Manuel,
part of the effort to write an abstraction layer is to factor out common
code and patterns and move this into a common base so that higher level
development becomes easier. a good example of this is persistent connections
handling, which is handled almost identically in all the db extensions.
another example is code implementing the different varieties of fetching
(fetch_array et al).
At the PHP level there are common patterns of use which may be aided by
support at lower levels. bulk fetches and resultset iteration to construct
markup and resultset caching are examples. in addition there are features
which when implemented at a core level, creates value which is automatically
available to users of each backend. while they can be implemented in PHP
itself, but make better sense as "core" functionality from a performance
standpoint. also, there is functionality not covered by the existing APIs,
hence unavailable totally to PHP ( eg fetching multiple resultsets from the
same query ).
in other words, i dont see the development of an abstraction layer in C
and PHP as competing propositions. the C layer just allows PHP developers to
be more productive and focus on higher levels of application development.
the current PHP abstractions fill the need not only of functionality, but of
code/work style and "aesthetic" preference. that need will persist whatever
happens at engine level.
One of the biggest pluses of PHP is that implementation of new
functionality is magnitudes faster than in "C". if a necessary feature is
needed, we dont have to depend on the release cycle of PHP to have it
included in a distribution. things in this business happens too quickly. and
on a personal/ideological level, many of the features included in Metabase
belong at the engine level (not that theyre not necessary). for instance, i
think anything touching automatic schema modification or generation should
absolutely be in user-space. use of PHP or standard extensions should not
automatically impose a methodology on a user. i like the work you did with
Metabase and think that the work being done on a new extension will make
your work easier by being able to focus more on portability issues, which is
Metabase's strength (imo) and which should be dealt with in PHP.
l0t3k
Manuel Lemos <[EMAIL PROTECTED]> wrote in message
[EMAIL PROTECTED]">news:[EMAIL PROTECTED]...
> Hello,
>
> Markus Fischer wrote:
> >
> > On Sat, Oct 27, 2001 at 12:29:48PM -0200, Manuel Lemos wrote :
> > > I am afraid that will never happen.
> >
> > A project for db abstraction as an extension is already in the
> > works. History discussions about it can be found on pear-dev
> > archiv.
>
> What I meant that it would never happen is the adoption of an existing
> database abstraction package.
>
>
> Regards,
> Manuel Lemos
--
PHP Development Mailing List <http://www.php.net/>
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
To contact the list administrators, e-mail: [EMAIL PROTECTED]
