[wp-hackers] Making WordPress database independent

Scott Kingsley Clark scott at skcdev.com
Mon Feb 7 20:05:35 UTC 2011


Great, that's what we need. People like you to stand up, put the
effort in, and let your voice be heard!

-Scott


On Feb 7, 1:57 pm, Piyush Mishra <m... at piyushmishra.com> wrote:
> I dont hope that they take it in, may be I do hope that they see and
> understand the need and may be put a different implementation or whatever
> suits the community more. I am not pro enough to make something that is
> perfect for everyone in its first go. but surely I love WP enough to push it
> in the direction I feel is right for its future.
>
> On Tue, Feb 8, 2011 at 1:00 AM, Scott Kingsley Clark <sc... at skcdev.com>wrote:
>
>
>
>
>
>
>
> > Actually, trunk is the current version of whatever's being developed,
> > right now that's 3.1 - though I must warn you that if you do fork this
> > on github and put your work into this - there is a very slim chance
> > your work will be considered for core. The only good you're probably
> > doing here is that you're giving other people a good starting point at
> > doing something similar. So don't walk into this thinking it's going
> > to get into core, walk into this as you're taking one of the first
> > steps towards showing how this can be done.
>
> > WP will likely remain on SVN for some time to come, though there are
> > internal talks of other options which involves a considerable amount
> > of time that may or may not make it feasible in the near future.
>
> > -Scott
>
> > On Feb 7, 1:26 pm, Piyush Mishra <m... at piyushmishra.com> wrote:
> > > Does the code athttp://core.svn.wordpress.org/trunk/remainat the
> > current
> > > development revision of WordPress? I can start the move from my end by
> > > cloning that into a github repo, work on it and developers who want the
> > > change can fork>update and send pull requests, if the core devs see the
> > move
> > > and understand the need of the change, they can move it back to svn/any
> > new
> > > versioning system that WP is planning? As Scott pointed out, "I think I
> > can"
>
> > > On Tue, Feb 8, 2011 at 12:46 AM, Scott Kingsley Clark <sc... at skcdev.com
> > >wrote:
>
> > > > Right, in my experience using a CMS is like buying a home. You know
> > > > you may eventually move so you're willing to make some concessions,
> > > > but if it's missing something you absolutely want then you're more
> > > > than likely going to look at other houses that have it over taking a
> > > > house that doesn't. For businesses, it's sort of the same thing since
> > > > people looking at a CMS look for certain features that they need, if a
> > > > CMS doesn't have it then they're going to move onto other ones that do
> > > > - and WP never knows about it because they glance at WP and it's
> > > > requirements and say "That can't do that", regardless of the existence
> > > > of some complicated installation process to get it to do what they
> > > > want. You could almost compare this to life before Custom Post Types,
> > > > in which you could just use Pages or Posts with categories for
> > > > everything. Yeah, you could get it done - but it's not an efficient
> > > > solution - it's a hack.
>
> > > > -Scott
>
> > > > On Feb 7, 1:09 pm, Piyush Mishra <m... at piyushmishra.com> wrote:
> > > > > oh, I dont think I can work my way through the talking with core devs
> > and
> > > > > contributors and explaining/realizing why or why not to support more
> > > > dbmses
> > > > > What I can do is, complete the layer on a fork of wordpress on github
> > > > with
> > > > > some help and may be hope for the core devs to pull that in as the
> > work
> > > > is
> > > > > already done. I can work on it as a GSoC project and lend a helping
> > hand.
>
> > > > > As far as I know, in regex vs rewrite, rewrite wins hands down in
> > > > > performance, cleaner codes, decoupling. and I dont need numbers/tests
> > to
> > > > > prove that. now the fact that almost all people use wordpress with
> > MySQL
> > > > and
> > > > > the core devs donot feel the need to implement the layer, imho is
> > kind of
> > > > > mutual. people wont know/use any product for something that the
> > product
> > > > > doesn't support out of the box / with some simplistic plugin. if they
> > > > need
> > > > > another dbms so bad and donot get a simple solution for it, they can
> > > > simply
> > > > > switch to another CMS/framework that supports it or if they really
> > love
> > > > wp,
> > > > > they'll probably make hacks better than adding db.php(obviously
> > putting
> > > > them
> > > > > out of the updates cycle) so this MySQL only trend can go on forever.
> > :(
>
> > > > > On Tue, Feb 8, 2011 at 12:21 AM, Scott Kingsley Clark <
> > sc... at skcdev.com
> > > > >wrote:
>
> > > > > > Some discussion has been made on the WordPress Core Trac regarding
> > > > > > supporting other database types, though it appears that the
> > direction
> > > > > > that WordPress core has taken is against supporting other database
> > > > > > types comes at the cost of increasing the amount of effort in
> > either
> > > > > > rewriting functions and / or RegEx-ing your way through the
> > barriers
> > > > > > that the MySQL only support WordPress currently restricts itself
> > to.
> > > > > > Yes, many developers have gone through this and built solutions of
> > > > > > many different types, but it doesn't compare to what real
> > integration
> > > > > > could bring, or additional provisions within Core for these
> > efforts.
>
> > > > > > Any attempt to abstract WordPress beyond MySQL will only waste core
> > > > > > dev time as they have to then explain why they won't be making any
> > > > > > changes. I apologize for my own time wasting on that end - I
> > believe
> > > > > > some sort of page in the Codex or About section of the WP.org
> > should
> > > > > > compile the full explanation of the reasoning behind the MySQL-only
> > > > > > support WP core has restricted itself to with some sort of evidence
> > in
> > > > > > favor of that restriction, offering up additional resources to
> > plugins
> > > > > > which were developed to support other engines that have been
> > > > > > recommended by core devs or other respectable devs / firms in the
> > WP
> > > > > > community.
>
> > > > > > There's much to be done, calling WP a full blown CMS right now is
> > > > > > tough, since many people's idea of 'full' can be different. In this
> > > > > > database discussion, a great question rises yet again as it has
> > over
> > > > > > the past many years of WP's life.
>
> > > > > > Will there be some form of redemption by core as more community
> > > > > > members speak out in favor of additional support? No one can tell,
> > at
> > > > > > least not for the very near future. We'll see how the  WordPress
> > > > > > landscape changes over the next few years. For now, your only hope
> > is
> > > > > > to prove them wrong -- that people want to use these other database
> > > > > > engines -- with solid proof of why it would be a good addition to
> > core
> > > > > > and the impact it could have in making WordPress better for more
> > > > > > people than just MySQL alone. Think of your task as being a
> > prosecutor
> > > > > > in a court case, WordPress is the defendant, and you've got to
> > > > > > persuade the jury (Core Devs and major contributors) to see the
> > future
> > > > > > and what they're doing by limiting growth.
>
> > > > > > That's my two dollars anyways.. I've been from the bottom to top of
> > > > > > the Core Dev ladder, the answer is the same all around -- They
> > simply
> > > > > > don't have this as a priority or even in sights for the next few
> > years
> > > > > > and for it to happen we will need to orchestrate this as a
> > community
> > > > > > effort to prove to them that it is a priority. I know right now, it
> > > > > > seems like we're protesting about the war in Vietnam (US History),
> > but
> > > > > > we can only make a difference if it's truly what the community
> > wants
> > > > > > and is the 'right' choice to expand the use of WordPress and it's
> > > > > > ability to perform to meet the needs of everyone who it can
> > benefit.
>
> > > > > > Good luck until then, just say "I think I can", sorry for the video
> > -
> > > > > > I had to:http://www.youtube.com/watch?v=LoVRV_xGEqc
>
> > > > > > -Scott
>
> > > > > > On Feb 7, 11:23 am, Piyush Mishra <m... at piyushmishra.com> wrote:
> > > > > > > May be using DAOs will help. we can push all functionality to the
> > DAO
> > > > of
> > > > > > the
> > > > > > > respective db layer and simple inbuilt functions from the
> > specific
> > > > vendor
> > > > > > > can be used where needed and in other databases, the same
> > features
> > > > can be
> > > > > > > implemented at the code level??
> > > > > > > Just an idea yet. but it seems good to me.
> > > > > > > 1 factory, 1 PDO or similar base class and few interfaces and
> > their
> > > > mysql
> > > > > > > implementation will get us to where we currently stand with
> > MySQL, we
> > > > > > might
> > > > > > > not actually need whole ORMs etc
>
> > > > > > > On Mon, Feb 7, 2011 at 10:33 PM, Brian Layman <
> > > > > > wp-hack... at thecodecave.com>wrote:
>
> > > > > > > > On 2/7/2011 11:49 AM, Piyush Mishra wrote:
>
> > > > > > > >> With NoSQL picking up craze, postgre gaining popularity and
> > > > WordPress
> > > > > > > >> moving
> > > > > > > >> towards being a complete CMS, its a move that we should make.
> > > > > > > >> Propel might be a good option, may be DAOs.
> > > > > > > >> But this option is good for progress of the CMS as a whole.
> > > > > > > >> I bet people will start discovering new areas for improvement
> > > > within
> > > > > > the
> > > > > > > >> core as they implement database abstraction.
>
> > > > > > > > Will they?  Admittedly anything that causes people to look at
> > code
> > > > > > again
> > > > > > > > can do that, but would it introduce any major improvements?
> > > > Remember -
> > > > > > every
> > > > > > > > feature that isn't supported in MySQL and all our "supported"
> > > > database
> > > > > > > > engines would be excluded from use..  Heck, we can't even
> > > > generically
> > > > > > use
> > > > > > > > features that require MySQL 5.0 yet (assuming 3.1 hasn't
> > dropped
> > > > before
> > > > > > I
> > > > > > > > hit send).
>
> > > > > > > > I'm really curious to see how having access to stored
> > procedures,
> > > > > > triggers,
> > > > > > > > views will affect the WordPress Core.  I wonder where that
> > rabbit
> > > > trail
> > > > > > will
> > > > > > > > lead us....
>
> read more »


More information about the wp-hackers mailing list