@media, day 2

Professional front-end engineering
Nate Koechley
An interesting, but fast-paced presentation from Nate Koechley of Yahoo. Its always really interesting to hear how 'big sites' manage work. The very title of this presentation also rings a bit of a change - the fact that there is more of a requirement emerging to fill that specialised knowledge gap between designers and back-end developers.
Nate mentioned how the team supports 'grades' of browsers, layering the 'richness' of the user experience according to the technology each grade supports. This enables the experience to degrade gracefully for older browsers, yet the content remains accessible to all. He promoted the use of tags such as 'cite', 'caption' and 'thead', proper use of declorations and inobtrusive javascript. Other interesting tips which really drive home the level of optimisation these popular sites require, included the method of horizonal alignment of CSS sprites rather than vertical, as this seemed to result in a much faster load time. Also, avoiding gaps between icons has a positive effect on optimisation. Oh and don't use @import for calling CSS files.

Building on the shoulders of giants
Jonathan Snook
A presentation about APIs/mashups/frameworks. Jonathan talked about how developers try to re-invent the wheel whereas designers re-use. With that in mind, he suggested that developers follow the 're-use' approach and look at what's already out there and established. He touched upon some of the main frameworks available.

For example: The Guadian & Dopplr
Marc Pacheco & Matt Bidulph
Marc focused on challenges the Guardian faces with CSS management. Due to the variety of layouts, many CSS files are used. In the development phase, the files are kept separate for easy maintenance and are called using @import. These are then merged when the site is published. He mentioned that in his experience researching CSS management, there was very little information available and very few tools that aid the control and maintenance of large numbers of CSS files. Particular issues discussed during the Q&A, were the locating of specific code in large files and the removal of redundant classes etc. Hmmm - am there are tools out there for that. I attended a Microsoft presentation on Expression Web and remember quite a feature was made of CSS management. Must investigate further...
Matt Bidulph focused on the use of APIs for his talk on the development of Dopplr. He made the point that an application is more interesting with other people's data in it. Dopplr uses a number of APIs to combine features and data. He mentioned how it can integrate with Google. Pretty interesting, and I can see how it can work for social netwoking sites, but not sure you would want to rely on third party data etc for business critical purposes. To ensure total control, you kind of need to re-invent the wheel to a degree, I guess.

Communicating best practices
Rachel Andrew, Paul Boag, Patrick H Lauke & Murray Rowan
I was a bit disappointed with this session, but I think that may have been because it covered similar points from previous years. Only a problem if you've been to a few of these events. And I think they are probably important points worth covering. Much of the session discussed how to present good practices to stakeholders, how to make them a selling point, backing up the reasons why. Or whether we should just follow this best practices anyway as matter of course. I would say a bit of both. Another point covered, was how did the panel handle clients that insisted on doing something that went against the grain. Answers to this was to inform the client of the reason, backing it up with articles from third parties if necessary, then making a point that you would perform the task but it was against advice and that once the project was completed, you would not wish to be associated with it/wouldn't want to include it in your portfolio. This often surprises clients, enforces a sense of integrity and drives home how much you disagree with a principle.

Richard Ishida
Global design: Characters, language and more
An interesting presentation discussing character encoding (never been quite sure of what all this truly meant, but have a vague reference point now!) He explained the importance of declaring the language in the header ie. lang="en" or xml:lang="en" and how Unicode encoding UTF-8 is favourable. It certainly explained why random characters might appear if the correct encoding wasn't declared. I will start looking at headers in websites to spot this now!