I didn’t take any notes, but JQuery looks pretty cool. Maybe even cooler than Prototype and Scriptaculous!
[tags]barcamplondon3, jquery, javascript, prototype, scriptaculous[/tags]
I didn’t take any notes, but JQuery looks pretty cool. Maybe even cooler than Prototype and Scriptaculous!
[tags]barcamplondon3, jquery, javascript, prototype, scriptaculous[/tags]
This was a hands-on session playing with lo-fi flip-book animation, based on the techniques on Cathy’s blog. Here’s my very poor effort:
[tags]barcamplondon3, animation, silly walks[/tags]
It’s all here!
[tags]barcamplondon3, lifestream, yahoopipes[/tags]
HTTP not ideal for asynchronous stuff (“Are we nearly there yet?” “No.” “Are we nearly there yet?” “No.” “Are we nearly there yet?” “No.”). Message-passing architecture is better (“Let me know when we’re there” … “We’re there!”).
Also useful for clusters, multi-core machines etc.
Erlang (also Scala, Haskell, E etc) are becoming interesting because they operate using messages. Not that different to pure OO. No shared state. Scalability doesn’t come for free, but it’s much easier.
Spread, MQ, JMS etc: publish and subscribe. Stick a message on a topic or queue and any subscribers receive it. Allows very loose coupling. One-to-many, with publisher not needing to know about subscribers.
Comet allows a connection to be held open between browser and server, rather than having to poll.
Jabber can be used to return future results for a search (for example). “There will always be more pictures of kittens.”
Worker queues remove expensive calculations from synchronous tasks (eg calculating Dopplr coincidences does your own immediately, and queues a job to calculate coincidences for all your contacts).
Using JQuery you can create your own events in a browser. Becomes more like traditional GUI programming.
Extreme late binding: combine data and code in application (eg javascript and data in browser – mashups).
The Internet is the computer.
First question: “How does Dopplr work?” :-)
Interesting point about hiring: go for the people who are good at the cool stuff (Erlang, Haskell, etc), because they’ll probably be really good at the workaday stuff too, and are obviously interested in the field.
[tags]barcamplondon3, messaging, erlang[/tags]
Caja compiles Javascript into different javascript, allowing you to put untrusted gadgets in trusted container pages (eg OpenSocial). Removes potentially evil code (eval etc), and passes in an object representing the global scope (document etc).
Ben’s writing a Caja wrapper for OpenSocial. Lots of hairy stuff with closures wrapping callbacks in functions, then wrapping the response in more functions. I haven’t really looked at OpenSocial yet, so some of this is flying merrily over my head.
[tags]barcamplondon3, caja, opensocial, javascript, gadgets[/tags]
Many pages on bbc.co.uk allow comments, but they tend to be trivial (“Mylene’s dress was a bit short”), are heavily moderated (which costs a lot of time), and are tied into the BBC site. Many of the comments people make are actually direct contacts regarding participation in the programmes, which end up going to the production team rather than on the page.
Mark is experimenting with allowing/forcing people to post their comments to their own blogs, and linking them back (like how the real Web works ;-) ). As well as solving some of the issues mentioned above, this would help the BBC fulfil its remit of encouraging participation in the Web (as opposed to treating the BBC site as a walled garden.
Apparently the message sinks in when people are shown Technorati results for their pages, when they realise how many people are commenting on them without them being aware of it.
In the Q&A afterwards, someone mentioned disemvowellment, which is an interesting technique I haven’t heard of before.
[tags]barcamplondon3, bbc, comments[/tags]
Screen-scraping HTML is evil. APIs are better, and can return structured XML, JSON or even SOAP (everyone laughs at the latter!)
If page is marked up semantically using microformats, an XSL stylesheet can convert it into RDF. Put the stylesheet on a profile page and link to it from the HTML (<head profile="...">), and triplr can generate triples from it in various formats. This can then be parsed using SPARQL.
The end result is custom microformats, possibly specific to your site, or something like class="nsfw" to mark up stuff that’s not safe for work, which Tom uses on his blog.
GRDDL allows
[tags]semanticweb, grddl, barcamplondon3[/tags]
Powered by WordPress