OSCON: Tuesday night, Wednesday keynotes

[mdlbear]

Last night's keynotes were Mark Shuttleworth (Mr. Ubuntu), Robert Lefkowitz, and Damian Conway.

Shuttleworth spoke without slides, bouncing back and forth between technical and community issues. Conway was, um..., indescribable. Quantum computing, special and general relativity, and some hilariously funny and totally faked Perl demos illustrating "positronic variables" that get set backwards in time.

Lefkowitz was the most interesting: he talked about software development methodology, using Quintilian's Instititutio Oratoria as a framework. The interesting thing about the open source process (in his version, which I don't necessarily agree with) is that it starts with a commit to version control, and includes users. There's no requirements phase; it's replaced by bug reports and feature requests.

Right now I'm in the morning keynotes; I'm probably not going to have much time to blog. Tim O'Reilly just came onstage. This is the 10th OSCon, and the 12th anniversary of the first Perl conference, which I attended in San Jose.

Net's getting laggy. *cheery wave*

Comments

Hi Steve!

[webmaven.livejournal.com]

I saw your cards on the table yesterday. I'm here too (speaking tomorrow). Got lunch plans?

Re: Hi Steve!

[mdlbear]

Hi - was planning to lunch at the con -- they feed you well, here. Join me?
I'll be down in the exhibit hall at the break.

Re: Hi Steve!

[mdlbear]

cell: 408   896   6133

Re: Hi Steve!

[webmaven.livejournal.com]

Mine is: 702 807 5557

I'll look for you during lunch.

I'm now sitting in the 'Opening Education' session, then in 'Web graphics and animation without Flash', then lunch.

Re: Hi Steve!

[mdlbear]

Got a late start on the first session due to a conversation in the exhibit hall; might get to the web graphics thing; looks interesting.

Re: Hi Steve!

[mdlbear]

I'm wearing a purple shirt.

[lapislaz.livejournal.com]

No requirements? What happens - someone just says "you guys start writing code - I'll go see what they want" ???

Oy.

[mdlbear]

What happens is that there's an existing project - one person, small group, even a proprietary product - that gets opened up after it's already proved that it can fill some existing niche. At that point, it acquires new users with new needs, and evolves by collaboration.

[technoshaman.livejournal.com]

Agile does this. You start with a backlog of feature requests, and when you've got something demo-able, you demo it, and they file requests and bugs against that, and it sorta snowballs until they're willing to pay you money for it.

I don't agree with it; unless the project is pure T-and-E based, or you have Godzilla for a project manager (somebody the client is just a little afraid of), you'll get screwed by scope creep so hard it'll come out your ears... (graphic, but true!) But it seems to be de riguer in telecoms, where the client knows they don't know exactly what they want but if they don't move yesterday the competition will beat them to market and then *they'll* be screwed and you won't get paid. *sigh*

[thnidu.livejournal.com]

Ah, mobile Web access. I finally got geeked enough to configure my T|X for the office's wireless servers so I can read LJ in the loo in the lunchroom, as I'm doing now.