Hi Group! Since I started working on a newspaper website upgrade two years ago - I've found a lot of good advice around here. So I thought I'd make a write-up on what I've done, as well as pose a question.
My workplace, Fria Tidningar, had a setup with a publishing system called Scoop, and an automatic web export to PHPnuke which was really starting to show its age when I came here 2.5 years ago. The company then published 4 newspapers, today 8, with different premises (4 of the papers are local) - but some shared content.
After evaluating different CMSes (primarily Drupal, Wordpress and Joomla) I decided on Drupal, and started pounding away on an automatic export system. The Scoop system was very limited in that it only provided exports in its own custom xml-format (with lots of formating data stripped away) at that time. These days it also provides a mangled version of NITF (though I have not had any closer look at it). Anyway, the Scoop xml-format is a tad unpredictable - so I had to make a perl script to parse it, go look for attached images, and write out a more NITF-ish xml-document more suitable for consumption by Drupal. In hindsight that may have been a bit stupid, with some duplication of effort as well as an extra step in the export chain - but I really had to parse the xml anyway to find attached images before upload, and we were at that time looking into the possibility of selling a newsfeed service, where we would be using NITF.
For the first Drupal (v5) installation, I coded a custom import module as well as a custom theme (based on Garland IIRC, but with pretty much everything changed). Modules we used were Image & Image Attach, Domain (for syndication between our related papers), Views, CCK and Taxonomy. CCK was used for extra text fields, and Taxonomy for sections and stuff. For the byline, I created a new user for every new name during import and asigned ownership to the user named exactly like the byline. This turned out to be a pretty bad decision in retrospect (the user table filled up with names of external writers, badly spelled names, names + descriptions, etc). The site worked though, and was put to use around march 2008.
Around february this year I started porting the entire site to Drupal 6 and made a new take on the news import. Apart from better coding practices, the bigest changes were to image/byline/owner handling. I made byline a taxonomy-field instead of autoregistering tons of users, which has greatly improved usefullness (the byline is of course clickable). I still autoregister users, but now it is the editor that entered the text into the publishing system - so when that happens it is because we hired a new editor. This also has the effect that an editor automatically owns all of his/her texts in Drupal. As for images, I still use the old Image module - but attach the images using CCK nodereferences. I hacked Image assist to make it add an entry to a noderef field, and display the icon beside text fields. I also coded a custom menu to help with things such as subscription forms.
As a starting point for our theme I used the Basic theme - which proved very nice to work with. I suppose I have bastardized the theme thoroughly by making it load images when provided noderefs, and opening the correct section menu. Still, it works fairly well.
In the last update, I added the WYSIWYG and CAPTCHA modules - in order to make both commenting and article authoring easier. And looking forward, the Drupal installation is starting to look like a good place to centralize the editorial workflow.
This last piece started to be a bit pressing when the distributor of our publishing system canceled our support contract without having made good on their delivery of the entire system. What we have is a file system-based administration system, as well as a number of quark-plugins to manage different aspects of the workflow. We are thinking of moving the workflow management into Drupal, and export the articles for Quark/InDesign/whatever.
Has anyone here tried using Drupal as a newspaper publishing hub? Do you know of any usable Quark/InDesign export plugins? (Or xml-export coupled with Quark/InDesign import plugins?)
Cheers!
/Filip
PS If you don't mind the Swedish, you can check out some of our sites at:
http://www.fria.nu
http://www.stockholmsfria.nu
http://www.goteborgsfria.nu

Comments
InDesign import plugins are
InDesign import plugins are basically a non-issue - the software works with any kind of xml. Export should be doable as well - I currently do it with a small module that acts on hook_nodeapi and updates an xml file every time a node is saved, which isn't too hard to make, but you should really look into modules that expand Views to provide xml output. I've worked with Views Datasource in the past, which was an ungodly mess, but there might be other modules around by now that do a better job of it.
I very much recommend the web-to-print approach above exporting from print to web. Your web content will always be more structured and thus easier to handle and import into InDesign than whatever you can export from InD.
Information.dk
Hi Filip,
http://www.information.dk has been using Drupal for quite some time now. 3 or 4 years I thing.
I found this wiki for modules for newspapers:
http://groups.drupal.org/node/5281
Last via a cached page of http://theopensourcenewspaper.org/ found at the bottom of information.dk
on google:
http://74.125.77.132/search?q=cache:sQutMRHX2cQJ:theopensourcenewspaper....
Spred Glaede :)
Flemming
Even a turtle reaches it´s goal...