Moving To Catalyst, Incrementally?

Bill Barry after.fallout at gmail.com
Thu Apr 30 03:21:13 UTC 2009


dieter.pearcey at gmail.com wrote:
> On Apr 28, 2:33 pm, "Gregary Hendricks" <ghendri... at novell.com> wrote:
>   
>> One of the things that they demonstrated at YAPC and OSCON was the benchmarks. Catalyst would likely speed things way up for any users still running mod_cgi, and would likely have speed benefits for hose running mod_perl. Probably the biggest speed benefit of the framework is that it consolodates all the database handling. My understanding of DBIx::Class is that it is really fast when compared to standard DBI. Max might have more to say on that though.
>>     
>
> I guess that's possible, but it would surprise me.  In particular, I
> find it hard to imagine that DBIx::Class would be *faster* than DBI.
>
> All of this stuff is optimized for development time.  You can always
> hand-optimize code to run faster than abstractions; it's just rarely
> worth the time it takes to do so.
>   
In my experience with NHibernate, it is often faster than hand written 
query based apps. This is because it caches things heavily and batches 
writes to the database. Properly configured it also almost universally 
uses proper indexes too. I assume DBIx::Class would do the same thing 
(memcached support, flush buffers, ...). The point being that, sure you 
could optimize a single query to be be faster, but that isn't the proper 
optimization point most of the time (it is at the 
batches-of-queries-level to reduce db roundtrips). 
_______________________________________________
dev-apps-bugzilla mailing list
dev-apps-bugzilla at lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-apps-bugzilla



More information about the developers mailing list