>===== Original Message From Avi Bryant <avi / beta4.com> =====
>On Wed, 18 Apr 2001, Erich Schubert wrote:
>
>> IOWA is surely an interesting development, but it lacks the most
>> important features (like deeply integrated caching), but it already has 
some
>> things i don't want:
>>
>> - i don't want the Ruby part to know anything about HTML or event to
>> modify my pages ( <a oid="log"> <--- i don't want this)
>
>That's a very interesting statement... I don't know if this is the forum
>to get into this, but the major design goal of IOWA was precisely the
>opposite: the Ruby part should know as much as possible about HTML so that
>the developer doesn't have to.  I find systems like PHP, etc, very
>difficult to deal with because the developer is constantly having to
>generate various pieces of html syntax by hand; I wanted to abstract all
>this away.  To put it another way - when the HTML template is being
>written by a web designer and not the developer, I would much rather that
>the IOWA library have intimate knowledge of the template's structure and
>the developer have none, than the other way around.

I would need to look at IOWA closely to say more, but my
preference is definitely that _neither_ the templating
solution nor the developer has any business trying to
understand HTML.  The developer shouldn't because the
developer doesn't want to become a bottleneck for the
designer.  The templating solution shouldn't because it
is very likely to need to deal with many kinds of text,
not just HTML.

The Perl world has many templating solutions.  The best
of these (by general acknowledgement) is the Template
Toolkit.  Their white paper is at

http://www.template-toolkit.org/tpc4/paper.html

Whether or not you like Perl, I recommend reading this
just to learn what the "current state of the art" is in
thinking about templating as a problem.

Cheers,
Ben