Re: [PDO] [RFC] An Idea for PDO 2

On 14.02.2008, at 22:33, Christopher Jones wrote:

Lukas Kahwe Smith wrote:

We'd support this. From my last set discussions internally, we'd

need

the spec to be under CLA.

Ok, thats a problem. I do not see much of a chance to convince anyone on php.net to do this. What other extensions have been written using an external (to PHP) specification?

You are missing the point here. A document that all drivers are expected to follow needs to be CLA free or it has no chance of being accepted at php.net. After all SQLite, MySQL, PostgreSQL, Firebird and friends would also be expected to follow those specs.

But lets say its not under a CLA, how is this different than you guys using the ZE2 API? Is it just because Oracle considers anything closer to RDBMS to be more of a risk? If we say that we do not expect much direct collaboration at all from any of the concerned vendors, is that still an issue? Using is different to contributing to a specification. And collaborating is different to working individually. (And the more that our legal team understood about PHP, the less they liked that management had already approved me contributing...) I think it's an issue for PHP if not all the data access providers can contribute.

My point is .. if the data providers that are concerned should simply not contribute. Then the data access providers would only be using a specification that is available under an OSS license. If there is concerns about patents, there is the same potential for hidden unlicensed patents in the ZE2 API.

From my point of view, the interesting crux of the PDO

discussions is

that the development and release _process_ is the key factor in determining what risk (and hence need for CLA's and licenses)

exists. Right. The base line is that anyone can setup a PEAR channel and distribute PHP or C code on their own infrastructure. We are looking if there is a way to more closely integrate you guys into the php.net infrastructure. Installation of PHP continues to be the number one barrier to entry issue I see, despite the maturation of various *AMP distros, Zend Core, and other prebuilt stacks. Reducing the barrier is the key reason to have PDO and its drivers integrated with PHP to a high level. It also makes sense to have a central place for documentation and bug reports.

Actually I do not see this issue. PHP is much easier to install any other competing web platform. That is its current secret to success. As such I do agree we should not make it that much harder. Which is why I agree that we need to work on making it easy for people to also install binaries via the PEAR/Pyrus installer. regards, Lukas