12.10.07
Flickr Upload DLL
Symbian have published my paper and code for a DLL which enables a Symbian application to get permission to upload images to a Flickr account via HTTP.
There are test apps which upload photos on S603.x and UIQ3.0.
12.09.07
Visualising large code bases using Treemaps
In a previous posting, I desribed a method for extracting LOC (lines of code) statistics from a large code base.
I’ve been playing around with Universtity of Maryland’s treemap visualisation tools. For those who don’t know what a Treemap is, the site has some good video tutorials or just google. There is a little bit of a learning curve to understanding and using them
Tremaps were invented by Ben Shneiderman (yep, same guy who created the Nasty-spiderman diagrams as we used to call them in school). Shneiderman wanted a tool to quickly visualise the large clusters of files on hard disk to find out where his space was really being wasted.
The use of Treemaps have extended to other domains with varying degrees of success, and my interest in treemaps was recently rekindled on seeing Tim O’Reilly use Treemap to visualise trends in the book market segments on his blog.
If we group LOC counts for each file by their directory hierarchy then it’s possible to visualise the components of a System in terms of their LOC contribution to the overall code base.
It’s a fascinating exercise. The picture below shows the Treemap for a multimillion line code base. (Note that I’ve anonymised the data by taking an MD5 of the strings, so it looks a bit less friendly than the real view).
12.04.07
Unskilled and unaware
I recently read this great little psychology paper on skill, and the difficulties in self assessment of incompetence.
http://www.apa.org/journals/features/psp7761121.pdf
“We propose that those with limited knowledge in a domain suffer a dual burden: Not only do they reach mistaken conclusions and make regrettable errors, but their incompetence robs them of the ability to realize it.”
Some of the observations seem obvious, for instance the paper asserts that skills which make someone competent in a domain are often the very same skills necessary to evaluate competence is that domain. In many fields the lack of meta cognition means that they are likely to have incorrect/inflated ideas about their own level of skill in relation to their peers.
As with all lab based psychology experiments, it is important to take care when extrapolating results to the real world.
A good example of false self skill delusion might be that an employee who joined a company at the same time as another may rate their skill to be similar to their peer rather than based on objective skill assessment.
For me, an interesting section of the paper is the section on “incompetence and the failure of feedback”. It says that it seems inconceivable that unskilled individuals manage to go so far without getting negative feedback about their skill. But I have found that objective skill assessments seem very rare outside academia and negative feedback even rarer.
I’ve always seen negative feedback as a useful self correcting measure for individuals. For instance if there is a guy in my department that I don’t know what he does with his time, I will say that.
The fact that negative feedback is often relayed by passive aggressive means such as lack of promotion, rather than direct feedback just opens the door to false attribution such as “I was having a bad period”, or “the boss preferred the other guy” – thus not really allowing the feedback to be accepted for self improvement feedback.
The word “incompetent” carries such a stigma, and the paper is careful to point out that it’s not a general assessment of a person, but incompetence refers to a domain specific skill/understanding.
The sorts of domain where this over estimation of skill might be prevalent are language skills, programming, problem solving skills and communication.
One thing that the paper doesn’t deal with is the result that reading the paper might have on one’s future self assessment.
12.02.07
Commoditisation of services
I had an interesting thought the other day while contemplating the mash up culture.
Every service is out there pushing their APIs and championing efforts from the community on innovative use of APIs.
APIs
Most APIs have some sort of limit associated with their usage. For instance a maximum number of requests per day, or a bounded dataset (e.g only return the last 100 comments).
The limit exists in order to oil the tracks for casual developers to innovate by using the APIs without having to sign contracts and service level agreements. If a developer came up with something which was popular which required much more bandwidth, then usually there is a fee/royalty to pay.
Using APIs for data migration
Consider a service which was once innovate but now is more of a commodity (like a gas supplier, or electricity). Let’s consider photo storage.
I’m currently with Flickr, but wizzy service fadphotoservice3.0 might come along and offer more or less the same service, but for free.
What’s stopping me from switching? Well, all my photos, my tags and my comments on Flickr. It’s far less hassle for me to continue paying Flickr.
But if you look at what’s happening in the UK, energy suppliers and banks go out of their way to make it easy for you to switch accounts. They know that the Griffin Saver account which you set up as a kid is probably the one you will have for life unless they find a economic for tempting you- usually an incredible rate.
If a service which offered everything that Flickr does for free, and offered to migrate my data (including preservation of comments by others) by having me authorise its use of the API, then it would remove most of larger obstacles for switching.
DRM on the desktop means that migrating music and videos is hard. But for the web, it looks like they are making it easier which is good for users.
If embracing the mash up culture means removing the API limits and spending lots of infrastructure, are incumbent services opening themselves up to the competition?