Why the GPL is bad for the world

A recent Slashdot article on the Verizon GPL infringement suit got me to thinking - the reciprocal nature of the GPL may be partly to blame for the whole SAAS movement, which itself encourages the continued concentration of power into the hands of a few.

You can deploy all the GPL'd software you want to a server - in that case you're not distributing binaries, you're just using them to deliver something else. But for companies that want to sell software, GPL is a poison pill. They can't sell software that anyone can just take for free. Preventing people from selling software will tend to concentrate power into the hands of the few organizations that can use economies of scale to barter their services for end-users' attention.

That's right - I'm saying that commercial software is actually the answer to unchecked consolidation of power in the software business. It's time people got used to paying for software again, with cash, rather than paying for it below the table (where consumers tend to get screwed).

3 comments:

Drew Kutchar said...

Yes, I have been thinking about this for a while too. That's the problem with what I like to call "the first generation" of open source licenses which are mainly based on the concept of software sale, which at this point is becoming less and less popular with SaaS.

I think that's the main driving factor for "Attribution" based licenses these days, such as the Common Public Attribution License and Yahoo Public License.

there was a good article about it on eweek.com

Anonymous said...

Slashdot publishes articles? :)

I'm not defending them, and share some of your concerns, but not for the same reason. The FSF takes great care to describe selling GPL software for a fee. There is no requirement in the GPL to make binaries available and freely redistributable.

The issue with BusyBox thing is, as I understand it, that they allegedly took the software, modified it and sold it as theirs without re-releasing the source. It appears Verizon subsequently re-sold that product, and thus could be brought into a law suite. If they have indemnification from BusyBox they'll be able to get out of it.

A better example would be TiVo, who was the darling of the Linux community when they were the underdog, but is now seen as evil because they don't embrace GPL for everything, and support DRM. TiVo does re-release the source they change. http://dynamic.tivo.com/linux/linux.asp (an ironic URL)

Oddly enough, the Linux community still supports IBM's Linux mainframe work, even though IBM requires the delivery of binary-only blobs to the kernel (which most see as a GPL violation in other situations). Even Linus Torvalds had commented publicly on this, saying he thought it was fine.

Personally, I see the whole SAAS thing more as an effect of a more wired net and people getting comfortable with using online services. The same thing is happening with finance (PayPal, online billpay), etc.

It is convenient for some though that they can wrap themselves in Open Source, yet make modifications, and only provide services and not have to re-release the source code.

I don't fault anyone for choosing to use the GPL, but I'm afraid far too many do without understanding what restrictions they're placing on others. Sun's Open Source guy, Simon Phipps, authored a doc specifically on this (intended for others) and solicited comments.

Personally, I'd rather use Mozilla, CDDL, or Apache style licenses for most things. They give both me and others MORE flexibility than the GPL.

Anonymous said...

I don't think the rise of SaaS is a result of the GPL at all, it's a matter of ease of deployment, the ubiquity of the browser, and the rise of network computing. If the GPL provided more lenient software distribution rights like the BSD license it doesn't necessarily mean we'd be using more desktop software to do the things we do at Flickr, Gmail, Facebook, etc. There are just so many advantages to web-based software. The fact that many respected web sites have XML APIs so one could and does write desktop software to interface with these services demonstrates this false correlation.