cocoon-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Neeme Praks" <ne...@one.lv>
Subject RE: Thoughts on a data-driven web site
Date Wed, 21 Jun 2000 19:16:48 GMT

I've been struggling with something similar, so I have couple of
questions/comments...

> -----Original Message-----
> From: Jonathan Stimmel [mailto:jon-lists@stimmel.net]
> Sent: Wednesday, June 21, 2000 8:31 PM

[snip]

> In order to fully support integrating content from (potentially)
> dozens of different data sources, I believe we need to use the
> following model for dynamic delivery:
> 
>  +-----------------------------+
>  |    user requests a page     |
>  +-----------------------------+
>                 |
>                 v
>  +-----------------------------+
>  | 1) identify content sources | <- (from index.xml, e.g.)
>  +-----------------------------+
>                 |
>                 v
>  +-----------------------------+
>  | 2) populate content         | <- (using a series of 
> lightweight taglibs)
>  +-----------------------------+
>                 |
>                 v
>  +-----------------------------+
>  | 3) perform rough layout     | <- (using XSL)
>  +-----------------------------+
>                 |
>                 v
>  +-----------------------------+
>  | 4) perform final "render"   | <- (using XSL)
>  +-----------------------------+

In which stage do you perform the actual information layout (meaning:
where do you specify how the content should be laid out on screen?)?

I came up with two solutions: combine it with stage 1 or stage 3.
If I would take the latter approach and combine it with stage 3 then I
would need to write specific stylesheets for each different page on my
site. I don't like that, but the more I think about it, it could turn
out to be the best solution.

The approach that I chose right now is to put some generic hints about
the layout of the page in stage 1. Then, at later stage I would turn
these hints into markup speficic to requested media / skin.

So, it would look something like this:

  +-----------------------------+
  |    user requests a page     |
  +-----------------------------+
                 |
                 v
  +-----------------------------+
  | 1) identify content sources | <- (from index.xml, e.g.)
  |  + identify logical layout  |
  +-----------------------------+
                 |
                 v
  +-----------------------------+
  | 2) populate content         | <- (using a series of lightweight
taglibs)
  +-----------------------------+
                 |
                 v
  +-----------------------------+
  | 4) perform final "render"   | <- (using XSL)
  +-----------------------------+


The alternative that I'm also thinking would be something similar to
yours:

  +-----------------------------+
  |    user requests a page     |
  +-----------------------------+
                 |
                 v
  +-----------------------------+
  | 1) identify content sources | <- (from index.xml, e.g.)
  +-----------------------------+
                 |
                 v
  +-----------------------------+
  | 2) populate content         | <- (using a series of lightweight
taglibs)
  +-----------------------------+
                 |
                 v
  +-----------------------------+
  | 3) identify media and make  | <- (using XSL)
  |    layout for that media    |
  +-----------------------------+
                 |
                 v
  +-----------------------------+
  | 4) perform final "render"   | <- (using XSL)
  |   (if needed for this media)|
  +-----------------------------+

But, as I said, the drawback of this is that you need separate
stylesheets for each page at stage 3 while the pervious approach can
reuse (possibliy only one) style.

Well... and this actually brings up another issue of reuse that I have
been thinking about... Is it really worth to try to reuse even the stage
1 (content sources)? After all, WAP client has very different
requirements than usual web browser and the same form that I would feed
to web browser is maybe not really suitable for WAP client...?

Maybe we all would be better off by building a separate pipeline for WAP
rather than trying to reuse the same pipeline for all the clients and
applying just a separate style at the end of the pipeline.

Cocoon2 sitemap and pipeline processing will make this easier for us,
but till then...

These turned out to be more or less random thoughts... ;-)

Neeme

Mime
View raw message