Tuesday, 11 September 2007

A Web of Data

Returning to the presentation made by Tom Coates, in my last post I extracted the "Your website is not your product" idea as that struck a particularly resonant chord. But there were plenty more ideas.

On the subject of meta data, Tom seemed to be in favor of retaining as much as possible (particularly if available for free as part of an automated process). You cannot easily second guess how your users will search and sift through your data, using Flickr as an example it was demonstrated just how many different ways that the photos can be sorted, obvious ones by user or tag and less obvious by Camera make/type and various camera settings when the photo was taken.

Tom also highlighted the idiocy of the taxonomy vs folksonomy arguments. Both have advantages so why not provide both, why not throw in parametric searching and anything else as well? This kind of reminds me of the way that small children absorb certain concepts. Ask them do you want X or icecream and they will see this as a choice that has to made. Once a certain age is reached and their mental processing of concepts they are familiar with improves they will think for a second and ask "errr can I have X and icecream please?".

This boagworld.com podcast talks briefly about Dconstruct07 in the news section towards the beginning (also has good things to say about Tom's presentation. Listening to this reminded me of a couple of things that I had forgotten or not noted, I think when they appear the podcasts of the Dconstruct07 will prove very useful for review. For now I have put the feed for the Dconstruct07 podcasts in the left column of this blog, but so far it only contains three pre-conference podcasts. Incidentally as a developer who is also interested in the front-end and user experiance I find the boagworld podcasts very helpful.

No comments: