Eclectic Brain Salad

Day 1 - Monday November 19th

This day started off slightly lukewarm - I was a bit hungover after a week of conferences, meetings and rock and roll, and Victoria station was totally knackered after the Queen deciding to throw a party (I was never much of a royalist, but now...) so I ended up being slightly late for the first session. I finally got to the really posh Church House conference centre, got my badge, and wandered in.

Dion Almaer and Ben Galbraith - The state of Ajax

It didn't surprise me to see the Ajaxian boys delivering the first keynote type speech, and it was a very high quality presentation. These guys obviously know the state of Ajax today better than anyone else does, and it showed. They talked about the different libraries available, how the technologies that combine to create Ajax apps have evolved, and other such things.

In the break after this talk, I was lucky enough to bump into such dashing gentleman as Bruce Lawson, Robert Nyman (love your sweater,) Jim O'Donnell, Christian Heilmann, PPK, Jeremy Keith, John Resig and Tom Hughes-Croucher. Much entertaining conversation followed (most of which is not repeatable here - thanks Bruce) and much coffee was drunk. By about 10am I felt close to human.

Mike Stenhouse - But I'm a bloody designer!

Mike's talk was nothing short of excellent - one of the best of the conference I'd say. It really highlighted the importance of making sure all of your team including the designers understand the basic principles of Ajax and how it can affect a web application. It affects the interface designers because of it's effect on accessibility, page loading, and the increase in possibilities for interactivity and interface effects.

Derek Featherstone - Real world accessibility for Ajax-enhanced web apps

I didn't catch all of Derek's talk, but what I did see was the same high quality delivery I've always witnessed him give in the past. It is so easy to render a webpage inaccessible using JavaScript; the Ajax subset is just as guilty. Derek showed some practical tips for keeping your web pages accessible even when Ajax is involved. We need to think about if we really need it each step of the way. If we do need it, we need to ensure that the application doesn't rely on it, but is merely enhanced by it. If functionality does rely on it, is there an accessible fallback available?

Lunch

The catering was very good, and included a very nice salmon dish, and some very rich chocolate cake, with cream drizzled over the top.

Stuart Langridge - How to destroy the web

Stuart's talk was by far the funniest talk of the entire conference, and also pretty practically useful! ;-) Stuart is at his strongest and most productive when he gets into a good rant about something, and this was basically what this talk was - a big rant about what to do if you want to destroy the web, and produce a bloated inaccessible mess...well, read this as "what not to do." He also called Christian Heilmann out for having Davey Crockett hair, compared John Resig to Kelly Osbourne, and repeatedly referred to the fact that we must use up all the bandwidth we have available by creating bloated web applications, otherwise people will just use it to download horse porn (with various slides full of farmyard pictures to back up the hilarity. Very entertaining, and also quite useful.

Christian Heilmann - Planning JavaScript and Ajax for larger teams

Christian is another good quality speaker that I've seen many times, and is always preaching good things about accessible JavaScript, playing nicely with others, and trying to work well with other JavaScripters and non-JavaScripters. The latter was basically the subject of this talk...in the context of Ajax of course. It was rich with tips on good code commenting to make sure others can understand what you are doing, planning your application carefully, using libraries carefully and making sure that people understand the basics of what they are doing for your application, etc.

PPK - Ajax at work: a case study

PPK obviously is a force to be reckoned with in the JavaScript world, and needs no introduction, so I was expecting a lot from this talk. He presented a case study - a royal family tree site - and talked about some useful stuff like how it was planned, why Ajax was needed for ths site, how it was made accessible, what data format was used and why, and a few other things. It was a bit marred however by the fact that he seemed to spend the majority of the talk going on about his interest nee obsession with history. It was interesting, but seemed slightly off-base. I didn't enjoy writing this, as I really like him and respect his work. A slight blip in a great career, I suspect.

Evening

So we wandered along to a bar close to the conference centre. Surprised? I bet you are. The beer was very cheap as in free, at least for the first couple of pints, and we hung around and chatted for a while - it was very interesting to talk to PPK and Wilfred Nas about their project, and also talk to Paul Duncan about his failed attempt at speaking Polish to the rather pretty barmaids (ok, should probably be "barpersons" in this day and age.)

We later headed out to a really nice Japanese place for a meal (myself, Bruce, Jim, PPK, Rob Nyman,) then back to the bar for a few more drinks. I ended up sleeping on Jim's sofa - many thanks for getting me off the street ;-)

Day 2 - Tuesday November 20th

Today I felt much better than the day before, which was good because this was the hardcore JavaScript day, so some of it was probably going to go over my head.

Brendan Eich - JavaScript 2 and the open web

Brendan's talk was pretty damn good - going over the point of JavaScript 2. He outlined what was going to change, and why, before showing some examples. He unsurprisingly included a dig or three at MS, commenting on their stance on ECMAScript 4. He also talked about the open web, and how we need to keep propagating it, or risk it being taken over by proprietary platforms like Flash and Silverlight. It was good to see him make self mentions of Opera, and our support for things like open <video> and 3d canvas. It was interesting to meet Brendan for the first time after the talk, and discuss the open web, and the fact that Mozilla and Opera seem to be standing side by side on some things, especially the ES4 thang.

John Resig - Building interactive prototypes with jQuery

John's talk was as expected - technical, full of code, and well presented. It was full of practical advice on using jQuery, examples of use, and it was very useful for the audience to have John there to ask questions to...Stuart Langridge made a very amusing point about the documentation being very hard to find things in.

Dan Webb - Metaprogramming JavaScript

Dan gave a very high end talk about metaprogramming JavaScript, ie creating applications so they actually alter their own functionality and code according to the context they are being used in. Very clever stuff, and this was one of the talks that went over my head the most ;-)

Alex Russell - Dojo 1.0: great experiences for everyone

I was initially a bit disappointed here, as I was expecting this to basically be the Dojo version of John's earlier talk. It sort of was, but he spent a lot of time at the beginning talking about some very abstract stuff. He finally started making sense after about 10 minutes, and talked about the aim of Dojo being to reduce complexity in web development - HTML/CSS/JS are pretty hard to learn well, and he thinks XMLis a bug, and the semantic web is a big faff. It is very hard to apply good semantics within web standards, although things like Microformats help. Semantics let you say what you mean and get what you want, but it is hard to do. I kind of disagreed with this point - surely it easy to appply good semantics in HTML, but many people just don't bother, or know what they are doing? He thinks that for applications, HTML and CSS are liabilities, and Ajax is a kind of get out of jail free card.

Alex's three ways to grow new semantics - standardisation (eg W3C and ECMA, which are byproducts of competition,) invention (eg JS upgrades or patches to pages,) and convention (eg Microformats, CSS design patterns.)

JavaScript semantics are disadvantaged, as you need a large investment for success. Dojo helps with this, and things like accessibility (it even supports ARIA) and i18n - it uses a tiny core and tightly controlled APIs. Ok, so not so disapppointed now, but it still felt a bit too esoteric; not hugely practical or useful, when compared to say, John Resig's jQuery talk.

Around this time I met Ann McMeekin, and some of the other really nice people from the RNIB.

Douglas Crockford - JavaScript: the good parts

Douglas' talk was very useful - he basically talked about what sucked about JavaScript, and how we can improve it by using what we have sensibly. JavaScript sucks because it has a lot of bad heritage from C, such as ++ -- operators, and switch. But it also has some really good features, like lambdas, loose typing, and inheritance. He also demonstrated some cool examples of inheritance, constructors, modeule patterns, and highlighted how bad globals are, before going on to talk about his JSLint tool. He proclaimed rather amusingly that it will hurt your feels as a JavaScript programmer, but it helps you to improve your bad habits by pointing out what is wrong. One example is that code style is argued about a lot, but it shouldn't be. This works well:

return {

false;

}

But this doesn't:

return

{

true;

}

JSLint helps you catch stuff like this. Finally he said that to fix JavaScript, we should deprecate it's weakest features, fix blunders, add new features that don't break syntax, and keep it simple and safe. We should add stuff like JSON support, and a safe eval. Another interesting point was made about perhaps adding a classical Ajax abstraction language for those who can't understand JavaScript.

Jeremy Keith, Brendan Eich, Stuart Langridge, Alex Russell and Douglas Crockford - Discussion panel

At the start of this talk, Bruce made the point that Jeremy Keith looks a bit like Professor Snape from the Harry Potter movies. This kept us in good humour for the last session. Jeremy asked the audience to put their hands up for some quick polls, before then reading out some atendee-submitted questions for the panel to discuss. One interesting point was that a lot more people seem to be using JavaScript libraries than when he asked the same question back at @Media 2005. He also asked if quite as many people would turn up if it was called @Media JavaScript - most of the room put their hands up. This shows that there were hardly any marketing type people at this event - they wouldn't have put their hands up, because they only go where the best buzzwords are.

I got Jeremy to ask the panel their feeling about cross device web development, and the iPhone only site madness, and what they think about cross device Ajax. The response was fairly mixed, but good, and it gave Opera a good mention. It was good to see that the general consensus was that there is no reason to shut anyone out, as we have the tools at our disposal. There were also other questions:

• Whether JavaScript libraries were going to become more standised? It is already hapening to a certain degree, and we will see more of it in the future.
• How will WCAG2 work with libraries? I didn't really catch this discussion, but it doesn't seem like it is going to be easy.
• With the client-side being so untrustworthy, how can JavaScript ever be secure? It can't really, but Alex Russell made the good point that security is never going to be a straight boolean "secure or not secure" - there are always levels of risk and mitigation. We just need to do the best we can. Douglas Crockford somewhere around here came out with the opinion that he thinks CSS sucks. Hmmm...interesting!
• Is JavaScript 2 too much like Java? Well, Brendan Eich obviously doesn't think so...another question was asked whether JavaScript 2 is just pandering to Google, cos they all seem to develop in Java ;-)
• Tom Hughes-Croucher asked how to get browser manufacturers to actually implement cool stuff in a consistent manner ;-) The answer my friend is to ask us about it. I'm working on it, ok?
• Are proprietary products and services still a hinderence to web development? Why not just use something like Silverlight? A bit of a deathly hush fell over the room at this point, but then Brenden Eich said that Flash is so ubiquitous that it might end up helping the future of the open web. I also about the fact that the ActionScript virtual machine has been open sourced - surely we are a lot closer to this than we might first think?
• Last question - is JavaScript still the world's most misunderstood language? Douglas Crockford answered a resounding yes!

This talk got a little strange towards the end, when Alex Russell decided to start attacking Yahoo for not being open source. This struck me as really odd, and it's the same kind of rubbish that we get aimed at us in Opera. We don't have an open source engine, but we think very open source, and the amount of documentation and tutorials and help forums and e-mail conversations we contribute to the community is surely a great contribution towards the open web? We also make a lot of our libraries and tools available as open source, and this is going to be on the increase in the future.

</body> </html>