[Chicago-talk] What exactly is a "Application Framework"

Jason Gessner jason at multiply.org
Tue Mar 29 07:49:32 PST 2005


On Mar 29, 2005, at 8:37 AM, Jim Thomason wrote:

>> Jim mentioned Basset get compared with CDBI a lot, but I wouldn't
> call CDBI a application
>> framework.  TT and Mason I'd call templating engines, good a 
>> generating
>> forms (web or paper) based on variables and input.  Maypole too
>
> For the record, I detest the comparison to CDBI. It comes about from
> people looking at the persistent object class and then saying, "Oh,
> this is just what CDBI does. Use that." And it's a silly argument
> since Basset provides much more functionality, but that's the bit
> people focus on, for some reason.
>

Why are you so defensive about the C::DBI comparison?  Data access is 
so crucial that people want something good, fast and simple.

Until you start trying to shoehorn C::DBI into being a reporting 
engine, it is quite a bit simpler than your basset examples (and 
requires no configuration file or editing of the module sources).  As 
for speed, again, it has its optimal uses, and its suboptimal uses.

If bassett is a totally general purpose application framework, then 
concentrate on being the glue between the MVC components, not rewriting 
them.  As glue it may be good.  But is your templating engine really 
better than mason, html::template or tt?  Probably not.  Is your DB 
access better than C::DBI?  Again, probably not.  If you concentrate on 
stripping out everything but the glue, perhaps your app can become as 
good as some of those apps.

As an example of a good domain specific application framework, Bryar 
provides a totally generic set of hooks to deal with its data 
persistence, front end and display components and consequently has many 
combinations of application setups available.  Despite some of its 
other issues, it is a very well designed framework for blog-esque 
publishing.

i wouldn't be so defensive.  You are putting your app out there and 
saying it is better than x.  ok, fine.  prove it.  Or prove even why it 
may be better than x in circumstance y.  Otherwise, relax a bit.  :)

-jason



More information about the Chicago-talk mailing list