RE: Cocoon

> -----Original Message-----
> From: Steven Dale
> Geoff,
>
> I am getting ready to play around with Cocoon with an eye towards
> Accessiblity.  I believe it would be a great tool if we can get it
> working in a best practices type of way.  Have you played with it?   I
> was involved a couple years ago when developing web content for both
> the PCs and WAP phones.  I think it would be great to tailor the page
> output based upon the user's preferences.
>
> -Steve
>

I intend to do much the same, but for me not only does the application
framework need to work properly, as well as the CMS, but there needs to be
ways that the CMS can collect enough metadata to extract from the source xml
for the transformed documents to be semantically rich and populate the
various attribute/elements with metadata that enrich them to user agents
addressing accessibility.

Not one commercial CMS collects metadata correctly, if at all.
Interwoven/Teamsite is the only one that allows a few fields for metadata,
but it is an EXTREMELY primitive approach.  It seems to me, very few
architects have seen how important this is to the whole CMS workflow.  It
astounds me.  Anyone who does this, immediately has huge advantages over the
rest, and it is not just in the ability to produce semantically rich
content, but to bring detailed workflow analysis.

This is also a problem working with MS Office and the discipline of putting
the right metadata in properties and using this meta repository.  The custom
fields in MS Office cannot be read on the scripting side, last time I wrote
anything to extract these fields.  I haven't tried it with Open Office, but
at least that developer community is open to feedback.

The other thing I love about the Cocoon architecture is the use of TCN
(transparent content negotiation), something that is SADLY underused in web
content delivery.  That is the other area I want to focus on in using
Cocoon.

I'd also like to get on the back of user agent developers and try and fix
the http header accuracy as far as the "Accept" headers are concern, so that
TCN can become a more viable option.  It has improved, since the Netscape 2
debacle that wrecked it, but it is still inconsistently supported (last time
I looked).

Geoff

Received on Wednesday, 23 June 2004 21:23:01 UTC