Douglas Crockford on Google Gears and the Mashup Problem

Douglas Crockford is always an interesting speaker. At AjaxWorld last week he gave a talk about the good parts (there are a few) and the bad parts (there are many) of the current JavaScript standard. (That talk was similar to this Yahoo! Video of the Keynote from the 2006 Konfabulator Developer Day).

My favorite pearl of wisdom from that talk: The best thing about JavaScript is that there have been no new design mistakes since 1999 (when spec was last updated).

In addition to being highly knowledgeable (Brendan Eich called him the Yoda of Lambda Programming and JavaScript, he “discovered” JSON) he’s also entertaining, funny, and thought provoking.

In this video, after about 10-12 minutes of broad background on why the fundamental nature of security on the web is broken, he dives into the specific problem of mashups, the same origin policy in JavaScript, the global namespace and shared DOM, and suggests a method for using Google Gears to craft a solution.

AjaxWorld West Presentation: Two Steps Forward, One Step Back

I presented earlier this morning at Ajax World West. The title of the presentation was “Two Steps Forward, One Step Back: Back to the Browser Wars.”

Not sure how valuable the slides will be in the absence of my commentary on them, but here they are:

• Two Steps Forward, One Step Back (4.3MB, in ODP format for OpenOffice)
• Two Steps Forward, One Step Back (3.3MB, in PDF format)

Thanks to those who attended and feel free to contact me with any questions.

This relationship is off to a bad start

Coming across Roger Dooley’s post about Sears and their privacy policy (Sears- Marketers vs Lawyers, with a tip of the hat to Make the Logo Bigger) I decided to go check out the site he references, My SHC Community.

Unfortunately, no such luck (cue the “No soup for you!” clip from Seinfeld):

Was the problem that I was running Firefox rather then Netscape (Netscape? Really?), or that I was running Linux?
(more…)

Represent

I’ve been catching up with videos since the release of the Miro player public preview. (And as I’ve had some traveling time, on trains, waiting for planes, etc).

Two recent videos stood out as worth sharing. Both focus on creative visualization, and are inspiring in terms of how some relatively simply changes in visual display of information can have a tremendous impact.
(more…)

Facebook and Firefox, Platforms, and Freedom

Very interesting set of conversations this morning at the O’Reilly Executive Briefing.

Tim O’Reilly interviewed Dave Morin from Facebook - they’re building on a LAMP stack, and have contributed some things back, but clearly the main core of facebook is not an open source project.

His basic response was that “We will continue to release as much as we can, when it makes sense.”

Two reasons why it might not make sense came up:

1. The functionality the code offers is so tied to your services as to not be useful to outside folks
2. The codebase isn’t mature or professional quality enough - not “ready” to be released

For example, he said “we want to make sure that when we release something it is something of value, and something that the community can use.”

Then O’Reilly interviewed Mike Shaver from Mozilla, along with Matt Gertner from AllPeers and Garrett Camp from StumbleUpon, talking about the Firefox platform for extensions.

The Mozilla approach, as I suppose one would expect, is entirely different: release everything.

We don’t provide a tightly controlled API we let people access a lot. If you write an extension, it is as though you were writing code in the browser itself.
What we did was we gave people possibility.
What you get with source access is a very rich, and sometimes messy, set of points of contact with the overall platform.

I wish O’Reilly had gone further down the path of this question. Rather than deciding on behalf of the community which pieces they are likely to find valuable, Firefox takes the approach of allowing the community to determine what is valuable. Rather than waiting for code to be “mature” to release it, they let the community help make it mature.

It’s the difference between a platform designed to be extensible - which really means developers can write applications to run on our platform, as in Facebook, and designing a platform to be an open platform for anyone to do anything.

Is the difference just that the Mozilla foundation is a non-profit community, and Facebook a for-profit company?