h a t l e s s GmbH |
|
|
am the NICKEL. need sapphire loam UNIT. contact contact LINT contact. are never ETYMOLOGY ruin ruin BELIE.
NEW! You know, I used to have some old stuff posted here. Maybe I'll put it back one of these days. By old stuff, I don't mean the archive. |
Sunday, July 29, 2001
Oh boy. Like thousands of college students before me, I set up a Shoutcast server on my web hosting site where I have unlimited bandwidth use, incidentally, so I can run it 24/7. Took about half an hour to get it runing. Maybe it's on right now. At the moment I have all my MP3s playing in random order, except for a couple of whole albums that would have loomed too large. The fact that it sounds like a coherent, programmed radio show as it hops between Iranian and Arabic pop music, Seattle punk, French chansons, hard bop and R+B might say something unflattering about radio program directors. The key seems to lie simply in not playing any bad music. Thursday, July 19, 2001
I should be an industry analyst. This from something I posted elsewhere in mid-February. Note the emphasized part of the last paragraph: So let's see. The labels can take what may amount to a 25% cut of Napster's revenue under Napster's terms. Or they can let Napster twist in the wind and go bankrupt, and buy the domain for a few hundred grand, maybe a few million if there's a bidding war. Then they can decide how much to charge if they want to keep it running at all, and keep 100% of any such fees.
This piece by Joel Spolsky is one of the more lucid looks into software development--and the ridiculous Internet craze--that I've read. It's both cheaper and more useful than any Jupiter report you'll pay $4000 to read this year. The big insight? That it takes ten years to make good software. The numbskulls at Slashdot find it outlandish and dismaying. The poster asks how a project can possibly survive those ten years, and what it means for open-source software. It doesn't mean anything. Frankly, it appears to be a solid rule of thumb when applied to any large piece of software, like my law that says that any personal computer that can run all the mainstream software on the market costs $800 per year. It was true in 1979, and it's true now, this nonsense about computers getting cheaper notwithstanding. Or Moore's Law, which is right on schedule and shows no sign of breaking. With regard to open-source software, it's holding up just fine, and living by the same rules. Linux has been on store shelves and CompSci students' PCs since 1994 or so, but Mr. Torvalds started work on it in 1990. It's been genuinely useful as a server operating system since 1996 or so, but didn't have any advantages over the "other" free Unix, FreeBSD, until the late 1990s. And right around that ten-year mark, with the release of the 2.4 kernel, it finally became fast and efficient enough on multiprocessor servers with big storage arrays to make sense as a platform for hosting corporate databases and J2EE application servers. In other words, it took ten years for it to reach unconditional functional parity with commercial server operating systems like SCO Unix, Solaris and IRIX on 1-4 CPU hardware. Before the ten-year mark, it was a niche OS even in the Unix world: great for low-level services like SMTP, DNS, file sharing, FTP and two-tier web serving, but you wouldn't really choose it to run a production Oracle or DB2 server or an EJB container. Now it's got good SMP, stable journaling filesystems, and threading good enough to run a fast JVM across multiple CPUs. It took ten years. If I may go off on a bit of a tangent: Web servers have been around for less than ten years. They're good. They're useful. But good ones are also a specialty item, reserved for server racks. And ones suitable for personal publishing out of the box are underpowered. Apache reaches back about eight years, if you go to its NCSA roots. Now look at Mac OS X. Its personal web server, with a friendly, elegant interface, is Apache, effectively signalling the beginning of the end of the distinction between what you can do with personal web services and datacenter web servers. With WebDAV support now bundled in and servlet, SOAP and EJB engines that install in minutes, Apache is now posied to become the basis for peer-to-peer applications. By 1998, at Spolsky's 5-year mark, Apache had become compelling to datacenters with unlimited budgets, and started displacing IIS and Netscape servers en masse at the largest web sites purely on its own technical merits. Now it's become both so flexible and so usable that it could well be on everything in a couple of years. When you look at Spolsky's assertion, that software is Good Enough for a substantial core user base after five years of development, and that it can reach a broader plateau of feature-complete goodness after ten, examples abound. Windows NT 3.51, the first useful iteration, was released around the five year mark. Windows 2000, the first agreeably stable, scalable version? Ten years. Simplistic and a bit arbitrary? Sure. But the lesson is there: it takes a long time for software to get its "natural" feature set, and on that long march a developer has to think realistically and pragmatically about the order in which features should be implemented. Are the features making it into 1.0 the most useful ones possible? For all his criticism of Groove, it looks to me like Ray Ozzie and his team are thinking these things through carefully. That they're getting so much attention for a slim 1.x product has as much to do with who is behind it as any marketing effort on Groove Networks' part. However, as you try the Groove client with some friends or colleagues, it's telling what they put in, and what they left out, especially considering the company's Lotus pedigree. You'll notice that the calendar--a group calendar-- doesn't have any invitation and conflict resolution features common to groupware for years. A smart decision: group scheduling features tend to go unused at the team or small-business level, and that's what Groove's initial audience is. Next, note the lack of rich-text editing tools for many functions. Most instant messaging tools let you set fonts, colors, emphasis and so forth, as do e-mail packages. Yet only a few of Groove's tools support rich text at all, and not for lack of readily available widgets they could have ordered from the Programmer's Paradise catalog. But what does Groove do well? It can deliver invitations to workspaces via e-mail and several popular instant messaging systems, automatically detecting which ones you have running. Your workspaces and all features--including voice chat--are able to reroute themselves past most firewalls transparently, no preference dialogs required. No mean feat, that. A simple pen-based whiteboard is there. T.120 application sharing isn't. The list goes on, summed up as: features everyone is likely to use are in. Features only a minority will use regularly aren't. They're following the 80%-20% rule just fine from what I'm seeing. If they keep the company lean and keep their sales expectations in check, Groove will be mighty nifty indeed in another few years. And it's mighty useful now--just not for a mass audience. And really, this early in the game, what is? Sunday, May 06, 2001
Notes from Bike New York, 2001:
Monday, April 16, 2001
Last weekend was the first nice one this year. Spring came late to New York. On Saturday, I pulled my bike out of the closet, inflated the tires, reattached a loose brake cable and finally got outside after 11. I'd figured on going to Coney Island and Brighton Beach, but it was pretty late to get started. So I took the ferry from Hoboken to Battery Park. The one other cyclist, who seemed pretty hardcore what with wraparound sunglasses and a reflective jacket, heading across the river was goiung to Central Park, and in 11 years living in and around the city had never been out to Coney Island. In any case, it was a nice ride out. They've finally got the bike and pedestrian lanes marked right on the Brooklyn Bridge, so there really was a more or less clear bike lane all the way across. Last year, they were marked one way at one end and the opposite way at the other, so there was just confusion. This time around, once I hit Brooklyn, I looked for a more direct route to Prospect Park and found one, a quick left downhill in heavy traffic and I found Flatbush Avenue. Past warehouses, stopped buses, concrete barriers, BAM and some auto body shops, I eventually came to Grand Army Plaza and the entrance to the park. Once you get out of the park and follow the signs to Ocean Parkway, you ineitably get a little lost, loop around a few blocks and find it. From there, it's a straight shot, eight miles of a bike lane set off on a grassy, median lined with park benches frequented by American-born Orthodox Jews and mostly-older mostly not-Orthodox emigres from the former Soviet Union; dolled-up women from Odessa, clumps of older men from Uzbekistan. I got out to Brighton Beach around lunchtime; the sun was out, and I'd taken off the sweatshirt back in the park. The cafés on the boardwalk aren't cheap, not by a long shot, but when it's that nice a day, it's better than Nathan's: big steins of Pilsner Urquell, a ceaesar salad, and kebabs with garlicky parsley-dusted Russian-style fries, looking out on the ocean. 1980s Soviet-era underground rock has finally passed into nostalgia. RBC Video, the CD and video store for rock-minded emigres has cut the rock section by half to make more room for the flood of commercial videotapes and DVDs from the motherland. The shops selling and renting homemade dubs of movies and TV shows are fading. And as for the golden age of Russian rock, it's now the stuff of cheap greatest-hits CDs. One interesting packaging move was a series of legal, Russian-made CDs filled with MP3s of 5-10 albums' worth of music, sometimes by one artist, sometimes of several likeminded ones. Banged the hell out of my knee around Ocean Parkway and Newkirk Avenue, trying to stop while rolling off the curb when a car ran the light. I couldn't get my feet out of the pedal clips and did a slo-mo fall on my right knee. Monday, February 05, 2001
Sunday, February 04, 2001
So, dig. Friday night I'm at this bar in midtown with a few people from where I used to work and a few people from the Fucked Company gathering that had just happened.
But that sister, she looked so put upon, smiling like a trouper and all, but every time the flash went off, she was relieved that maybe it was over, this picture-taking.
At which point I'd take the picture, and she'd howl in anguish. "Stop! Oh please stop!" After three or four of these, she was somewhere between despair and real anger. Needless to say, each picture was better than the last. See, the thing is, good-natured party pictures are always dull as rocks, of fleeting interest only to the people in them. But candid snapshots of hostile people have real emotion, complexity, grace. Saturday, January 06, 2001
aaaaaahhhhh-CHOO! Some rules to live by:
Saturday, December 09, 2000
I've been looking for an archived copy of this for a few years now and I found it this morning: Date: Mon, 11 Apr 1994 09:46:00 EST MHS: Source date is: 10-Apr-94 17:44 EDT LINES COMPOSED UPON THE DEATH OF MR. COBAIN: I. THE BENEFIT CONCERT II. THE TRIBUTE ALBUM (Small-label version) III. THE TRIBUTE ALBUM (Warner-Brothers version) IV. WHAT WILL GROHL AND NOVOSELIC DO? V. WHAT ABOUT THE WIFE AND KID? VI. MEANWHILE, BACK IN SEATTLE: ========================= Tuesday, November 14, 2000
©1994-2001 s.m. koppelman. copy this and i'll hunt you down and kill you.
|