tfnico's blog

Just some quick notes here. This article (more like a memo) encourages an object oriented approach to creating content (reusable, etc), in hoping that the content elements can be reassembled by demand in an almost magical process. This is definitly what will become of content in the future (and sounds somewhat like Boiko's definition of content = information elements put together), but is still a bit abstract sci-fi. Maybe in a few years..

An experiment in making a CMS based on P2P technology. A good article, scientific. Several good points. Pretty technical. The obvious benefit of having a distributed repository is scalability, but it is suggested that content ownership/quality is increased since the content is local. The authors make a point of using open standards, though not a very good one (java and XML). JXTA is one of the more exciting and relevant standards. The experimental implementation of the system had not been completed at the time of writing, and I can't find any continued effort on the net. Bit of a shame actually, sounded like a very reasonable implementation. A good idea and a good architecture, but I guess our centralized servers are still strong enough for handling all content.

Subtitle: Does Systematisation of the Knowledge Base have a Positive or Negative Effect on Organizational Culture? The article starts off with a pretty stupid question. Of course, I would believe the answer most of the time is yes, and the conclusion of the article doesn't differ. The article considers two sorts of KM/CM software: Workflow systems and CM systems. Funnily enough, XML CMSes is the exact term, and already it is shown how the author likes sticking his finger into irrelevant technical details (so what if the CMS is based on XML?). However silly the purpose of the paper, there is a nice view on the value of workflow systems (cause process is knowledge too). All in all, this article says nothing else than "IT affects orgnizations! Wow! Even positively most of the time!" and is pretty worthless.

Also a very unscientific article. Very subjective. The bottom line here is that content should be written independant of languages (good luck), or at least make accessible enough information so that content can be translated without loosing value. A finger is pointed on several issues in translation/internationalization/localization that you might not have thought about, but all in all, this article is thin, subjective and not very useful.

A nice intro to CMSes. Not a scientific article, but easily read. Plone is similar to Magnolia in many ways, more powerful in some areas, weaker in others. Very content-specific domain model. Can not be abstracted into anything else than content. Nice that the CMS has themes, as well as new-content-wizards and template editor. They call wiki-text structured text . But why wiki when you can wysiwyg? :) Overall, nice intro to Plone, and some good ideas for WCMS in general (requirements). I could've used Plone/Zope as a candidate in my comparison, but my scope is too small I think (will stick to the java ones).

Interesting material to use for analyzing the value of a WCMS. Might use it myself for comparing the candidates. The Analytical Hierarchy Process (AHP) is a set of principles or a framework for evaluating investments. To soften the subjectiveness of the framework, a wide range of sensitivity analyses are performed. A framework is presented, a bit similar to the one in Management Challenges of the New Infrastructure, only less complex (Input -> Process -Output). The framework is mostly applicable for doing pair-comparisons, so the candiates should be reduced to a handful before doing the benchmark. Funny how this article regards WCMS to be what most thinkers consider to be ECM.. Overall, this article uses alot of text to get to the point, but the framework presented is quite good, and I will probably use it.

Today I'm taking the train back down south for my summer vacation. Since it's been a while since my last post, I'll do some more thoughts here before I'm off. I am planning to spend a good bit of the summer reading through Bob Boiko's CMS bible, as well as putting together a scheduler-application with JSF, and perhaps glue it together with Magnolia. Will also be making some much needed custom Magnolia templates. I was reading the CMS bible the other night, and I read something about sorting content in to binary and "normal" files. Binary files being "all 1's and 0's", like pictures and media files, but also files of a closed format. A word file or an Excel workbook is a binary file. During the course of knowledge management this spring, I learned that heaps of information which contents are allready at an atomic level (there is no easy way to aggregate elements out of a word document), we call unstructured data . An email, which content c...

A asked me to write an abstract for this conference . It seems to be very KM-centric, and I know that KM people like to distance themselves from CMS-stuff, so I'm not sure they'll like it. KM-consultants like talking about CMS systems as if they are trivial pieces of sotware which were perfected in the 90'ies, but they aren't. There's still a lot of work do be done in that camp, and KM folks need to acknowledge this effort.. Anyhow, the paper can be found here .

Got some valuable links from A on how to evalute the ROI of implementing CMS, among other things: Merant report on CMS ROI IT strategy and WCMS Also a very interesting detailed description of java implemented "knowledge center"

Looking through som old blog postings done by Martin Fowler (whose posts I find most interesting, both in an academic and practical way), I found this one . I consider building on those thoughts when defending my thesis' approach. Martin Fowler asks whether open source projects can be actual R&D. I ask (overstating the problem): Do all the brittle R&D efforts done through students' master thesi need to be 100 pages of paper theory?

According to one guest lecturer at a course about knowledge management, Åsmund Mæhle, their first attempt at creating a web portal for governmental CoP's failed because they tried to sort information in a hierarchal manner. Eg. they would put everything related to one municipality in under one node, but later failed to connect it to other municipalities having similar practices. Their solution was to have a map-like, associative structure of information: Concepts and relations (a hierarchy typically only has one sort of relation: is-child-of). Of course, an associative directory will be less structured, more difficult to implement and (arguebly) harder to maintain. Certain websites require a rigid structure (legal documents, laws and rules) to guarantee no navigational mishaps or confusion. My imminent conclusion is that larger sites, especially the portals, needs to declare certain parts of their structure an associative one, leaving a strict core with important documents in a hie...

Spring 2005 - Gather requirements - Develop solution Autumn 2005 - Deploy solution - Extend functionality Spring 2006 - Put together feedback, notes, reports - Results in a thesis

As I presented my thesis, I had to define a scope. The world of CMS is a large, complex and deals with multiple scales and platforms, ranging from global enterprise content management (with proper intranets, DAM, DRM, and the lot) to small single site web publisher software. To reduce the scope I laid my approach to the following areas: -One platform (java) -One market (open source) Focusing on these two, the second in particular (open source can mean many things) can be risky, but it's better than enveloping the whole CMS business. The other day I was contacted by Benjamin Mestrallet, founder of the eXo platform, as he had naturally taken some interest in this blog. I realized that eXo fits quite nicely into the scope of my thesis. This means that I will have to analyse the eXo platform in the same way I have started off with Magnolia. I know there are some more open source WCMS around (http://java-source.net/open-source/content-managment-systems), but I try to remind myself that ...

I presented my thesis (as far as I've got) today for A and the rest of her students. The presentation (available on my workspace) went fairly well. I was, however, vague on a couple of points: The approach which I'm going to use (not just explicit, but a theoretical approach). What sort of feedback will I get from the testing sites? Quantiative and/or qualitative? Quantitative will be in form of statistics that we have already. After setting up the new WCMS, we can measure if activity on the webs of our customers increases. I can measure the amount of support we do for our customers. I can do qualitive interviews with the users before and after. What sort of theory will I produce? -Compare to excisting CMS theory -Compare to portal theory -Add my own theory!

This is a from an internal discussion at work.. Jahia seems nice on the demo, but they are quite outdated on the open source part (using old stuff like struts and ant, no JSR-170). Same thing goes for the very popular Liferay. Swedish SiteVision is also good, but Objectware have alread got the under their hood. Magnolia as well, but the (like us) are based on persisting the content with Jackrabbit, and at this point there is no DBMS support there (though this is right around the corner). Magnolia's developers, the commercial company Obinary, are offering their customers an other JSR-170 compliant repository: CRX from Day. I suspect that the CRX is heavily based on Jackrabbit, since the Day developers are the core of the Jackrabbit crew. Here's an incomplete overview: Company Open Source produt Commercial product ------------------------------------------------------------- Obinary Magnolia Compass Day Jackrabbit CRX S...