Re: [PDO] wiki clean up and potential in person meeting
[email protected] (Wez Furlong) Tue, 29 Dec 2009 12:07:38 -0500
| Newsgroups | php.pdo |
|---|---|
| Message-ID | <[email protected]> |
On Dec 29, 2009, at 7:05 AM, Lester Caine wrote: > Matteo Beccati wrote: >>> Without a GROUP of people who are interested in developing PDO, I >>> am of >>> the opinion that it should actually be pulled back out of the core >>> since >>> the generic drivers ARE being developed further, but some projects >>> ARE >>> also converting to PDO in the believe that it already provides a >>> cross >>> database compatibility which is not really the case? >> I respectfully disagree: you're just biased because you are >> frustrated by the status of PDO_Firebird. Other drivers (SQLite, >> MySQL, PgSQL, OCI, DB2 maybe?) work pretty well and apart from some >> issues (e.g. CLOBs support in PDO_OCI) I still think that PDO >> should be core. It's PDO_Firebird that should be pulled out until >> someone can really start taking care of it. > > Along with OCI and DB2? and what about MSSQL. The CLOB problem is > fundamental to all of these databases .... if there was a way to get > around it, then I might be able to get somewhat further on using > pdo_firebird. > > The oracle drive is currently suffering from the same problems as > firebird, and it's apparently a generic problem with the way PDO has > be straight jacketed to provide an unrealistically limited set of > facilities. I've been educating folks about PDO for quite some time, and part of my presentation is asking people which databases they are using. You need to face the fact that no one (that I have spoken to) at numerous locations in the USA and in Europe uses firebird, and that it is rare for those folks to have even heard of it. The firebird driver needs a maintainer, but given that open source development is primarily itch-driven, and given that "no one" uses firebird (except you, it seems), it seems unlikely to me that it is going to progress any further. I think it is time to quit belly-aching about the state of that driver; either step up and do something about it by working with the community, or let it go. Regarding CLOBs and BLOBs, it is actually a pretty trivial change to enable support for those as distinct entities, but no one has yet scratched the itch and made it happen. PDO is not straight jacketed as you suggest; you have always had an extremely polarized view based on the sorry state of the firebird driver. > As far as I can see, ONLY sqlite, mysql and postgresql drivers > actually work to a level that they can be used transparently ( with > some fairly major restrictions on the SQL ). This was never a design goal for PDO. Data _access_ abstraction not _database_ abstraction. While some adjustments can be made to make the latter easier, it is an extremely challenging proposition, and I believe it is a mistake to make that the driving goal at the cost of exposing driver specific features. --Wez.