I need to minimize the use of my computer for a few days, so Hack the Planet probably won't be updated.
IBM is getting ready to release a transcoding server product. Considering that transcoding has been possible for several years, I'm surprised that no startups have beaten them to market.
You might have thought Jini was just software, but now it's a government, too. It has a bicameral legislature with semi-direct representation and an executive technical oversight committe. Will democratic software work?
Christopher Browne posted an interesting message about configuration formats, XML, and databases to Slashdot.
IronDoc development has resumed after a 10-month hiatus, although the status page has not yet been updated.
According to the latest update of the Mozilla roadmap, the developer beta release has been pushed back to February.
JavaWorld has an interesting article about the Java Community Process. The JCP moves so slowly; out of 37 accepted JSRs, only one spec is available so far.
Is it just me, or does this sound like Terrasoft has a non-compete agreement with Apple? "Staats says the warm relationship between his company and Apple management exists largely because Terrasoft targets its Yellow Dog Linux distribution toward the core of the Linux user market: Internet service providers, Unix system administrators and research laboratories. This leaves Apple to tend its traditional markets: home users, schools and graphics systems users." Will we ever actually see the "Gone Home" desktop edition of Yellow Dog? Read the rest in Upside.
"The strategy of putting all device drivers in the kernel tree is fundamentally unscalable. It has been adequate until now, but with the exponential proliferation of device drivers we've seen, the time will come, I'd guess in the next two years or so, when people will realize that this just won't work any more. No one is going to want to hand propagate kernel API changes through 1000 (or 5000) device drivers in the kernel tree just to get a clean build, or even to wade through that many kernel config options to find the ones they want." -- Dave Hinds
Mark Schumann's endorsement for this site is cleverly concise: "It's fun, it loads quickly, and you learn things."
Google officially launched today and they added a new "GoogleScout" feature that tells you what links are "related" to others. Interestingly enough, my home page is related to John Brockman's site and the Woz. All the links related to Hack the Planet seem to be Weblogs!
Today's Alertbox is pretty good.
eTunnels might be a good idea, although I'm not sure why you have to have an account on a central server to use it. Since their site is low in information, I suggest heading for the walkthrough.
Oh my. The Merlin has 4 PowerPC 7400s at 333 MHz with 2MB 1:1 L2 cache per CPU, and it supports NUMA. Too bad only the military can afford it.
QScheme is a new GPLed Scheme that claims to be pretty fast. Looks like it's time for some *Scheme benchmarks.
I noticed today that the UT CS system has CodeWarrior for Linux on it. I didn't have a chance to use it much, but the dialogs are a disaster. If only Latitude supported Gtk+...
Normally I wouldn't pay attention to anything from Microsoft called a "digital dashboard", but it actually looks like it might be a useful idea.
Do any readers out there know anything about the Charon filesystem?
John Brockman talks to Roger Schank about the disrespected student. "School isn't really about learning at all. It is about certification. ...Education ought not be a competition. Learning should be fun. The stress that students endure in school and the arbitrariness and general lack of real world relevance of what they learn make learning anything but fun. ...Most computer science professors are not familiar with the commercial packages that are in use on a daily basis in industry and even if they happen to know them, they consider them to be of little intellectual interest."
Somehow the crickets are getting into my apartment again. :-(
Motorola has a new AltiVec site that looks helpful, well-designed, and free of marketing nonsense.
I read Ender's Shadow last night. If you liked the first four, you'll probably like this one.
President Clinton is expected to loosen the US crypto export rules today, but as Declan McCullagh points out, the little guy still looses. "Thursday's White House announcement loosening encryption import standards may make it easier for big businesses, but it won't help anyone who wants to distribute software freely on the Web."
Prompted by one of today's Slashdot articles, I started reading through the Internet2 Web site to see if they have actually accomplished anything yet. It turns out that they have: The Distributed Storage Initiative is an Akamai-like setup for universities that mirrors CPAN, MetaLab, and some other stuff. If you want to install Linux in a dorm room over the Net, DSI is probably a good bet. The upcoming Megaconference multi-way video conference also looks insteresting.
Maybe they're on to something at Epinions. If I like the same books and Web sites as Naval Ravikant, maybe I'll like the same TV, too. (It looks like Jorn changed the photo on his page to a less frightening one...)
Jorn also has a Weblog FAQ; I'm glad to see him encouraging Webloggers. (Pages that are linked from several Weblogs seem to annoy some people. Since I only read 3 or 4 logs, there's no way for me to know what the others are pointing to. Likewise, I can't assume that readers of this site read the same sites that I do.)
This is kind of interesting: As part of the Java Community Process, Sun has put up diffs between the 1.2 spec and the proposed 1.3 spec. Unfortunately, 1.3 has been in the works for months already, so it's pretty questionable whether feedback at this point would have any effect. Also, there are tons of spurious diffs that only fix misspellings and typos, making it difficult to find things that actually changed.
An early access version of JAAS is available. "The Java Authentication and Authorization Service (JAAS) is a Java package that enables services to authenticate and enforce access controls upon users. It implements a Java version of the standard Pluggable Authentication Module (PAM) framework, and extends the access control architecture of the Java 2 Platform in a compatible fashion to support user-based authorization."
InfoWorld's Dylan Tweney thinks DNS may be replaced by a "friendly" system like RealNames. Unfortunately, this overlooks an important fact: DNS is hierarchical; most friendly systems are not. I can afford one domain, but I shouldn't be forced to buy a separate friendly name (and pollute the namespace) for every one of my computers.
Dan Lyke: "Personally I'm pushing for a real content management system to eradicate all those out-of-date web pages, but while the world needs that, I'm not sure that they want it." I'm not so sure either.
Jeremy Bowers has a great paper on Third Voice. I'm about to go to bed, but hopefully tomorrow (or tonight, depending on when you read this) I'll have time to write my thoughts on how to do annotation services right. The summary: Make comments into first-class resources and separate metadata storage from processing using open protocols. For some possible solutions to the scaling issues (what if everyone uses annotations?), look at K. Eric Drexler's paper "Hypertext Publishing and the Evolution of Knowledge". I'm sure I'll also mention CritLink (a little better than Third Voice IMO, but we can do even better) and maybe get around to reputation servers. (I know Bruce Sterling gave a talk somewhere about reputation servers; any readers want to find it for me?)
There's also an interesting Third Voice discussion on UserLand's discussion group. "This has gone too far." -- Dave Winer
After Amiga Inc. gave up on the hardware business, a group of Amiga lovers from QNX, Metrowerks, REBOL, Met@box, Phase 5, KOSH, Amino(?), etc. have created the Phoenix Consortium. All that has come from the Amiga community in the last few years is vapor; I can't imagine why it would be different this time. An interview with Fleecy Moss has the details if you're inclined to believe them.
Aironet has redesigned their Web site, possibly to cash in on the upcoming AirPort rush. Too bad they screwed it up: the site forces you to open a new window!
I just noticed that I am an "Elite Member" on AppleInsider. Scary.
In the now-it-can-be-told department: SOAP has arrived.
Wow, there is nothing happening on the weekend.
If you're as bored as I am, check out Elgin Park, especially "What I Can't Do".
The project's finished. <phew/>
In Salon, Scott Rosenberg looks at the good and bad of network computing. I think his analysis of free Web services is particularly important. If the service is free, that means the company probably doesn't see you as a customer, but as a sucker.
Why would someone from Transmeta be porting the GNU toolchain to PicoJava?
Here's a good paper on the design of Dents, the only DNS server that I've seen that cares about efficiency. "Ironically, automated interfaces and graphical interfaces have the same requirements, and textfile-based interfaces are hostile to both. (For some reason, many people think that textfile-interfaces are friendly to automated environments; of course, nothing could be farther from the truth.)"
The JDC has a transcript of the chat with MAJC designer Marc Tremblay.
I saw some other good links today and yesterday, but I don't remember them now. Oh well.
WebSprocket has released some new information about their far-out embedded Java development tools.
Sun announced the Sun Ray 1 Java terminal today. Besides a partially-translucent case and a smart card slot, the Sun Ray is actually less powerful than a JavaStation. Basically it's equivalent to a really really dumb VNC terminal.
I finally found a good Sun rumor weblog.
I was just thinking: Lydstrom, TiVo, DTCP, DTV, digital cable, 1394; I wonder when an equivalent to the frox will arrive. It's got to happen sooner or later.
"Web designers doing weblogs is kind of like the old joke about student film makers: they tend produce films about being a student film maker, filled with struggling actors playing struggling actors and frustrated writers playing frustrated writers." -- Matthew Haughey (I can't help but thinking that that sounds a lot like Shakespeare in Love.)
No news today. I have to write a CPU scheduler (with a partner :-< ) before friday, so don't expect much the rest of this week.
In case you didn't read it when it was in Wired, I suggest you spend a few minutes with Brian W. Aldiss's "Supertoys Last All Summer Long".
"You ask such silly questions... Nobody knows what 'real' really means."
An early access release of Solaris 8 seems to be available. Anybody out there have details?
There's also a new pre-6.1 Red Hat beta, including (surprise, surprise) a graphical installer. It's nice to see that all that IPO money is going to good uses.
Dan might be interested in this after he gets back from the desert: The Digital Transmission Content Protection (DTCP) scheme will apparently be protecting all those cool video-over-FireWire links in home theaters of the future. I'm not sure which is worse: the fact that it's super-proprietary or the fact that it uses "viral" CRLs which spread automatically from device to device, allowing the 5C group to remotely shut down any DTCP device.
Amid plenty of complaints about new entertainment-labelling measures in Congress, only Evan Vetere sees the good side: "As a teenager, I want better labeling of content so I can tell at a glance while scanning DirecTV's film listings which shows contain nudity, which contain wanton violence, and which are just pansy-ass dramas."
My DNS provider was apparently down for a while today. :-(
Ooh, I found the crypto snake-oil FAQ. I can't go a month without seeing another snake-oil system, especially since that includes all copy protection schemes.
Emacs runs on Macs now. Something is really scary about that.
Total Impact has some interesting NUMA PowerPC multiprocessing PCI cards and they're planning to run Linux on them (somehow).
Here's the real explanation for the Windows CryptoAPI backdoor. More comments: Paul Crowley
I didn't lend much credence to the rumors that Microware would sue Apple over OS-9, but it finally happened.
Meanwhile, some G3 users are not happy that Apple is blocking third-party G4 upgrades. It's all about respect.
It's an interesting day in Web-land; the guys running Slashdot apparently don't know how to configure their DNS, and I couldn't get through to nirvana.userland.com (again).
X Appeal takes a look at an August build of MacOS X. It's looking a lot more like MacOS.
(This may not be true after all, but here's my analysis anyway.) I knew this would happen eventually: Red Hat may be restricting the use of their trademark, specifically on stuff that isn't "official". I see this as a good move, since trademarks prevent the bad forms of fragmentation (where you buy something called "Red Hat" which isn't actually compatible with Red Hat). I had a feeling this might happen when I noticed that VA Linux Systems bundles "VA Linux OS" with their systems; apparently since they install updates and a customized kernel, they can't call it Red Hat.
There's an interesting discussion about HTTP compression on UserLand's discussion group. Even though it may not intuitively seem that way, compression comes out ahead in every actual test that I've seen. Yes, modems have compression, but it's not nearly as efficient as deflate. If your Web server already has some kind of cache, then you can cache compressed variants of your pages so that there is no increase in latency. I don't think that compression is the most important issue on the Web (by a long shot), but it can be beneficial.