I've been quiet so far about the PDO/CLA issues because I wanted to think a bit more about it so the reply would not be an impulsive one. I am copying it here for additional comments:
I've been deliberating looking at all the comments regarding PDO2 and the CLA proposal to allow for some more thinking-before-writing. My first impression was not much unlike Pierre's though. Now that I have spend some time thinking about it, I am replying with my thoughts and comments. They are my personal thoughts on this and can not be attributed to any other entity.
First of all, it's good that some attention is give to PDO. PDO is useful, but also has many rough edges in the ways that it behaves different depending on different database connection libraries. This is highly annoying, even more because it's not really documented. Anything to improve this is good, and support from the vendors themselves definitely would help here. However, love to PDO should be given in a similar way as to the rest of PHP. I disagree, but understand, how PDO2 got started. It's not that much an issue to start with a small group to talk about certain new things, but it should have been possible for other people to join as well.
The outcome of this process is however what went into the wrong hole for me. From the start with Wez' commit to create a closed off module in our CVS repository I was skeptical. I've now seen the proposed CLA and PDO license and it seems my scepticism is justified. PHP has been an Open Source project for more than 10 years now, and I've spend a good deal of my spare time in the last 6 years on it. Never was there any need for a CLA although there were some minor IP related issues which got cleared up pretty quickly. It happens, and a CLA can not prevent this. An example here is that when one of the CLA-signed PHP developers talks at a conference with another person (that did not sign the CLA) on some PDO related issues, the ideas that this other person brings up can not be used as it's not own contribution. This effectively stops discussing relevant technical issues with peers at conferences for example. I know the CLA only talks about actually stuff that makes it in (into the spec or code), but can you really rightfully claim it your contribution if somebody else suggested it orally? Another issue here is that we can't really have a public bug tracker for PDO where people can put in patches. The PDO developers (that signed a CLA) can not even look for issues as they might be tainted with evil patches. This practise is in spirit against Open Source as we've been practising it in the PHP project.
For now this CLA is only covering PDO, but once this precedence is set, it can easily expand over the rest of PHP as "it works fine for PDO". Another expansion could easily turn a CLA into an NDA if you really make it look black.
Another issue I see with the CLA is the patent clause. Where I live (Norway) and where I've lived (The Netherlands, Italy) there are no software patents. The practise does simply not exist, and for good reasons. Thoughts are free, and should stay free. It's absurd that trivial patents such as amazon's "One click patent" are even considered to be granted. Because software patents are such a moronic thing I would never agree to sign anything that mentions that I "warrant that the submission of My Contribution will include accurate details of" "related patents" "of which I am aware off". Although the FAQ writes that we're not expected to do a patent search, nothing is done to prevent the litigation against individual contributers in case their contribution was to be covered by a US software patent. This CLA does not give this protection to the contributors that is mentioned in the FAQ. It merely serves the interests of the big vendors which have 10.000s of patents themselves. In order to do real good for Open Source, those patents should be provided to the Open Source community free-of-charge. When IBM, MS and other big ones (Sun, here is your chance) provide their patents portfolio to Open Source projects other companies will think twice before trying to sue PHP (or their contributors) for patent infringement.
Besides some of the more legal issues, I've also concerns about developer interest in maintaining, or contributing, towards CLA covered bits of code. For eZ's projects we also have a CLA. There have been a few occasions where this CLA prevented people from contributing code. This is not what Open Source is about. It's not only about being able to use code freely, it's just as much as making it easy to contribute back. A CLA hinders this process, even more in the cases where simple updates (API changes in PHP f.e., proto updates, operating system support improvements) can not be done by every PHP contributor because they didn't sign a CLA. I do know that there were not many contributors actually looking at PDO before, but I think that has no influence on whether a CLA is good or not. I think the lack of developers is more because of the mostly (undocumented) complex workings of the extension and its drivers.
Besides this, the PHP project is our project, and if the big vendors want to make money through the PHP project by making PHP connect to their products better, I don't see why they should not follow our rules instead of dictating their own. There have been plenty of occasions as well where code (tests, fixes) were already contributed by either IBM and Oracle anyway.
I also have an issue with the license for the PDO parts. Although the FAQ writes correctly that there are some parts of PHP that do not fall under the PHP License, all of those parts have not been written for the PHP project specifically - they were always adopted from other sources. PDO is part of the PHP project, and should therefore not come with its own license that has to be OSI approved again.
I hope that I didn't forget anything in this longish email, but it should be clear that I'm totally against having a CLA on any part of PHP.