I have used Template-Toolkit, and the time used for parsing those templates used to double the whole time the program needed to run, so the speed of template parsing could be important.
Teddy ----- Original Message ----- From: "Jonathan Vanasco" <[EMAIL PROTECTED]> To: "mod_perl List" <modperl@perl.apache.org> Sent: Wednesday, September 07, 2005 20:00 PM Subject: Re: a faster html::template? > > Just to reiterate - the difference in templating time is very likely > negligible when you factor in your application logic. > > ie: > on my tests a HTML::Template file interpolated in .00x, while a > HTML::Template::JIT interpolated in .000y > but my application logic takes .025 to work > > .25x vs .250y is not a large difference -- and compared with the ease > in writing and bugfixing templates on some of the slower > implementations, it just became a much better decision to ignore > templating speed > > that said, i chose Petal because TAL is (*#&$ great and supported under > many languages -- for example, if a stats reporting page is running too > slow, I can use the same template and just have apache proxypass that > url to a TwistedPython daemon that crunches numbers way faster than > perl can. > > It's REALLY easy to get caught up in creating the most efficient > templating system. I fall into that trap at least 2x a year. > > Just remember that the templating is only a fraction of your > application, and is likely to be the fastest part of your application > by a significant amount. > > It's good to have a friend who you believe is a better coder than you > and occasionally ask him to belittle you on your every idea to try and > keep you in check. > > > > On Sep 7, 2005, at 10:49 AM, Perrin Harkins wrote: > > > That's the trade-off: Compiled drops some of the H::T features, which > > may be a problem if you are trying to use existing templates with it. > > JIT is pretty true to the H::T feature set, and that may be part of > > what > > slows it down. It's still many times faster than H::T though. >
