[GLLUG] Retrieving a row from mysql
jason D. Justman
justmanj at msu.edu
Wed Feb 18 23:09:21 EST 2004
not having the strong background over adodb as you probably do, one of
the constraints i do see is the lack of "dumb" loadbalancing for non
scalable (read: mysql) databases. i'm not sure if you are familiar with
the "attempt" of loadbalancing that mysql has - read only for the slaves
and updates to the masters - but as of now, i dont see these design
"constraints" built into the mysql driver. granted, if you are only
serving up one/two pages a day its not a big deal.
i am surprised to see a strong presentation of global code in the
adodb.inc.php - it seems like a singleton pattern would have been a good
design choice instead of a long series of define and global vars (i cant
say i'm perfect either, its such an easy crutch when you need one piece
of information back and forth).
but beyond intricies of the implementation an abstraction layer. i do
prefer a lightweight, transparent record container that is extensible
without a custom adaptor class to bridge ado's recordset with subsequent
class operations. whereas ado only provides a record set container -
and the only return of data is either into an array or std_obj (php's
version of std::). perhaps the ability for automatic callbacks for
meta-type management from
http://phplens.com/lens/adodb/docs-adodb.htm#ex2 in moving the
transformation logic into its respective class.
plus, im not too sure how i feel with a 2k parent class to be inherited
from a pseudo-factory pattern (would not caching of
multiple-instantation of the same driver be a possibility) would work
with performance, and only have to open a new connection?. i do
recognize there is a runtime driver loader which is fine, i've been
anxiously waiting php5s autoload runtime code. but seeing as how apache
on os x likes to crash under our current load - im hesitant to try
tuckmmcache even on our pseudo-production machines.
its a great db abstraction layer for it's bredth and clever handing of
meta information - but it seems almost as if the adoconnection class is
somewhat monolithic (integration of db operations inside of a connection
class, likewise tight coupling of statistics).
j
Hampton, Rodney wrote:
> Don't want to hijack the thread, but I would like to hear why you
> aren't enamored with ADODB (not to be confused with microsoft's ADO
> technology)
> http://php.weblogs.com/ADODB
>
>
> -----Original Message-----
> From: jason D. Justman [mailto:justmanj at msu.edu]
> Sent: Tuesday, February 17, 2004 11:36 PM
> To: Seth Bembeneck
> Cc: 'GGLUG'
> Subject: Re: [GLLUG] Retrieving a row from mysql
>
> as rodney hampton suggested about using a database abstraction
> mechanism (or, if you are really 'reet, write your own if you ever
> need to do load balancing with mysql). ado sends shivvers down my
> spine, but the pear's dbi design is pretty decent. you'll find
> that doing:
>
>------------------------------------------------------------------------
>
>_______________________________________________
>linux-user mailing list
>linux-user at egr.msu.edu
>http://www.egr.msu.edu/mailman/listinfo/linux-user
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.egr.msu.edu/mailman-archives/linux-user/attachments/20040218/15d89200/attachment.htm
More information about the linux-user
mailing list