29 Sep 2010 JoeNotCharles   » (Journeyer)

W2E Day 2: Afternoon Presentations

What to Expect From Browsers in the Next Five Years: A Perspective
Panel discussion

This was a great discussion, although it really didn’t touch on its ostensible topic much! It was more a discussion of the current state of the browser. The panelists were:

Douglas Crockford, creator of JSON, active on the ECMAscript committee, and Javascript evangelist at Yahoo
Alex Russell, whose workshop I already went to on day 1
Brendan Eich, original inventor of Javascript and Mozilla CTO
Håkon Wium Lie, creator of CSS and Opera CTO

The discussion moved way too fast to take detailed notes, so here are some quotes (paraphrased):

On the IE9 beta:
Hakon: “I’ve dedicated my life to improving IE”
Doug: “They have no ECMAscript 5 – there’s no provision in the standard for subsets meaning they’ve chosen to be noncompliant”
Alex: “It’s not available on XP, which is a concern because XP still has so many users”

On how to differentiate a browser:
Hakon: “Opera has a good emphasis on JS performance, but getting on every device in Asia and Africa are another goal.”
Brendan: “Mozilla is behind on performance but improvements are ongoing. Integration with the user’s identity is what I’m interested in (password management, etc so that the browser has your info and doles it out to sites rather than the user passing info to sites individuallly).”
Doug: “Do we work AT ALL – getting better.” (I have no idea what the context of this was any more.)
Alex: “Developer productivity. The big bottleneck is now network behaviour: we need more expressive API’s because using javascript to address HTML and CSS shortcomings makes the web less semantic.”

On data privacy with the browser knowing your identity:
Hakon: “I’m scared about evercookie. I would rather have the browser shield the user.”
Brendan: “The browser should be the first point of trust, providing API’s to the user, instead of Facebook providing API’s.”
Doug: “XSS is the biggest security problem: all scripts share a common global context, and there is a complicated language abd protocol stack with different conventions at each level making it impossible to reason about. HTML5 was irresponsible in adding new networking, local storage, and DB access before fixing XSS.”
Brendan: “We need to fix incrementally, or nobody will adopt. HTML5 has some security features.”
Alex: “We have theoretical solutions. The battle will be good vs usable: it has to be a system that everyone can use because that is the web’s strength.”

On the use of “HTML5″ to refer to a set of features, many of which are actually CSS additions, rather than the actual spec:
Hakon: “This is a marketing problem: people want one name. The solution to ensure MS supports these features is ACID tests.”
Alex: “This is unproductive because lag time on creating ACID tests and getting browsers in hands is very long. I’m more interested in dynamics of getting browsers upgraded and released quickly.”
Doug: “This gets worse in the short term for web developers because the differences between browsers will increase, but longterm IE6 will die and it will become better.”
Brendan: “We need to use JS libraries for extensibility, because building everything into the browser is too slow to deploy.”
Hakon: “Yes, we need to use JS as a sandbox and move successes into the declarative side.”

On apps vs the Web:
Hakon “Native apps will be a footnote.”
Alex: “The Chrome web store let’s you monetize a web page. Targeting fixed hardware let’s you target the edge of a device, but as hardware improves we can burn cycles and the Web becomes a cost reduction to avoid writing things 5 times.”

On Google’s NACL:
Somebody, I didn’t write who (maybe the moderator): “It’s a promising research and prototyping story to get better performance out of the browser.”
Brendan: “It’s too complex. People here are calling it ActiveG. Nobody wants it. I don’t want pthreads running in my browser.”

On the lack of audience questions:
Moderator: “I don’t think you understand – this man invented CSS. If you have any questions about why your transforms aren’t working, ask them now!”

On developer tools:
Everybody: “Our browser has a tool built in. This is its name.”

On XML compatibility:
Doug: “XML is obsolete, didn’t you get the memo?”
Brendan: “Firefox uses heavy XML. We bought into it heavily, but now we’ve ripped a bunch of it out.”

On Mozilla’s evolving role now that every browser is debuting new user features, and the fight to get IE to follow web standards is won:
Brendan: “We represent the user first and only.”
Hakon: “Competition is great. We need several rendering engines to verify the models.”
Brendan: “People think engines are too complex to maintain several, but it’s not true.”

On standards bodies:
Brendan: “We need smart people who can do both practical and theory to serve on them, and we need them to work together.” Doug: “ECMAcript is the most important standard because if the others are broken, JS is the workaround.”

Some genius in the audience asked the best question of the session: “What’s the biggest unfilled hole in the HTML/CSS/Javascript stack?” On that:
Doug: “Security.”
Alex: “Integration. There are leaky seams between standards, and too many standards groups that don’t cooperate.”
Hakon: “CSS has an object model coming to help with that.”
Brendan: “The security problem lies in the leaky schemes.”

On plugin privacy concerns (users uniquely identifiable by their list of installed plugins):
Brendan: “We’re going to start turning off old plugins.”
Alex: “We’re trying to reduce the surface area of plugins. We need to leak some info so that they can be instantiated, but at least we need to hide their upgrade path and only keep the most recent versions installed.”

10 Things You Never Knew About Opera
by Håkon Wium Lie

I wanted to see this after seeing Hakon at the last panel, but it turned out to be pretty useless: basically a big ad for Opera. I guess that’s expected from the title, but I was hoping for some interesting new technology they incorporated or something. Instead it was all, “We employ 500 engineers,” and, “We’re big in Russia.” The only new thing mentioned was Opera Unite, which didn’t work when he tried to demo it. (My big question about it was how it dealt with firewalls and NAT – apparently it doesn’t, which is pretty useless.)

David Kaneda mentioned in the morning that he was required to make his presentation about general technology and not his particular product. I guess sponsors are immune.

Personal, Relevant, Connected: Designing Integrated Mobile Experiences for Apps and Web
by some drones from Microsoft

This was even worse. A lot of people walked out. I won’t waste my time by elaborating.

(While I’m attending this conference on behalf of Research In Motion, this blog and its contents are my personal opinions and do not represent the views of my employers. No, I would not like a pig.)


Syndicated 2010-09-29 21:28:51 from I Am Not Charles

Latest blog entries     Older blog entries

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!