Wednesday, May 31, 2006

Using the Blog as an online research tool

I usually don't like meta-blogging (blogging about the concept of blogs), because a lot of other people do it .

But today John Udell raised an interesting example of interaction between (in lack of a better general term) enthusiasts as a result of the openness of the web (all the web2.0 thingies like blogs and podcasts, their links, comments and trackbacks). This reminded me of that I dared to write "I've experimented with the use of a blog as an online research tool" in the Method-chapter in my thesis, and continued:
While it lacks structure and rigorousness of this thesis, the blog is still chronological through time, and in a way, it represents the research in a more honest way. It also performs the role of a dynamic research tool, as updated resources are available through my blogroll and linkroll.
Kristin Helen was the fellow-CMS enthusiast that opened my eyes for getting away with this (by pointing me to this document (1), and now she has in fact published an entire master thesis (2) about it! Congratulations :)

Now, how did the blog affect my research?

Primarily, it's a notebook. Its a place for me where I can note down stuff before I forget about it.

It's also a way to get feedback. A blog quickly attracts other enthusiasts due to the transparency of the open web (web2.0, ugh), depending on the activity of the blogger (see below). Like in Udell's example, I've found trackbacks to my own posts, whose blogs again have posts with interesting comments from other bloggers who I subscribe to and so on and so forth. I didn't get too many comments over the last few months, but I did get invited to a conference, and got some hints and nudges. It is also somewhat motivating to see traffic on yer blog, which means that somebody in the world actually take an interest in what you're researching for your master's (I know some other thesi that don't enjoy this interest).

How to do it?

To get into the blogosphere and get yer researchin' started, do the following:

  • Write. Do it regularly and write interesting/relevant stuff on one topic (use tags if you write on several topics). It's more tempting to subscribe to a stable blog.
  • Read. Use Bloglines or another feed-aggregator every day. As an example of the daily digest: these are my subscriptions
  • Comment. When you comment you leave a link back to your own blog, which is probably about a related topic. The author of the blog where you commented is likely to have a look at yours.
  • Trackback. Write about what other people write about. I doubt John Udell will notice that this post is a trackback of his post (he probably gets a hundred trackbacks every day), but maybe Kristin will? (no offense :D )
  • Search and surf. Check out technorati and for blogs on similar subjects.

So the thing you have to figure out is how big a limit you should put on your blog-activity. I spend half-an-hour reading through my subscriptions (comment when I find something particularly interesting, trackback when I find something immensely interesting), and write ~1 post a week, which takes 10-30 minutes.

(1) Mortensen, T., Walker, J. 2002, "Blogging thoughts:personal publication as an online research tool", Researching ICTs in Context, Intermedia/UniPub, ch. 11, p. 249-279

(2) Andersen, K. H. 2004, "Student’s Use of Weblogs: Weblogs for Collaboration in an Educational Setting" Department of Information Science, University of Bergen, Norway

Tuesday, May 23, 2006

InfoQ running on an open source CMS?

jmettraux noted that Alex Popescu is the technological architect of the new InfoQ developer community site and suggests that it is running on Magnolia. If this is the case, it is one excellent reference site for the Magnolia project. Alex commented that he will be writing about the architecture in some coming articles, but I can't wait with taking a few guesses. I think it's a mash-up of different systems (seeing they're gonna use Jive for forums).
  • There's not a uniform url-pattern on the links. One .jsp, one .html, some .action (although this need not mean anything)
  • The .action urls are very WebWork'ish, but could of course be JSF (event oriented). The use of DWR points towards no web framework in particular. Alex has also made a comment supporting the notion of WebWork/Struts Ti. On the other hand, the Magnolia developers are moving towards JSF/Facelets/Freemarker.
  • The is forbidden, so there is definetly some Magnolia stuff going on here :)
  • The structure of the content (given that they're not trying to confuse us too much with url-rewriting) seems to be divided into news and articles (content). The communities and topics seem to be meta-data search patterns, which could be using XPath queries to search through the news and content. Could they be using JCR-reference nodes to populate the rightColumn boxes and the categories? Just speculations.
All in all, the site is not too revealing on its inner workings. Allthemore interesting it will be when/if Alex writes something about it.

Sunday, May 21, 2006

Email and Content Management

Seth Gottlieb wrote a good post on Email and Content Management

I never miss the opportunity to evangelize the use of our KM-tools instead of e-mail (and sometimes even verbal contact). Some effective techniques I used at work:

* Refuse to do any task colleagues sent me by email, insisting that they issue a ticket in our task-tracker instead.

* Never reply with informative emails. Rather write a wiki-page on the subject and send the link to the correspondents.

* Each time a colleague seeks to explain how something works, or teach me something, I insist that they write a wiki-page instead, and send me the link.

We shouldn't abandon email. It's a quick and responsive tool that creates alot of dialog and content productivity. KM-people know that face-to-face meetings is the best way to convey knowledge. Phone, instant messaging and even email are closer to face-to-face than a CMS.

Any company should have set conventions of how different collaboration tools are used. Here are some example keywords:

* Email: Drafting and suggestions
* Forum: Discussions
* IM/Chat/voice: Quick notifications, brainstorming
* Blogs: External presentation for promotion
* CMS: Formal promotion, customer/user documentation

One could also divide the strategy into extra (public, open) and intra (internal company biz), and perhaps even a strategy for how communication with customers are done. We use a different kind of ticket-tracker for our customer-support (OTRS) than we use for internal business tickets (Trac).

Tuesday, May 16, 2006

The Paradox between Functionality and Extensibility

(Stupid blogger/writeley. Blog publish still not working.)

Apoorv's post reminded me of the extensibility requirements of a WCMS. While his approach seems to be a more pragmatic find-requirements-and-select-best-solution, I prefer the generic but extensible solutions (probably because I'm a developer/programmer at heart). Apoorv also differentiates between customization and extension, but I find it hard to draw a clear line between these two. Here's some lines on extensibility from my thesis:

The final and most important requirement of the WCMS stems from a single principle. It is impossible to predefine all requirements for a WCMS. Each year new concepts, ideas and methods are introduced to the World Wide Web, and web-sites must change the way they deliver content, content managers must change the way they produce content, and developers must change the WCMS to allow the requisite changes.

We propose that extensibility is the most important requirement of WCMS, because they have no definite set of requirements. A WCMS is an abstract information system, and users will not properly realize potential functionality before the WCMS has been put to use. The requirements are indefinite, and more functionality will be demanded as time goes by, and this is why extensibility is such a crucial requirement.

[On our old proprietary CMS]

Whenever there is a new requirement and the customer is willing to pay, the functionality is sown into the existing code-base of the project.

I suggest that the improvements of functionality reduces the extensibility of the product. More code makes it harder to read and use existing code, and every decision made in the certain case of one customer makes the module unusable in the eyes of another customer. For example, the customer had a very specific requirement that each web-shop product be editable by its distributor or producer. To make this work in our CMS, the products were simply divided by distributor, and correct access privileges were granted to one distributor user for her products, respectively. For another customer, the distributor-product scheme could be completely useless, and the entire web-shop would have to be re-factored for such a purpose.

The question becomes whether the case be any different with an open source system. One point was that the we avoided changing the CMS source directly, as any changes made here would be overwritten (or have to be dealt with) on rolling out the next version of the CMS. As a result the extension was entirely made outside the core source in its own module.

The restriction of extensibility can be avoided by having a central architecture which is unnecessary to change. Such an architecture would have to be generic to adopt to possible use-cases (like the web-shop for instance), and we hope that has been achieved in the JCR. The content nodes are flexible enough to be changed into web-shop products, and the developers are confident that they would withstand other WCMS extending transformations as well.


Here's some example extensions from our customers (removed this from the thesis):


An organization selling membership-articles used their website for displaying product information on their merchandise. At one point, they decide it would be better if merchandise can be bought directly off the site and require a webshop functionality built into their WCMS.

Web service

The Norwegian Association of Allergic People daily update their website with the most recent pollen-forecast. They realize their members would like historic information and require the page to connect to a repository of pollen-numbers, where queries can produce graphs of pollen levels.

Integrating external services

A trading company see the need to feed various stock quotes on to their front page automatically. A service reading information from another site needs to be set up and connected to the WCMS.


A news-company comes up with a new category of news which should be displayed for paying-customers only. Customers gain access to this privileged part of the WCMS by entering a personal user name and password on the site.

Integrating local services

A news company decides that its web page should also display information from a data sheet (which is being constantly updated) directly onto their site. The WCMS needs a data sheet reading service which flushes its content onto a web page at regular intervals.

These are just some example applications that can be implemented outside the WCMS, but sooner or later they need to interconnect in some way or the other, and the system will have to be modified to allow this connection.

Monday, May 15, 2006

Google Trends

Interesting application.

It seems that most searches are done on, which is a bit surprising (lotsa Danes on the CMS-scene, but they're still a small country).

Also somewhat worrying that the graph for content management is steadily sinking. We need more hype :)

Monday, May 08, 2006

Some writing tips

Having delivered the thesis, I've figured out how I'm going to use this blog.

I'm going to continue posting here, writing about web content management, but also about software development in general and other stuff. But the first thing I want to note down are my thoughts on how it was to write the thesis, sharing some writing tips with you.

Writing English

My first advice is: Don't. Unless its your native language, of course. I have had the pleasure of a bi-lingual upbringing, and seeing the English of many other Norwegians, I'm sorry to that generally, Norwegians write lousy English. Its not that we make grammatical or syntactical errors, but the feeling of the sentence is just.. wrong. And don't be fooled: Word or any other another spell-checker is not good enough. I've gone over another thesis which was already "fixed" by Word, and I still managed to empty a red-ink pen in the process.

Another point is that you don't *think* in English. You keep falling back into articulating difficult sentences in your native language, and then you have to translate it. This breaks down productivity, and you also get those Norwegian-English sentences.

If you still have to write English, make sure you know a native speaker who will go over the thesis for you and point out any mistakes and sentences you should reformulate. Train. Write English, and read English books aloud.

Writing flow

Write down what you think. Use notebooks, and get things down. You can always remove things later. Or don't delete things, just move them over to a "snippets" document, saving your thoughts in case you want to include them later (you could always blog about them).

Don't read your thesis too many times. You get tired of it. Write things once, and move on to the next section. Write through what you wrote a month ago. This will give you a fresh view of the content and allow you to

Turn off the spell-checker when you are writing. Those red-lines are also disrupting your writing flow.

Sentence rythm

Bad sentence structures easily bore the reader. Think short-medium-long-medium-short and so on.

Don't use too many "helping" verbs. The subject verbs the object. To will, to be, to have, are helping verbs you can put around the verb, but avoid this. Put the power of the senctence in the verb. Can't think of any good examples right now, but I might get back to that. Also, reduce the amount of adjectives, and put the meaning of them in the verb if this is possible.

Structure and Disposition

Write a disposition. You'll change it later, removing or adding chapters as you see how things are coming along, but a good disposition is the skeleton of your thesis. Write through it many times, each time adding a little more flesh to the bone (ugh, nice picture).

Document-driven development

Plan, and keep track of your changes. I kept a Todo-list. When I completed an item on the list, I moved it to the changelog . Each time I thought of something I should write about, I added it to the Todo-list. I also made checklist, where I added things each time I though "Oh. I'll have to see if that thing is done through the whole document". It was something like this

  • Use of modification/extensibility/extendability/customization or variations of these
  • Use of company/organisation/corporation
  • Use of "you", "I", "we" etc.
  • Upper Cases in Headings
  • Who am I? Programmer/developer/technician/researcher/author , we, I, our, us, me,
  • Who is the user? user/author/content manager/customer/visitor
  • A/applet, J/java
  • Check that all h3 are default formatting (some small ones)
  • Remove apostrophes' (Don't use apostrophes in formal documents.)

Red flags

Stay clear of these, or reformulate them:
  • very
  • little
  • think
  • I thought
  • easy
  • hard
  • seem

The reader

Learn the mind of the reader. I read this book which had some good points on how your thesis is read. For example, the reader wants to see a crystallized research question.

Any how, that's all I've got time for now. I've got some other ideas as well, but I'll have to add them later.