Focus on the frontend, increase percevied performance.
Script block page execution, they block document.ready event. Cuzillion is a tool to test execution flow. There’s a set limit of requests for each domain, ‘connection pool’ is the tech term, so use domain sharding. images.example.com/ scripts.example.com point to the same resource. Then you can queue up resources and they won’t block, because they served from different domains.
There’s a full breakdown of techniques in the slideshow but using JS to insert JS into the head won’t block, so you can use a main script to trigger loading other scripts. Useful but it gets tech if you need to load a library before another script, e.g. jQuery before jQuery UI. Suspect somebody needs to write a library to do this better. Pay attention to race conditions.
iframes the most expensive DOM element another excuse not use them. iframes also block document ready, which is hard when your serving 3rd party ads. A way around this is to wait for document ready then insert the iframes, they load after the important JS and don’t block your UI. Just bundle them in with (#foo).html = “adcode”
There were also some recommendations about using mod_delate although that spends CPU time, so if you’ve got heavy traffic it might not be applicable. Also Google do some clever stuff with flushing, where part of the page is sent to the browser even if the rest of the page is in flux. It’s interesting, but then they’ve got control of the whole stack. They could/do probably run a modded up apache which is speed focused.
“Google want the web to be instantaneous” interesting choice of words, not fast, or quick but immediate. It’s a utility.
SXSWi: Everything You Know About Web Design Is Wrong
We kicked off our first Substrakt SXSW with Everything You Know About Web Design Is Wrong, I think we came out of it with a fresh perspective and focus on a very familiar topic.
Here’s a quick video summary:
raw notes: (my comments in bold)
media driven by tech not artists
master the tech and have good content = win
web needs to develop a language/grammar, film used to suck but then we developed ideas about how to use the camera to tell a story
artist driven instead of tech driven
list of sites that would be just as good if they were printed
above the fold is a big fail won’t be the last time I say that
don’t know what the grammar will be, but here are some emerging patterns
random voyerism
examples like flickrvision because we like to watch see also: flickr.com/explore/clock/ which I think is the best example of this atm
we like to watch people, even if were not connected to them people are all the same
another example is found magazine, which is a collection of found photographs we have the ability to construct a narrative from non-narrative forms
self-aware, but controlled, content (i.e. content with metadata)
Metadata is content that knows itself better than we do.
user created content
The web is about a single user and the choices they make.
They control the content
ambient awareness
trival and profound
twitter as a portrait again with the network narrative thing, also relevant to Transaction Analysis
experiential content
Rollercoaster is not the track, it’s the experience
Experience as the content
Design is not about making something look pretty, it’s the whole stack, visual design is a means to an end, real design solves problems. Design needs to happen at the beginning not just at the end, and it needs to be jambalaya mix everyone in. Use an expert in the context to explore and protect the user experience
Here’s the answer but your asking the wrong question
there was a bit of talk about being allowed to fail and this usual learn from your failures vibe, I don’t agree, you need to learn from success not how to get it wrong
If you’d like to digest the whole thing on your iPod then grab the podcast, everything was filmed but actually I can’t seem to track down the full version anywhere.
In March, Jim and I will be off to SXSW. There’s some fantastic panels lined up, I’m already having a hard time picking which things to go to, everything sounds a bit too interesting.
I know that Pete Ashton and Antonio Gould are headed in that direction, but who else from around Birmingham is going over? It would be nice to have a bit of a meetup in Austin.
If you’re headed over then add me on Dopplr so we can start making plans!
