All entries for Friday 09 November 2007
November 09, 2007
So, now the sessions are over, what kind of things have stuck in my brain? There are a few key themes that I’ll be taking home:
1) Commoditization: All sorts of things that used to be distinctive features, are now becoming commodities. There’s no point expending your own resource on things that you can get from somewhere else, be they hardware resources (ec2/s3/joyent etc), or data (flickr/gmaps/facebook/every other site with an API, or identity services (openID/OAuth)). Re-use what you can, and devote more effort to the things that make you unique. Bear in mind, though, that there are always drawbacks to using someone else’s stuff; make sure you understand what they are, and what your escape strategy is, before you commit.
2) Focusing on the user experience: And, perhaps more importantly, focus on what the user wants to do. Our users don’t, by and large, want to make web pages. They want to do fabulous research, or be recognized as great teachers, or build their reputation as a brilliant journalist or fascinating blogger (is there a difference?), or whatever.
I’m not sure I really understand the things that our users do, and the reasons why they want to do them, well enough to really deliver the best user experience we possibly could.
2.1) Look further afield for ways that the user experience can be improved. Several people have pointed out the synergies between the user-experience elements in game design and the web space; looking at some of the elements around progression that games feature, would probably benefit some of our applications.
3) There’s a world of Federated Identity evolving quite separately to JISC’s Shibboleth roll-out. Fortunately, we don’t have to choose between the two; it seems as if we should be able to have both.
4) Good operations, ( assuming you’re not just handing the whole lot over to amazon :-) ) is an engineering problem. If you want to develop a good operating environment, the people you need are developers; Sysadmins are customers of this development process.
5) Mobile Stuff. Mobiles are the obvious extension point for web 2.0, but there are various different takes on this. On the one hand, you can look to serve upcoming smartphones, which are basically fully-featured browsers which just have a slightly smaller screen than usual. Or you can target the millions and millions of PAYG phones with little or no web connectivity, just basic SMS and maybe WAP. Or you can go for the middle-ground, with very limited HTML and low-bandwidth connections. Whatever you go for though, the interaction model is very different, with people looking for short, limited-attention, result-focussed interactions rather than the long-running dedicated attention that you get when users are sat in front of a PC.