« January 2006 | Main | March 2006 »

Online Summative Assessment

Good podcast here about delivering summative assessment in the University of Vermont. Specifically about delivery on Blackboard simultaneously to a large number of students. Information on question import(Respondus), network used(wired), load testing(essential, start 16 students at a time), security(secure browser), analysis (extract info from xml), multimedia(real need for audio and video), question randomization and unexpected problems caused by heavy simultaneous use of the content system.

Many of the technical challenges had roots in heavy client/server interaction. That's a problem with architectures that treat the client as a dumb terminal and do all the processing on the server. Assessment delivery systems can reduce network errors and improve client responsiveness by embracing the rich client model - the assessment is delivered at the start of the exam and the results collected at the end. Simple network interactions like those are reliable even for large numbers of simultaneous connections.

Posted by Alexander on February 24, 2006 | Online Summative Assessment | Comments (0) | TrackBack

Jobs not recruiters

Most people, especially in tech, have wrestled with job sites at one time or another. They've been around since the beginning of the internet which makes it all the more surprising that they still fail so badly. This Slashdot discussion gives some great ideas for improvement. Many comments say job sites should disintermediate recruiters. Looks like an area ripe for a Web 2.0 solution.

Main points -
Actual jobs rather than recruiters.
Direct contact between the employer and employee.
Some way to penalize fake jobs / inflated resumes.
Feedback on hiring companies from employees.
Specified salaries.

Posted by Alexander on February 21, 2006 | Jobs not recruiters | Comments (0) | TrackBack

ClickRank to replace PageRank?

Matt McAlister asks what's going to replace PageRank.

The hyperlink was a vote in the search-driven Internet. Now I'm dependent on a new currency - human action. The click is much more potent than the existence of a link.

This is a logical progression and Google is already on top of it. Google Analytics is a free service for websites. The webmaster puts a piece of Javascript on each page and Google records each pageview. It provides the figures and some analysis of the traffic. Google gets the same data. The more websites that add Google Analytics, the better its view on which links are being clicked and how users are moving from one page to another. With this information, it can rank based not only on the number of clicks and page views, but the amount of time a user spends on a page. That's the supreme indicator of the value of a search result.

Posted by Alexander on February 20, 2006 | ClickRank to replace PageRank? | Comments (1) | TrackBack

Don't Be Evil = Be Trustworthy

After a hard disk upgrade last week, I took a leaf out of Jon Udell's blog and started routing all my mail through Gmail. It's a nice solution, it allows me to funnel all my email from sexdecillion different accounts into a single box. It lets me send from my own domain name, and I trust that Google will let me export my archives if ever want to switch to a different service. Trust is an important word for Google. You can understand nearly everything Google does by reference to that single word.


Google must have been watching the scene in Minority Report where Tom Cruise walks into 'The Gap' and his eyes are recognized by the computer system. It greets him 'personally' and has recommendations for him. There's a similar scene in Total Recall, where a billboard recognizes Arnold Schwarzenegger and presents personalized adverts. I think that's the end in a game that's just beginning now. It starts with personalized ads on the web and with email, moves onto personalized ads in the living room, and from there onto personalized ads in public spaces. None of these steps can be forced though, they are all permission based, all depend on trust.

Part of this trust must also come from the fact that they have two postgraduates running the show, rather than career CEOs

Consider Google's recent refusal to hand over information on searches to the Department of Justice. I trust a company that takes privacy that seriously. Consider Google's recent entry into China. I trust a company that levels with me if it is censoring my search results. (Then I can go and try to use the real Google China search engine). This is all in marked contrast to Yahoo! who seem to be falling over themselves to send search results to the DOJ and seem to have a fax hotline to Beijing to send the names of dissident journalists. (They have to use fax, as e-mail wouldn't get through the firewall y'see).

So I've taken the first step along the Google Web 2.0 offering because I believe they're trustworthy. Part of this trust must also come from the fact that they have two postgraduates running the show, rather than career CEOs. Google is just not your typical corporate citizen.


Let me balance all this a bit with some of my other experiences of Google this week so that you'll know I'm not just some zealous Google fanboy. I installed the 'Google Pack' to ease the burden of constantly updating software. This install was on a fresh Windows XP machine. Google Earth didn't work at all. Google Toolbar for Firefox cause it to freeze for half a second every two seconds. It also installed Google Desktop, but I didn't really know what that was doing, and then the EFF told me it was sending my documents to Google so I shut that down double time. The other applications were mostly other companies applications. I conclude that Google isn't much good at desktop software no matter how many Ph.Ds they've got.

I also tried to subscribe to Google Analytics to try it with Flash and to see the level of integration with Google Adwords which I also use. It seems like it's very oversubscribed, with nobody new being taken on since the middle of January.

Posted by Alexander on February 13, 2006 | Don't Be Evil = Be Trustworthy | Comments (1) | TrackBack

A Google Nobody

This comment on Slashdot gave me reason to stop and think today.

noted physicist Dr. Franklin Felber will present

"Franklin Felber" [google.dk] has less than 40 hits on google. For that reason I very much doubt he is a noted physicist. By association, I am not going to take his claims seriously...

What Slashdot geeks are doing today, everyone will be doing in 5 years. And less than 40 hits on google will certainly not make you a 'noted physicist' ...and less than 5 hits may mean you're not a noted anything.

note - 19/02/06 People are arriving here after searching for 'Franklin Felber'. I don't have anything to say about FF or his ideas. For all I know he's the next Einstein. This post is about how people are using the web to assess credibility.

Posted by Alexander on February 11, 2006 | A Google Nobody | Comments (3) | TrackBack

Cash Value of Website Content

"Content is King!" is something smart people have been saying about the web (even before the Nasdaq gave everyone a right kick in the nads in March 2000). I've been pondering today the exact value of content. Content is like a capital investment - you invest once and you generate a return on the investment over time - return on cash is about 4% at the moment, property about 6%, the stock market about 11%*. By measuring the return on the investment of content, it should be possible to place a capital value on it.

Specifically, I've been working out the exact dollar value of having a quiz on a website. I host something called 'The Aggression Questionnaire' on-line as a sample - it's been established at the same URL for about a year and it now attracts 600 users per month - not a huge amount, but it's not original content and is available elsewhere. People who use it often locate it using keywords like 'aggression', 'aggression questionnaire', etc.

To put a cash value on those eyeballs, let's compare what it would cost to buy those visitors from Google. Some keywords, like mesotheliona will cost $84.08 per click whereas it might not cost so much to buy a keyword click on a phrase like 'free ways to get free stuff for free'. For a word like 'Aggression', it's not heavily contested, so we'll say it's at the lower end, about 5 cents. So the calculation is 600 (clicks) * 12 (months) * 0.05 = $360 per year.

That's only a dollar a day. It doesn't sound like a lot, but what do you call something that generates $360 a year? Capital. How much would you need in a savings account to generate $360 per year? About $9000. Depending on what kind of return on investment you think content should make, its capital value is around $3000 to $9000.

For the calculation to hold, the keywords have to be the kind of keywords that you think you'd actually buy. The Aggression Questionnaire isn't worth that much to me - I just put it up as a sample. I don't have anything else to offer visitors interested in aggression. But if I ran anger management courses for Corporates with employees that were busy beating each other into a state of non-productive inertia, then that's exactly how to place a value it.

So what quiz or questionnaire could you write this afternoon that would attract visitors interested in your product or service?

* 11% over time. Or so they say. Typically these calculations don't include stock markets that disappear entirely, like the stock markets in Eastern Europe before World War 2. If you'd invested £100 in the Eastern European market in 1930, you'd have precisely nothing to show for it today.

Posted by Alexander on February 06, 2006 | Cash Value of Website Content | Comments (1) | TrackBack

Settlers of Catan

I've been playing a lot of a board game called 'Settlers Of Catan' recently. I picked it up over Christmas following some recommendations on a thread on The Business of Software. I'm not a big fan of board games usually - most are equivalent to Ludo, with no decisions to make - or others like Monopoly where learning a simple strategy will suffice for the whole game (buy everything you land on, if you really need to know).


Settlers of Catan really allows for some tough strategic decisions. More than that though, it heavily rewards successful trading with other players in the game. It's a zero-sum game in the sense that the're only one winner, but the transactions are truly non-zero sum. There's elements of negotiation, renegotiation, influence, persuasion, concilliation, reconcilliation, trade, co-operation, trust and of course, competition - everything that should exist in a multi-player game and usually doesn't. I mention it here because I think it's truly educational as well as entertaining - it provides an opportunity to practice and hone skills that can usually only be practiced in situations that have bigger consequences than losing a game or two.

Posted by Alexander on February 02, 2006 | Settlers of Catan | Comments (2) | TrackBack

Blog Changes

I've established a separate blog for Question Writer. I've been worried for a while that this blog has been getting too commercial.

I used to subscribe to Hugh McLeod's blog. It turned me onto Seth Godin and the whole Cluetrain Manifesto. I've a lot to thank him for. (Thank you, Hugh.) But then he started pimping a wine called Stormhoek (even sent me a free bottle). He blogged it incessantly, and every time I saw the word 'Stormhoek' a little part of me knew I was about to lose 15 seconds that I'd never get back. So I unsubscribed from the whole feed to avoid it.

I don't want that to happen here. So for Question Writer users, who want to know about every new feature and new bug fix, there's a blog for that now.

Posted by Alexander on February 01, 2006 | Blog Changes | Comments (1) | TrackBack