On Mon, Jun 18, 2001 at 11:08:06PM +0100, David Cantrell wrote:
> On Mon, Jun 18, 2001 at 08:24:13PM +0100, Matthew Byng-Maddick wrote:
> > On Mon, Jun 18, 2001 at 07:54:36PM +0100, David Cantrell wrote:
> > > It is NOT POSSIBLE to completely divorce presentation/application.
> > > So you end up with all sorts of languages made up to be mixed in with
> > > the presentation - like PHP and the mini-language of TT. Why are
> > > those OK (I'm thinking specifically of TT - we all know PHP sucks for
> > > other reasons) but plain ol' perl isn't?
> > Ohmigod, I'm agreeing with Cantrell on something!!
> What am I doing wrong? ;-)
The next few paras. :-)
> Seriously, I agree 100% that you should strive to seperate application
> from your presentation as much as possible, but seeing that you can not
> do this entirely, you may as well embed perl in your HTML and save
> yourself the trouble of inventing a whole new wheel.
I have done both. I have to say, that ultimately, I think I prefer the
code/data separation, and certainly started looking more at that kind
of thing, having had to maintain templates I once wrote which were mixed.
Those templates only had calls into modules elsewhere, so the main code
was in modules, but it's the same argument, ultimately as accessor methods
in OO.
> You can still stick your business logic elsewhere and have that called
> by the perl embedded in the templates.
Exactly what we did. It still ended up being unmaintainable.
> > Despite having written an embedded perl templating system, I'm now very
> > much in favour of one where the tags are just delimiters as far as possible.
> > Thus I think things like HTML::Template are actually better than TT2,
> > precisely because the toy language in TT2 is just as bad as embedding code.
> > See my point about SQL, as it's related to this.
> Think of SQL as being a cross-language extension to the 'host' language
> and you'll feel much better about it :-)
I think the smiley indicates the point of this. :-) And there, you use
things like stored procedures which go a long way towards that separation.
MBM
--
Matthew Byng-Maddick <[EMAIL PROTECTED]> http://colondot.net/
