Newspaper site reformation 1.1
Okay. I’m pretty sure I’m going to build a new system to act as the CMS of the new newspaper sites.
Since I’m doing so, I may as well choose the technologies that I feel most confident with.
In this case the web pages will be built with PHP and Apache, and the database will be Postgresql. These will all run on FreeBSD.
Why build instead of using Drupal or Joomla? It is a tough call, but I have yet to see a CMS that isn’t story centric, or even more importantly, “site centric.”
Yes, you can achieve a great degree of “community” with a lot of CMS’s out there, especially with a blogging platform like Wordpress, which I seriously considered for many good reasons.
But the future, IMHO, is in distributed systems, and I don’t see anything that is built specifically with that in mind.
I’m basing the system on RSS 2.0 and all of the data that will be treated as atomic units, whether it was created by a staff reporter or photographer, a site visitor, or it comes from another system or site entirely.
If it sounds like an aggregator, well it is, but it’s a very smart aggregator, because I don’t mean to suggest that we won’t be displaying our original content with some prominence. But that’s just one view of the content. OPML will be the skeletal structure that suports the multiple views. We may even try to implement checkbox news. ; )
You see, it’s based upon a feed-centric view of items, so if the feed is from the editorial department and has a category of sports, we know what it is.
If the feed is from elsewhere, we know what it is.
Ultimately this is based upon a web-centric view of things, not a company-centric.
Also of note, is the idea of groups, or groupings, as Stowe Boyd might more aptly describe them. AdHocracies, is the way I’ve put it in the past. Stowe hinted at this some time back with his blog trees idea.
The concept here is that a feed, or a blog and its items are a thread of expanding contributions. I’ll be using SSE to help that along. Two-way news is something I’ve been ineterested in for a while, but have yet to hammer down.
Greg Narain told me (at Ajax World) that it was too complicated and I agree that is an issue.
The only way this can work is if it’s transparent to the people using it. It’s just got to make sense from a user interface perspective right off the bat, or it won’t fly.
I know. Good luck.
A lot of this decision was based upon the ideas that the projectvrm folks have been throwing around lately. In addition to a distributed approach to the content, we’ll be supporting OpenID and other distributed identity ideas like XRI/inames. (like =matthew)
Lastly, along the same lines as VRM, is the idea of user attention data, but It’s way too early to tell in what ways we can make that useful. But i think we will.