you think about the advantage Google and other colossal sites have, it's that you have powerful features, and a lot of data - OUR DATA. But it's the SAAS from giants that could really decimate freedom in the market. It's MS Office style lock-in again but minus the files in proprietary formats, it'll all be tucked away in their databases and customised to their closed source application features. Unless they have ODF export, I don't know if thats the case.

If that becomes the dominant platform and domains and email are hosted there too, will I be able to offer any development service to a locked in prospect beyond graphic design and saas customisation, and could even that get locked down with deals like distributorship's where only one company has the rights to do that in my geographical area?

At one level it's all the more reason why we should ensure FS develops, so even if the monsters take the code but keep their special sauce secret, at least we'll have good code, some chance.

There could be an answer that saves us all, and it revolves around the question - how much does SAAS have to be centralised? An idea popped into my head on nuxified regarding spam, that just as peer to peer networks share files, server apps could do similar with blacklists and heuristics to filter spam. But lets take that further, into the realm of every type of web app for example cms's.

If I'm searching it's usually English language sites on specific topics. Just like a p2p index points people in the right direction why can't servers in s2s networks point me there? If digg, reddit et al shared their indices I can find what I want without going near google. They could do similar with ad placement. An API could allow cms apps to network and expose the deep web using proper semantics, eg Dublin Core and specialised ontologies. That could be very useful, and these could be done with AGPL.

It's fair to say that these operators could be just as bad for privacy as the current lot, but then maybe central semantic indices could be provided by governments as public utilities.

Maybe I shouldn't be so concerned about the saas market, after all, surely it would be easy to make inroads and capture market share by giving people what they want, powerful features yet respect for privacy. Maybe a fairer deal is the only edge a startup needs to get in there and be successful. I'm tired now, I'll think more later.