Wednesday, May 17, 2006

Java on the server, AJAX on the client

I'd blogged earlier about how Netscape and Sun fought to the death over a way to rich web clients, while Microsoft stole the prize.

Today the battle has been rejoined. Google has released their Java based (server) AJAX Framework. The Slashdot reviews are basically favorable.

Wow. We could have done a LOT with this 6 years ago. Today's independent small development shops now have some very powerful tools, tools that may deliver the vision that Sun and Netscape destroyed. Microsoft cannot be happy today.

Google does so do lock-in. Fool.

The Economist needs a CPU upgrade. They're not only channelling the Wall Street Journal's moronic editorial pages, they're completely mangling the most important concept of modern business -- lock-in:
Information technology | Is Google the new Microsoft? | Economist.com

... Try to avoid using Microsoft's software for a day, particularly if you work in an office, and you will have difficulty; but surviving a day without Google is relatively easy. It has strong competitors in all the markets in which it operates: search, online advertising, mapping, software services, and so on. Large firms such as Yahoo!, which previously farmed searches out to Google, have switched to other technologies. Google's market share in search has fallen from a high of around 80% to around 50% today. Perhaps the clearest evidence that Google's continued dominance is not inevitable is the fate of AltaVista, the former top dog in internet search. Who remembers it today?

Without a proprietary lock-in to protect its dominant position, Google will have to work hard to stay on top.
Anyone try changing an email address lately? Move one's blog from blogger? Switch from Gmail? Google is far more open than Yahoo! or AOL (they support POP transfers from the mailbox), but it's not like you can move a 1 GB image repository from Picasa to Flickr in a keystroke. Data ownership is a far bigger lock-in than mere software UI. Just ask any corporation running SAP.

In any case Microsoft's lock-in was never Word (really), it was .doc. That's an immensely important difference.

How can The Economist miss something as fundamental as this?

There's a lot more to this topic than I can discuss in a few moments, but here's an exercise for the reader. Describe how the following items relate to lock-in and how a manufacturer can obsolete a product they seem not to control ... (hint. Stop making the stylus.)
1. Apple's new magnetic power cord attachment.
2. Any proprietary battery for any device.
3. Any charger for a device.
4. Any patented connector. (Think iPod connector.)
5. Any stylus or any other non-generic consummable that gets lost or broken.
For extra credit note that Apple's lock-in is not merely the FairPlay DRM, it's that nothing but an iPod can sync with iTunes.

Update 5/17: Hey, is the Guardian reading my blog?
... Received wisdom says there's no lock-in on the web, with rival search engines just a click away. But if you develop a Google search habit, and Google has your email and address book, calendar, news feeds, bookmarks (in Notebook), back-up files (in Gdrive, soon) and other data (in Base), and if it handles your voice calls, organises your photos and so on, then you're going to find it increasingly tedious to switch. That's the idea.
Seriously, Jack Schofield must have written this independently of me, but it is noteworthy that The Guardian has pipped The Economist. And not for the first time. Jack has a blog btw, which I've added to my bloglines list.

The power of Google and Wikipedia

I recalled that an evolutionary biologist of some sort had once used an exotic word from architecture to describe something in biology. What was the word?

I cheated. My wife remembered Stephen Jay Gould's name, but nothing about the term -- though I'd have found his name easily enough on Google.

A Google search on him suggested Wikipedia. Wikipedia had the term:
Stephen Jay Gould - Wikipedia, the free encyclopedia:

...use of the architectural word 'spandrel' in an evolutionary context, using it to mean a feature of an organism that exists as a necessary consequence of other features and not built piece by piece by natural selection...
Of course it's easy to then research spandrel. This must all seem mundane to young-uns, but it's still miraculous to me.

Manipulating memory by photo display - unexpected results

An experiment I began a year ago has had some unexpected results:
Gordon's Notes: Happiness is a selective memory - manipulating memory for good and for profit

My approach to creating a selectively-false and happy set of memories is a large collection of family photos that cycle across our array of computer displays. These leverage the principle of selective reinforcement of memory -- given two proximate events, unbalanced reinforcement of one will decrease retrieval of the second. It as though as one memory grows it usurps the foundation of its "neighbor" memories. In this experiment the happy photos selectively blur away all other events.

Truth is fundamentally overrated in our current universe.
It turns out there's a reason humans don't look at old photo albums all that often. I have now thousands of family related images spanning about 100 years that cycle, at various times, on up to 5 displays. I find that images that are older than about 1-2 years, which is roughly the memory range that I live in, are often unsettling. For the children, 6 months is about the limit.

Even edited for happy events, the pictures show beloved pets that are gone, loved ones that are gone, friendships that have sundered, children that are now different (too quickly), our younger prettier selves.. It all it is a richness of living that we cannot truly contain, we are not "made" to remember...

It feels all too much like a source of wisdom, and I have a considerable fear of wisdom (the price seems always high). It seems to enforce a perspective I otherwise lack, and changes my thinking ...

And yet I am addicted to it. I will likely add tens of thousands of additional legacy images over the next few years ...

Unexpected results indeed!

Screening for berry aneurysms -- genomics

My PubMed RSS, embedded in Bloglines, continues to work beautifully.

Today's harvest from a month of fishing included 19 articles on Berry aneurysms. Much of the new activity is in identifying genes that predispose to forming berry aneurysms. As well as furthering the understanding of the disorder, it suggests that we may one day be able to identify persons at risk for subarachnoid hemorrhage and thus screen them by MRI. Whether this will add anything to our current screening criteria (two first degree relatives with SAH or BA) remains to be seen.

Biology of aging: telomeres predict your age

Fascinating: FuturePundit: Telomeres Wear Down Quicker In Men Than Women. It will be interesting to apply these techniques to different breeds of dog (age at different rates) and to identify slow aging humans. It will also be interesting to study middle-aged animals over a period of 1-2 years. I think we'll find that telomere erosion occurs in "bursts" rather than as a continuous process. The old saw about "he aged 10 years in a night" will turn out to be somewhat true. Then we'll study the mechanism of "aging bursts" and identify interventions.

Open source Java and a classic strategic blunder

Sun, a company with 3.98 extremities in the grave, finally makes an interesting move. They claim they'll open up Java, giving away one of their gems (Java is still important on servers).
Chief Says Sun Plans to Offer Open-Source Version of Java - New York Times

One of his first appearances since taking the helm of the struggling company three weeks ago, the executive, Jonathan Schwartz, told a gathering of software developers here that Sun viewed open-source software as a major part of its turnaround strategy...
It's unlikely to save Sun, but at least a part of their legacy might live on. Sun is following the path of the 1970s computer hardware companies that once ringed the twin cities -- shrinking and becoming a "services" company.

The story reminds me of one of the great strategic blunders of the 1990s. Netscape and Sun fought bitterly over who'd deliver "thicker client" capabilities to the browser. Netscape had an Ajax-like vision of JavaScript (no relation to Java) and enhanced markup with session management (not sure whether they had the asynchronous model though). Sun wanted everything, including the browser to run in a Java Runtime Engine. Actually, they wanted the browser to go away.

Meanwhile Microsoft, in one of Gates' most brilliant and ruthless moves, blew away all of its old software strategy and embraced, extended, and corrupted both the browser and Java. Gates is an evil force for computer geeks like me, but there's no denying his strategic brilliance. Nobody could shoot their own horse like Gates in his prime.

Sun destroyed Java on the Windows desktop with an unending series of tiny updates -- each of which destroyed a generation of installed software. Sometimes there were unmarked incompatible versions!

Sun delivered a second fatal shot when instead of building a cache solution for applets they persisted in a ridiculous vision of ubiquitous high speed computing. Applets should always be instantly available. Oh, and forget version control.

In the meantime, Microsoft sprinted while Netscape and Sun fought like psychotic wolverines.

Sun and Netscape destroyed one another. A classic lesson in strategic incompetence. Bottom line -- never forget that the 'enemy of my enemy is my friend'. It's better to take half of a trillion dollar market than nothing at all ...

Tuesday, May 16, 2006

100 years: likelihood of human survival

Pharyngula writes:
Pharyngula: The Seed Crystal Ball:

Our Seed Overlords have submitted yet another question to their blogulous oracle, i.e., us: Will the "human" race be around in 100 years?

I don't think it's a particularly good question, I'm afraid. The answer is simply "yes". If the question were about prairie chickens, cheetahs, or chimpanzees, it would be a more challenging question, but with a population of 6.5 billion of us, I don't think there's much doubt. We'll be here. The only question is what state we and the world will be in. I'll speculate a bit on possible outcomes.

To which someone responds (comments):
... 100 years is now a very LONG time. In terms of the accelerating rates of change in our de facto environment (which is technological and increasingly virtual) it's probably equivalent to 30,000 years of pre-10K BCE living.

I can't imagine any ecosystem catastrophe that would wipe out humans. So the options are:

1. Wild tech: Gray-Goo nano disaster, some weird vaccum energy thingie, etc. I'd put these all down as unlikely.

2. Engineered plague: A pimply 15 yo in southern China is upset by social rejection and engineers a virus using his home biotech kit that wipes out humanity. I'd put this at 20% when you consider all the folks that might try this.

3. AI: If we ever did produce a sentient AI, and if it turned out that 'intelligence' scaled to an IQ of, say, 100,000, then all bets are off on everything. I'd put that down as a 40% probability over a course of 100 years.

Since we have to survive both 2 and 3 to make the 100 year mark I'd say:

.8 * .6 = .48 or less than a 50% survival probability.

PS. The Fermi Paradox (aka the great silence) is not comforting here.

Monday, May 15, 2006

The surveillance society: Us and Them

The surveillanace society. It's about "us" and "them". Arkin says this well.
Early Warning by William M. Arkin - washingtonpost.com

... The government's position is that if you are "innocent," you have nothing to hide. It is a new version of 'you are either with us or against us.' Massive monitoring is of course meant to find terrorists; I completely believe that this is not some 1960's enemies list politically motivated effort. But these post 9/11 programs signal a new and different problem.

People of Middle Eastern and South Asian descent and Muslims are potential terrorists, machine selected as "of interest."

Throw in there callers and travelers to Pakistan and Saudi Arabia, recipients of wire transfers, purchasers of fertilizer, flight school attendees. These are the new guilty until proven innocent.

Innocent means of course mostly white, mostly Christian Americans who accept that the government knows best and that the national security state is only after the bad guys and would never apply its new found capacities in any illegitimate way.

The government and its new seamless surveillance culture are build ing a digital dividing line, even in our own society. The assumption is one of an enemy in our midst.

The government's failure to provide for domestic tranquility and basic security in our homes is rewarded by more power for the government; "innocent" Americans are increasingly primed and frightened to accept that greater government surveillance is required by the realities of infiltration, ceding even more power. It is as much a way of thinking as it is a way of life...
I wired cash to my sister in Canada a few months ago. I was electonically interrogated to match me and my transaction to an identify database entry. For all I know that alone added me to a watch list - at least added a few points to my Threat Score. This is how it goes now. Everything you do now, you can justifiably ask yourself -- will this put me on The List?

Just being Muslim may take you halfway to The List. Supporting Greenpeace might add some points too. Who you know, who your family knows ...

Being a white evangelical right wing Bush supporter will tend to keep you off The List.

Anyone remember the internment of the Japanese-Americans? That was about 63 years ago; I have a rather good book at home that was self-published by a child caught up in that despicable American mistake. I wonder if one of the forty remaining journalists in America might be troubled to interview a survivors of that time...

There's an election coming. How you vote will matter, again (sigh).

Gore 2008

Gore the bore. Ok, now that we've got that out of the way...
Talking Points Memo: by Joshua Micah Marshall: May 14, 2006 - May 20, 2006 Archives

Andrew Sullivan just published an email from a reader who says it'll be Al Gore in 2008 for the Democrats, not Hillary. I could see it. I could totally see it.

I don't think Hillary is anywhere near as strong as she looks or as people seem to think she is. And Gore would be formidable.
I believe him when I read that he doesn't feel compelled to run. I believe Tipper Gore is deadset against it.

However, I also recall that his daughter wants him to run again. Daughters can be very persuasive. He'd have my vote of course. If he runs, I'd expect a late entry.

Troops at the border? Humor is indicated

And now, for a much needed humor break.Shrill hysterical laughter is the only response to Bush's lastest brainstorm -- deploying the national guard along the Mexican border. Damn, what about all those snowbacks [1]?

[1] I saw this term first used in an early 20th century newspaper article. I wish I'd noted the reference. It may be an older term than "wetback"; illegal french canadian immigration was a big deal in the northeast back then. I wonder how my wife's Grandfather ended up there ...

Secondary outsourcing

I think it was Gary Trudeau (Doonesbury) who first popularized a programmer covertly outsourcing his work to India. A similar thing is happening in schools (DeLong).

These students, when they go to work, will use the same techniques. DeLong is annoyed with the reporter's gullibility, but it's neat to see a widely predicted future happening. There's a generational gap here -- I lack the outsourcing skills these college kids will have. I have, however, long had ideas about other uses for this sort of resource network ...

PS. I think Vernor Vinge's Fast Times at Fairmount High is our best current guide to the world ahead.

Are cars overbuilt?

I was chatting with a colleague about how long modern cars last, when it occurred to me that today's cars mostly die of trauma. It doesn't matter that a car will run for 15 years; given our aging drivers it'll be smacked within 10 years.

So are cars overbuilt? Should we be buying cars based on the costs of replacing body panels? A side blow to a van with motorized doors may be prohibitively expensive to repair, but a van with manual doors might survive the same accident.

Sunday, May 14, 2006

Rove to resign?

Rove to resign post indictment?. I'd rate this one as high as 50%, based on a reading of the original source. There's a real journalist at the start of this blog thread.

Saturday, May 13, 2006

The Descent of Dog

I'm fascinated by the evolution of the dog, a creature with a remarkably flexible genome and a survival strategy the cuckoo must envy. The BBC gives us an update:
BBC NEWS | Science/Nature | Exploring the wolves in dogs' clothing

.... According to DNA studies, domestic dogs owe their origins to a wolf cub that probably fell into the hands of humans some 40,000 years ago somewhere in Southeast Asia.

... "Originally, according to DNA samples, it would appear that the domestic dog is most closely related to the grey wolf.

"Point-two-percent is the difference between domestic dog DNA and grey wolf DNA, whereas the difference between coyote DNA and dog DNA is 4%.

... Experts are divided on how wolves first entered the lives of humans. Some believe that a band of hunter gatherers took wolf cubs back to their caves, perhaps to act as early guard dogs.

Others believe that wolves adopted people - creeping ever closer to human settlements to scavenge on discarded food.

By about 15,000 years ago, at the time of the last Ice Age, they were probably living alongside humans, perhaps retrieving wild animals felled with axes or bows and arrows.

The earliest archaeological evidence of dogs as true pets dates back to 12,000 to 14,000 years ago.

That is the estimated age of the remains of an old woman holding a puppy in her hand excavated in what is now Israel.

... Inside a glass case is the skeleton of an Early Bronze Age dog dug up in Tell el-Duweir near Jerusalem - the site of the ancient city of Lachish.

"Finds like these are really interesting for scientists because they give us clues as to how and when dogs became domesticated," says Dowswell.

"Although this is a relatively late example - examples have been found dating back as far as 15,000 years ago, the end of the last Ice Age - it seems that the dog skeletons they are finding are similar to wolf skeletons but much smaller.

"As a reduction in size is one of the first signs of domestication, and the fact that it's been found in a human settlement, this shows us pretty clearly that that's what's going on."
There's a very long interval between when "dogs" divide from wolves and when they become pets. Shortly after they're known to be pets humans begin building civilizations.