« amerika 24/7 | Main | from the blogosphere »
May 03, 2003
wanting to piss so badly
here's a current problem with blogs, and the problems it leads to: bandwidth.
i really want to piss on Dr. Alan Kay's o'reilly etech presentation for a number of reasons. smugness, hollier then though tone, and the misleading nature of some of his details. as i reposted below "dude, it must suck to be 40 years ahead of your time!" except that i think he's dead wrong. Englebart was 40 years to soon, in some ways. In what Kay presented, Kay wasn't. Maybe.
yet i'm finding it hard to unlimber the canons on Kay. you'll notice I'm using caps here. Dr. Alan Kay is one of my heros. he was a total fuck up till he met his first computer, he coined the term personal computer in the 60's, he pretty much invented a lot of modern object oriented programming. and he was one of two pioneers that pushed for children's computer usage, the other being Papert at MIT (who invented what became the LEGO Brainstorms kit in the 70's.) he's the father of the Xerox Alto. more importantly he's the father of the still unbuilt Dynabook, a concept so powerful it fathered everything from the modern laptop and PDA/PIM to the cell phone and WiFi wireless at starbucks. so powerful toshiba calls their some of their high line laptop Dynabooks everywhere but in america, but still leaves the moniker hidden on parts of the case. The concept is so powerful, Microsoft has tried to build them twice, and Bricklin went bust trying to do it 12 years ago (Kay's presentation was down on a Microsoft Tablet PC, except for the OS, as close as you can come to a real Dynabook--and you don't know how much it galls me to say that.)
but all most of the issues i have with the presentation are checked by not having seen it. unlike a MS Powerpoint presentation, Kay used a dynamic, fluid video presentation that involved everything from archived footage of Englebart demonstrating interactive computing, networking, windowed operating systems and mice in the 60's (heyday of paper tape and punchcards) to real time computer generated 3d images, that he literally programmed right in front of the audience. there are a few audio and video collections of the presentation, but they're being kept bandwidth limited. pulling down a 650 meg pile of vids from lisa rein's archived page at 1k/sec will take a week or so. i wasn't able to get the first 50meg chunk in 5 hours last night, on a high speed connection. and i can't just download his software and try it my self. while he talks at length about squeak, he did the presentation in something new. which he removed from the website because of slashdotting--before the Etech convention. so if i do take affront at what i've read of his presentation, I'm missing all the salient details. all i can do is piss on quibbling little bits, do a partial ad hominum on my hero, and make an analogy:
the same thing happens whenever you watch or listen to the tv or radio. you aren't get all the details, you don't even have the bandwidth to do so. everything is mediated.
i can say is he did lie once. Squeak isn't an OS, it needs to be hosted on an OS. he dodges this in some corrections he posted to doctorow's notes. claiming operating systems are passee is horsehit. OS's exist so that dreamers and application designers don't have to write their own drivers. there was a point when they did, if they were writing for dos in the early days, and it was ugly. every application you bought had to have it's own driver set for everything from the screen to the printer, and not all of them had the drivers you needed.
Kay is also an academic in outlook. his major notion of all you need is an object oriented language falls short when your a secretary who needs to do a mailing list merge by first writing the word processor. to a certain extent, this is the unix notion of small tools easily linked. the reality of the workplace is building and linking tools tools takes a back seat to getting the job done cheaply. which is why MS is pushing into schools. then there's the possibility that if your secretary can wrote a mail merge and word processor, she might decide she's too good to be a secretary and leaving only the truly ignorant to be secretaries, and once again necessitating pre-made applications. can't imagine the captains of industry liking that, but it's a minor point. because if Kay is right, after the usual mosaic generation of change, everyone will be a programmer.
this doesn't just apply to secretaries. i can't imagine most people being able to or wanting to roll their own software. in Kay's defense, using squeak doesn't truly preclude selling applications written in it. but it could...
then there's the intellectual property claim. smalltalk, on which squeak and croquet is based, has been encumbered for over twenty years by xerox. there's no way it could have caught on, because no one but him and the few people who could afford the license fees ever used it. it's only recently thast clones/lookalikes of it have popped up. and he wasn't exactly the worlds greatest evangelist for it, no matter how great his live presentations are supposed to be. he could have pushed for it's release, or a change in licensing. IBM was a better evangelist of the smalltalk languages, to the point of building all it's integrated development systems around the smalltalk browser. while squeak is open source, or so they say, the software he used for the presentation, croquet isn't. well, it isn't even available to play with.
my final complaint with Kay's presentation is a subtle one. his proof of concept is entirely based on extrememly smart academic users, and children. children are capable of things adults aren't. for example i used to be able to read 3 books a day. if you start a kid young enough with a computer, she's capable of amazing things, especially if she has a predilection for problem solving. Kay also begs the question of how large his sample base was, and almost certainly the kids work shown to the audience was that of Kay's prodigies...
but i'll never know. no bandwidth on my end, no bandwidth on the transmitting end. guess i'll just have to go to etech next year, and hope he shows again.
Squeak can be found at Squeak.org
croquet is not available for download.
The Unix Hater's Handbook is availible as a PDF from one of the authors. the last chapter includes a portion of "The Rise of Worse is Better," by Richard P. Gabriel. an interesting article on why academic "perfect" systems like small talk and squeak fail
Posted by parody at May 3, 2003 04:24 PM
Trackback Pings
TrackBack URL for this entry:
http://www.ffej.org/cgi/mt/mt-tb.cgi/598