[thelist] Idea: Distributed CMS.

James Baster evolt at thesinner.co.uk
Tue Feb 25 17:00:01 CST 2003

Heres info about an Idea I have: I'm posting it here get feedback on it. I
do have techinical thoughts on how this would work, but that's for later:
for now, what do you all think of the idea? Would people find it useful -
would it work?



The Sinner is a web site I run at my university; its a typical community
website with messageboards, articles, polls and a calendar. Users can put
polls in messages, and under each article users can post messages. Nothing
special. (No link provided becuase it's a working site and I dont want it
disrupted; and it really has absolutely no unique features.)

But imagine that at each uni there was a similiar site: wouldn't it be great
if users from one site could sign in and participate on other sites in this
group? Wouldn't it be great if articles from one university that had
relevance at other universitys could be passed around? Wouldn't it be great
if threads from one sites messageboards could be shared on messageboards on
other sites?

I see this site being used by small local groups sharing a certain intrest
who want to form a larger network: for example, local sports clubs could run
there own small local sites but if something of national intrest happened,
like a national competion, content relating to this could be shared across
all the websites.

My Idea

Firstly, it is  important to note this I do not see a central database with
mini websites. In this system each site is its own master - a peer to peer
network in effect. This enables each individual site to keep control in
there hands. For instance, if a roleplaying club joins a national group of
roleplayers and later on disagrees with them on some fundamental idea, they
can if they choose simply leave the group and still own all there own
content, domain name and website.

Articles, Calendar events, threads on messageboards and polls are all
content that should be allowed to be pulled across. The site admins could
choose between having other sites simply link back to them, or they can pass
there content across and let the other site display it in there template.

This means from the persepctive of the user the end result is smoother and
more consistent. Issues have to be taken into consideration as maybe the
content providing site would like some recognition or financial reward for
the content. Options include passing a link such as "Article orginally
appeared on ..." or a advert to display along with the content or to have a
complete logging system so bills can be worked out later. Of course, some
sites might abuse this and only display the content and not the advert; the
only solution to this is to hope the providing site realises and stops
dealing with the offending site.

Current Work

"This kind of thing happens all the time!" is a fair cry here: yes, but
mainly on a very formal footing. http://www.theregister.co.uk presents
articles from securityfocus. ujournal lets users add  livejournal users to
there friends list. but these are very formal arangments done by big
companies who can afford to pay programmers to set up the link; its hard for
small sites to do the same.

RSS Feeds (http://backend.userland.com/rss#whatIsRss) are the odvious thing
to bring up; however they only allow other sites to grab headlines from your
site and thus provide a link back to your pages. This is all very well but I
see far more than this.

While RDF (http://www.w3.org/RDF/) seems to be a comprehensive standard for
describing information, as far as I can see there is not neccisarily a way
to pass information throught a bunch of websites automatically. (And
admitadly I havent looked at RDF in any kind of depth at all.) My idea would
need to include this.

Dupral (http://www.drupal.org/) is a CMS System that allows people with an
account on one dupral run site to log in to another dupral run site; but the
only actual content sharing I know of it does is via RSS feeds. Again I see
far more than this. As a format to describe information passed between the
websites in my idea, RDF might hold possibilities. In any case, XML would
definetly be recommended. Its flexible and many open source librarys exist
to manipulate it.

James Baster
james at thesinner dot co dot uk

More information about the thelist mailing list