Caching what?

Open content's decentralised caching is interesting, but it assumes that your local link is not the bottleneck. I suppose if the source is someone else's local link this may well be true.

Berkeley & Stanford should talk

Jaron Lanier suggests a 2-dimensional, pattern-recognition based computing for greater resilience. Pat Hanrahan explains that modern Graphics coprocessor cards are now massively parallel computers, so this could be realised without loading the main CPU much.

Radio needs a 1st Amendment

Bob Frankston writes on how the associations of the word 'radio' prevent the FCC and others from understanding the abundance that could exist with smarter technology. This is why the word 'wireless' is back in vogue, a different term being needed when talking about networking. For me it evokes big valve radio sets.
Perhaps we need to revive the term Aether?

The coming dominance of IP - almost understood

Computer World says:
IP Devours All
Yes!
The Internet Protocol is like Pac-Man, and it will eat everything within the next 10 years. -- Hossein Eslambolchi, CTO, AT&T
Yes!
Frame relay will become obsolete and be replaced by IP within 10 years. Already, 50% of companies are now planning to move to IP. As with most trends, large companies will make the transition first, followed by midmarket and small businesses.
Cobblers. Most small businesses are runnign IP, not frame relay.
The transition will occur because all of the applications in the computing world use IP.
Yes!
Today we have to convert that protocol into frames, and we don't have the flexibility to give frames different priorities. IP offers the flexibility of establishing different priorities based on the application or the user. -- Andrew Borgstrom, CEO, T-Systems North America, Lisle, Ill.
No! - Priority is a mistake. It isn't really part of IP.

Publishers are Crap

Arnold Kling misses the irony:
I am not saying that free content is crap. I am saying that unfiltered content is crap. TechCentralStation filters content. I am still open to the charge that my essay reflects its title, but at least one editor must have thought otherwise.

The irony is that the filters that lead me to his Content is Crap article (various weblogs) have nothing to do with his publisher, the supposed filter. I read occasion articles on TCS, but I get there via a blog's deep link, and have never browsed the top page.

I don't think Bayesian filters are the answer either; trust models could be a useful aid, but creating economic incentives for helpful selection is the ultimate answer.

Of course, the mroe subtle irony is that anything that calls itself 'Content' likely is crap.

Derivative works - an alternative to Public domain

Give the Eldred decision, building new works on public domain ones remains problematic. I'd like to propose an alternative conceptual route.
In the word of digital media, the Copyright lobby's focus on copying seems very quaint. Making another copy of a work is trivially easy these days, so I'd like to suggest a modification to the idea of derivative works and the doctrine of first sale.
Derivative works should be allowed, as long as the full price of the source work is included in the price, and paid to the Creator of the source work (as long as the purchaser has not already bought it).
This avoids the whole commercial/non-commercial dichotomy that makes Creative Commons licenses so complicated, and has the nice property that works that derive from culturally widespread memes will be more attractive to fans. It is a neat answer to Ry Rivard's exponential growth issue and AKMA's call:
Discredit this decision by exercising our gifts and imaginations in public, without claiming exclusive prerogatives over those ideas� eternal destiny; share music, words, images, performances in person and online; but don�t make the mistake of thinking that Eldred v. Ashcroft closes the case on copyright.

I've discussed it in more detail at mediAgora

Eldred fails

This blog is black today as the US Supreme Court decided 7-2 against overturning the 1998 copyright extension .
Time to start a campaign for a congressional bill repealing it instead.

Hanging on

I was listening to Disc 2 of Pink Floyd's 'Echoes' in the car taking the boys to a movie, and track 1 is of course the very long version of 'Shine On You Crazy Diamond'.
Tonight I came across Chris Locke quoting it in a long piece about uncertainty, and it struck me - Chris Locke is the Syd Barrett of Cluetrain. Not sure where that leaves Doc and Dave, let alone the elusive Rick. (He exploded in a bizarre gardening accident?)
On the certainty thing, there is a time-honoured physicists joke:

With Newtonian Mechanics, you can't solve the 3-body problem (Sun, Earth, Moon)
With Einsteinian (Relativistic) Mechanics, you can't solve the 2-body problem
With Quantum Mechanics, you can't solve the 1-body problem (Uncertainty)
And Relativistic Quantum mechanics can't solve the 0-body problem (vacuum state).

Now we have computers, we don't need to solve equations any more, we can just run simulations instead, which means we can explore some more interesting problems, but can't know we have solved them. Wolfram's Principle of Computational Equivalence says that the really interesting problems can be solved by programs, but they take as long to run as watching what happens anyway.

"Hanging on in quiet desperation is the English way"

.lit Crit

Dorothea also sent a version of Down & Out in Microsoft Reader format (.lit). so I tried out Microsoft Reader.
First of all, it's only available for Windows, so I couldn't run it on my Macs.
The install nagged me about activation. I'm not going there...
I can't resize the window.
ClearType on my CRT just looks fuzzy and slightly fringed. I bet it's good on an LCD though.
Font size choices are very stiff and clunky. I can't change the typeface choice (It's an OK serif face vaguely like Baskerville).
It full-justifies and auto-hyphenates the text, but doesn't use dangling hyphens, and even has hyphens hang over page boundaries. Widow and Orphan control seems rudimentary. I got orphans in the licence agreement; if I turn the font size up (which doesn't make it resize the window width to compensate for line length) it orphans the second lines of the quote's titles. I bet it really blows with something more complex.
I can't turn this off to get ragged right. This really looks bad in the intro bit with URLs in and the indented contractual bits.

There isn't a way to navigate through the chapters within the text. The complex next/previous page thingie has buttons that look like they should do this , but they are disabled.

The way the Find box pops up and dances about is really nasty. Why does it have to be on top of the text at all?

Looks like the state of the art here hasn't advanced much if at all since the Voyager Books 12 years ago.

Dorothea explains what would be needed to add full OEBPS support to a browser.

Down & Out is out and downloadable

Cory Doctorow's first novel, 'Down and Out in the Magic Kingdom' is published today.
Cory has posted the entire text of the book online for downloading under a Creative Commons licence today too.
I really liked the book myself (I reviewed it in September).

Over the New Year break, I had some interesting email conversations with Cory and Dorothea Salo about the nature of eBooks. Cory argued that ASCII is best, both because of universality, and also because of the increasing inverse correlation between polished presentation and interesting content (when was the last time you get an HTML email you wanted to read?)
I pointed out that before I read his book I munged it into HTML with some regexp's. As Dorothea has previously championed the OEBPS format as having maximum flexibility, I asked her about this, and she heroically created OEBPS-compliant XHTML in very short order. I recommend her version if you want to read it digitally, but do buy the print book too.

MacWhirled

I went to visit MacWorld today. The new PowerBooks are very nice, the big one lighter and less extreme than it looks in the pictures. The show seemed very lively to me, with as usual some interesting things around the periphery, away from the obvious companies.

I found the elusive Video iPod - a FireWire drive that will record DV straight from the camera, for later editing.

Doc pointed me to Starry Night, a great simulation of the Universe, now on OS X as well as 9 & Windows.

I also saw CustomFlix, which I describe as CafePress for video. You send them a DVD-R of your movie and $50, and they host the webpage advertising it, and fulfill orders for DVD and VHS copies of it, for $10 a go, with you setting the sales price and being sent the profits. They'd be a great candidate for adopting the mediAgora model.

I also had a long and interesting conversation with Doc, touching on Digital ID, Zeroconf, mediAgora, Zipf distributions, annotating video, video hyperlinks, the inverse correlation between polished presentation and content, Blogger & Safari and other things I may get around to mentioning later.

Jack "My Precious" Valenti

Marek points to a curious parallel between the ancient, paranoid, yet curiously compelling synthetic creature, guarding it's precious and Gollum.
Not sure if this is fair to Gollum, as Gollum did lose the ring after all, whereas Valenti's members have only been made enormously richer by his imagined depradations.

Fins des si�cles

Neil Gaiman bids us read these two essays, in this order.
Enoch Soames: A memory of the Eighteen-Nineties by Max Beerbohm.
A memory of the Nineteen-Nineties by Teller.

He was so right.

Never type an IP address again

Bob Frankston has written a cogent piece asking for more simplicity in networks, particularly when connecting lots of devices dynamically.
This sounds exactly the problem space that Zeroconf solves.
In fact, Stuart Cheshire has described solutions to Bob's problems as examples of a possible future for Zeroconf.

Explicable weblogs

Bob Cringely finds weblogs inexplicable and wants a way to search them.
Weblogs let amateurs pontificate and think in public, just like Bob does for money. It is fun, after all. It would be nice to get paid too.
Unlike mailing lists, chatrooms and Usenet, they don't get polluted by spam and trolls because they are controlled by individuals.
You don't need to read them all, you just need to find a few you like and read those - the searching is done by individual selection.
And Google does index them very well - the individually made links in weblogs are a key part of what Google consolidates into a consensus about what is worth reading. If the weblogs say something interesting on a topic, Google will find it.

Once you find a weblog you like, use this tool to find others.

Look for followups here

The way to read them is an RSS aggregator like NetNewsWire.