Longhorn Alerts vs. OS X Alerts
Interesting and entertaining discussion about alerts over on Matthew Thomas's Blog:
http://mpt.phrasewise.com/2003/03/09#a476
Just ran across this blog, some choice tidbits:
Larry Ellison: “Star Office is almost usable … I play with it, and it’s not awful.”
Brendan Eich, Dave Hyatt: “It’s clear to us that Mozilla needs a new roadmap … The major changes after [version] 1.4 involve switching to Phoenix and Thunderbird.”
Well, it’s about frigging time.
Tim Bray: “We have so much processing power and so much memory and such great tools and we still fail, most times, to produce things that are fun to use.”
"There’s a fascinating discussion going on amongst the Gnome hackers at the moment. (Well, I find it fascinating; your mileage may vary.) You see, it looks like they’re getting rid of the “Quit” menu item from their programs, on account of its cruftiness...
...With Mozilla development, I learnt to recognize this as the Jack Straws Effect. In the game of jack straws, the idea is to lift one of the plastic pieces off the pile without moving any of the others, which usually means you need to remove all the pieces above it first. This tends to be slow and laborious work — which, in the game, is kinda the point.
The equivalent in volunteer software development is people insisting that improvement A can only be made after larger improvements B and C have been completed, otherwise the software might temporarily be less usable overall. Complying with such demands allows volunteer software releases to happen more frequently than those of professional software — since (modulo stupidity and bugs) the very latest code is always at least as good as the last stable release. But as with the jack straws game, this makes for slow and laborious work.
Frustratingly, it means that as far as the user interface is concerned, even with the same number of developer hours, volunteer software will improve more slowly than professional software. This is because the professionals are happy to check in usability changes in any order, without worrying constantly about the current coherence of the interface, until the release date for the next version approaches. And when it does approach, they’ll be much happier to back out half-baked changes if necessary, since they’re being paid either way..."
I highly recommend checking out this blog. Lots of interesting tidbits and insight into the Linux / OpenSource world.
Update: LOL. Just noticed these great story links on the side: