> From: Perrin Harkins [mailto:[EMAIL PROTECTED]] 
> Sent: 01 June 2002 00:17
> To: Rob Nagler
> Cc: [EMAIL PROTECTED]

> The same template?  How does the layout manager help with that?  
> Does it modify the template?  It would make more sense to me if 
> this were a sort of abstraction to factor out common layout 
> ideas from multiple templates.

The layout manager is more complex than a static template. It takes 
a collection of data and meta-data and creates a document tree that 
is not a specific rendering flavour - the DOM might subsequently
be rendered in NS4.0 HTML, IE5.01+ HTML, Excel, PDF etc. without 
requiring (every page) x (every format) x (every option) templates.

> ...
> That doesn't require a layout manager though.  Simple templating is 
> fine for that, i.e. two different templates (views).

In most circumstances it is ok to have templates that are 
specific to one output flavour, possible with some simple alts.

But what if I have 40 pages that I want to render in: IE5+HTML, 
NS4-HTML, Excel, HP-PCL5 and PDF - and I also want SIMPLE views 
of each page. This would result in 40 x 5 x 2 = 400 templates. 
For every change in a basic page, I have to keep 9 other templates 
in other languages in step.

If the number of output formats and page variants is not trivial, 
it might make more sense to have an active component that creates 
the document tree, and passes it to simple components that know how 
to render a DOM in the desired output format.

All in all, I think the MVC+template approach described is fine for
most of the type of dynamic sites currently available on the web.

But I work for a small ASP and we have to manage rapidly changing 
complex dynamic sites and complex user reporting requirements, with
minimal work. We don't have any HTML specialists, but rather a few 
very good technical generalists. I think that for us a meta-data 
layout driven approach will be more effective. 

We do however plan to borrow all the good ideas we can find from any 
other model, including MVC+template. Our planned move to mod_perl is
mostly due to the benefits we have seen from out Perl data processing
and reporting - hence the desire to join both the web and DP infra at
the hip.

Regards
Jeff


Reply via email to