Re: [PDO] Re: [PHP-DEV] [PATCH] New PDO methods for PostgreSQL driver

On 03.06.2010, at 21:05, Lester Caine wrote:

> progress. Without a clean higher level library that we can standardise on, PDO is holding back progress rather than leading the way.


Right, at this point I must agree. I mean I spend a lot of time in MDB2 creating a unified API and I can confirm its possible. At the same time I as thrilled to not have to do all of this work in userland anymore when PDO came around.

But at this point the fact of the matter is that this lingering PDO is a brick on our legs.

Note here that one of the core ideas of PDO was to reduce the amount of code had to maintain and write to get a given RDBMS supported. Given that most vendors seem to prefer to invest their time in their own extensions, despite the fact that for quite some time all major libraries have adopted PDO as the go to interface (albeit many also providing adapters for the old extensions, there by sort of eliminating the benefit of not having to do these kinds of native ext to unified api aka pdo api wrappers). However it seems this benefit is not significant enough or the unwillingness to work on a common codebase by vendors (in light of their desire for a CLA) seems to indicate the approach has failed.

Now of course if people would start dedicating time and working on a PDO that really becomes able to replace the existing extensions: all the better.

But maybe a more realistic approach is just defining a set of OO interfaces that native extension authors can easily use to wrap their existing functionality. I would assume this would reduce the effort required to transform the existing logic.

regards,
Lukas Kahwe Smith
[email protected]