Tuesday, June 07, 2005

Best analysis on Apple's MacTel switch (OS X on x86, aka OS X86)

Apple shifts to Intel: what is all the fuss about? | The Register

The Register is improving. I have to add them back to my bloglines subscription. This is the best analysis I've read thus far on the OS X IBM -> Intel switch (My preferred name: OS X86). The conclusion is pretty upbeat. I think Apple will have a bad year for PowerBook and Tower sales, but some of that business will divert to iMac (if only they had a reassuring fix for the heating problem, but that's why they're going to Intel) and Mac Mini sales. I do expect their stock to take a serious hit, but they have money in the bank. It's not a happy move, but they didn't have much choice.

This commentary is pretty good too. The claim is that Apple will outsource almost all of its manufacturing to Intel, presumably retaining their design expertise and (of course) the OS. One wonders then if Apple will allow Intel to sell Apple's sophisticated designs as Windows machines -- for a price of course. Such machines would emulate Apple's 'stealth' strategy, if purchased via Intel there might be an option to 'switch' the buyers with a second sell of the OS. I think it very unlikely, however, that Apple would allow any old PC vendor to run OS X. They'll have lots of ways to prevent that from happening.

Update: My own thoughts: ... and one Palladium to rule them.

I've posted previously about Intel's Palladium and DRM. Palladium binds a computer's identity to all of its transactions [1], and allows enforcement of very powerful Digital Rights Management. There exists a possible American and Chinese future in which it's quite legal to own a machine gun (a pathetic toy really, good only for killing a few people) but utterly illegal to own a non-Palladium computer. Palladium is where Homeland Security (security), Corporate America (security), healtchare industries (security) and Hollywood (entertainment) agree completely; it will be mandated one day for all corporate and governmental computer purchases. It will be required, eventually, for iTunes like distribution of high-definition movies and entertainment.

All of them. Intel owns Palladium. It could be forced to share it, but for now they control it.

Palladium, under some other name, is the future.

Ultimately, I think this decision is really about Palladium.

It will be interesting to see what Cringely writes.

[1] Binding a person to the transactions is more challenging. Palladium technology could be used to mandate biometric identification, but that's a weaker link.

No classic on Intel Macs

Daring Fireball: Classic Not Supported on Intel-Based Macs:

Quite a bit of children's educational software is still only available for classic. The emulation layer for Mac Intel appears to emulate a G3, so no AltiVec emulation either.

This is a very dangerous time for Apple. I would not want to be holding shares now; their sales will drop quite severely until we learn what the new machines will bring.

Computer Quiz: why would moving a PCI card cause attached hard drives to fail?

The saga of the failing Vantec was passing into memory. It was time to hook it up to my system again. Alas, the IDE cable didn't quite reach (old story). I had to move the drive controller (Paradise IDE card) to a slot closer to the top of my case. A simple procedure.

All was well. I started my backup. My system locked up completely. Not even a blue screen. What was going on?

At first it seemed the problem was Retrospect; it causes me much pain. I tried my favorite disk diagnostic -- running a defrag utility. Started ok. Next morning -- locked up. Was the drive overheating in the new location? Had I bent a pin? Was the IDE cable bad? Did the card not like the new slot? Was it another Norton Antivirus fiasco? Since drive letters had changed, had that cased a problem (I had the swap file on one of the secondary drives) with XP?

I remembered my XP admin events tool. There I saw these fatal errors:

- The device, \Device\Scsi\ultra1, did not respond within the timeout period. (XP treats the Paradise IDE controller as a SCSI controller.)
- An error was detected on device \Device\Harddisk2\D during a paging operation. (that's because of my swap file)

Hardware? Overheating? Software? Two long nights and a new air condition installation before I figured it out.

When you move a PCI Card in Windows, XP loads new device drivers for it (just as XP still loads drivers when you move your printer's USB cable from one port to another). In the old days XP loaded a copy of the current drivers. Post SP2 XP loads, it seems, the latest signed drivers. The signed drivers for this IDE card don't support large (>135 GB drives); I'd previously updated to the vendor's unsigned drivers. I forced XP to install the new drivers (you have choose the manual install by location, XP SP2 refuses to install unsigned drivers by any other means).

Problem solved.

And some people think TV shows are responsible for stressed brains and elevated IQ. I don't know how I figured this one out.

Monday, June 06, 2005

Newsweek's Baghdad bureau chief: It's bad in Iraq

Good Intentions Gone Bad - Newsweek World News - MSNBC.com

This reminds me of the letter home from the WSJ correspondent that contradicted what her own newspaper was publishing.


June 13 issue - Two years ago I went to Iraq as an unabashed believer in toppling Saddam Hussein. I knew his regime well from previous visits; WMDs or no, ridding the world of Saddam would surely be for the best, and America's good intentions would carry the day. What went wrong? A lot, but the biggest turning point was the Abu Ghraib scandal. Since April 2004 the liberation of Iraq has become a desperate exercise in damage control. The abuse of prisoners at Abu Ghraib alienated a broad swath of the Iraqi public. On top of that, it didn't work. There is no evidence that all the mistreatment and humiliation saved a single American life or led to the capture of any major terrorist, despite claims by the military that the prison produced "actionable intelligence."

The most shocking thing about Abu Ghraib was not the behavior of U.S. troops, but the incompetence of their leaders. Against the conduct of the Lynndie Englands and the Charles Graners, I'll gladly set the honesty and courage of Specialist Joseph Darby, the young MP who reported the abuse. A few soldiers will always do bad things. That's why you need competent officers, who know what the men and women under their command are capable of—and make sure it doesn't happen.

Living and working in Iraq, it's hard not to succumb to despair. At last count America has pumped at least $7 billion into reconstruction projects, with little to show for it but the hostility of ordinary Iraqis, who still have an 18 percent unemployment rate. Most of the cash goes to U.S. contractors who spend much of it on personal security. Basic services like electricity, water and sewers still aren't up to prewar levels. Electricity is especially vital in a country where summer temperatures commonly reach 125 degrees Fahrenheit. Yet only 15 percent of Iraqis have reliable electrical service. In the capital, where it counts most, it's only 4 percent.

The most powerful army in human history can't even protect a two-mile stretch of road. The Airport Highway connects both the international airport and Baghdad's main American military base, Camp Victory, to the city center. At night U.S. troops secure the road for the use of dignitaries; they close it to traffic and shoot at any unauthorized vehicles. More troops and more helicopters could help make the whole country safer. Instead the Pentagon has been drawing down the number of helicopters. And America never deployed nearly enough soldiers. They couldn't stop the orgy of looting that followed Saddam's fall. Now their primary mission is self-defense at any cost—which only deepens Iraqis' resentment.

The four-square-mile Green Zone, the one place in Baghdad where foreigners are reasonably safe, could be a showcase of American values and abilities. Instead the American enclave is a trash-strewn wasteland of Mad Max-style fortifications. The traffic lights don't work because no one has bothered to fix them. The garbage rarely gets collected. Some of the worst ambassadors in U.S. history are the GIs at the Green Zone's checkpoints. They've repeatedly punched Iraqi ministers, accidentally shot at visiting dignitaries and behave (even on good days) with all the courtesy of nightclub bouncers—to Americans and Iraqis alike. Not that U.S. soldiers in Iraq have much to smile about. They're overworked, much ignored on the home front and widely despised in Iraq, with little to look forward to but the distant end of their tours—and in most cases, another tour soon to follow. Many are reservists who, when they get home, often face the wreckage of careers and family.

I can't say how it will end. Iraq now has an elected government, popular at least among Shiites and Kurds, who give it strong approval ratings. There's even some hope that the Sunni minority will join the constitutional process. Iraqi security forces continue to get better trained and equipped. But Iraqis have such a long way to go, and there are so many ways for things to get even worse. I'm not one of those who think America should pull out immediately. There's no real choice but to stay, probably for many years to come. The question isn't "When will America pull out?"; it's "How bad a mess can we afford to leave behind?" All I can say is this: last one out, please turn on the lights.

Will a volunteer army fight the Forever War?

US Officers Plot Exit Strategy from the Open-Ended 'War on Terror' (LA Times 5/22)

The Forever War was a classic science fiction story:
Yet Tuohey, who was promoted to captain upon returning to Ft. Hood, said he was not sure whether he would stay in the Army when his commitment ended next year. He said he was tempted to work on Wall Street.

It's not the money he's after. It's the fact that an Army that was gutted after the Cold War was promising him a future of perpetual deployments fighting a war that could last for decades.

That is not a future he is sure he can commit to.

"What's the end point?" he asked. "When do you declare victory?"
The LA Times article goes beyond the well known problems with recruiting high school graduates to the potentially more severe problem of retaining young officers. These officers are reasonably comfortable with their mission and their command, but they can't face repeated counter-insurgency deployments. They article doesn't go into why, but I assume that most rational warriors have one or two of these types of engagements in them -- but not an unending variety. (I don't believe I could manage one of them, myself).

Internal exile: the fate of those who fall

In a Transparent Society those who carry the scarlet letter are doomed to internal exile:
Barred From the Long Haul - New York Times

... Requiring drivers to have background checks before receiving hazardous material certifications makes perfect sense. But the law, as interpreted by the Transportation Security Administration, singles out law-abiding ex-offenders whose criminal records have nothing to do with terrorism or national security. The new rules, which went into effect at the end of May, could potentially worsen a national trucking shortage and kill off valuable training programs that bring former convicts back into the work force by teaching them to drive trucks and helping them obtain the necessary licenses.
A local paper ran an article in the same theme today. We seem to be flying by the 1950s towards the 1630s.

I never thought I'd see the day when I really missed Newt Gingrich. By comparison to the puritan troglodytes who rule us now Newt was a paragon of reason and thoughtfulness.

Sunday, June 05, 2005

Fixing Consumer Reports: what would I do?

I wrote earlier about where I think Consumer Reports is failing to deliver value: Faughnan's Notes: Consumer Reports: the paradox of a consumer organization that hates its customers.

So what would I do if I ran the place? (I'm sure you're keen to know.) Well, if it were possible to provide CR with feedback, this is the plan I'd propose. (Maybe someone else will create this business?).

1. Launch a new initiative called 'Quality First'. Plan a marketing campaign around it.

2. Quality First has two components, one for cooperating companies, the other for all the rest.

2a. Cooperating Companies: The Quality Audit
- cooperating vendors agree to provide an auditing firm with the data they have on return rates, defect rates, service satisfaction, etc. They agree to survey customers who have service done on their satisfaction with the work. They agree to participate in staged service calls run by consumer reports.
- cooperating vendors get special placement in all CR reports. They are 'Quality First Gold Members'. All ratings of all products have a new column called 'Quality Gold'; participants get a check, others don't. Quality Gold members get a separate call out rating.
- in Phase II of the program participation in the Quality Audit, and the 'Quality Ranking (see 2b) become important factors in overall rankings.

2b. Non-cooperating companies: Quality Ranking Estimated
- non-participating vendors don't get the Quality Gold benefits.
- non-participating vendors still get quality estimates done by reviewing Amazon user reports, web complaint sites, CR member surveys, etc. When in doubt they get low quality rankings. If they don't like it, they can join the Gold program. Complaints must be weighted by sales volume of course.

3. Quality Rankings
- Quality rankings are calculated based on warrantee periods, service records and surveys. They become a strongly weighted part of product rankings.

There, I've done CR's business plan for them. If they take it, it could revitalize their business. Too bad there's no way to send it to them!

Consumer Reports: the paradox of a consumer organization that hates its customers

Consumer Reports Ratings and recommendations available at ConsumerReports.org

This would be funny, if it weren't so irritating. I think there are broader lessons here than a single dysfunctional company. Many of us feel the pain of "brand reputation and quality" in a world where price competition is severe and consumers are overwhelmed with choices and making consistently "wrong" (IMHO) value decisions. Consider the story of Consumer Reports and my quest for a reliable air conditioner.

Consumer Reports is an old (once non-profit?) consumer oriented organization. It's not ad supported, I pay to get access to their web site. I went there looking for a quality of service and repair record information on air conditioners [1]. It might be there, but I couldn't find it. So I decided I was annoyed enough to send them some feedback.

So I find it quite amazing that, after 15 minutes of searching, I couldn't find any way to send them feedback on CR's services [1]. Maybe it's there, but it's professionally hidden (they've outsourced their web support to a separate company). I gave up. I expect that kind of thing from Amazon.com, but from Consumer Reports?

Did they get bought out by a some evil entity? Alas, the more likely explanation is that they're far more interested in their newly launched health services rating and education program than they are in providing the services I'm interested in. I suspect a deeply introspective and dysfunctional service company in a non-competitive niche. I wish they'd provide the services I need, but they so dominate this niche that I'm probably still better off paying for the crummy service they provide.

Of course give me an alternative and I'll be gone in a heartbeat ...

Quality of Consumer Reports and their outsourced customer support operation. Quality of Samsung air conditioners. Quality of Best Buy and their outsourced repair services. Getting data and providing feedback. It's a theme.

---

[1] The feedback was that I don't care about 'air conditioner' test attributes as much as I care about reliability records and quality of service; CR doesn't provide the key data I want. In this case we'd bought a highly rated Samsung air conditioner from Best Buy. It worked for about a year, then it apparently lost coolant. We had it serviced under warrantee by a Best Buy contractor, after a one week repair, and upon later reinstallation, it still doesn't work. We don't have time to waste on this stuff, so we wrote off both Best Buy and Samsung (sorry guys, you only get one chance to deeply annoy us). I went to CR looking for some repair and service measures and couldn't find them. No wonder vendors don't bother to invest in quality!

Update 6/5: To be fair to CR, I wrote out a business plan for their recovery. In a not-unrelated event, today at the gym my one year old delicately handled SONY walkman died on me. I'm replacing it for now with a much abused "Jensen" device that I bought at a garage sale, it was made in China perhaps 8 years ago. Grr. The aggravation of disposable devices is never-ending.

Saturday, June 04, 2005

Self-replicating device -- another step on the road

CNN.com - The machine that can copy anything - Jun 2, 2005

A device that can supposedly create some of its own most critical components. If it can also create other interesting devices, in theory it could allow massive distributed manufacturing. It's not 'grey-goo' nano-stuff, it still needs conventional raw materials and human activity.

Remember that quaint 1970s book by Toffler called 'Future Shock'? He greatly underestimated the ability to humans to adopt to change. The Future Shocks he described described are now beneath our everyday notice.

I wonder what our adaptive limit will be?

Friday, June 03, 2005

Apple has crummy server performance?

AnandTech: No more mysteries: Apple's G5 versus x86, Mac OS X versus Linux

A very technical analysis finds the G5 to be a reasonable workstation choice, but OS X fares poorly as a server solution.
The server performance of the Apple platform is, however, catastrophic. When we asked Apple for a reaction, they told us that some database vendors, Sybase and Oracle, have found a way around the threading problems. We'll try Sybase later, but frankly, we are very sceptical. The whole 'multi-threaded Mach microkernel trapped inside a monolithic FreeBSD cocoon with several threading wrappers and coarse-grained threading access to the kernel', with a 'backwards compatibility' millstone around its neck sounds like a bad fusion recipe for performance.

Workstation apps will hardly mind, but the performance of server applications depends greatly on the threading, signalling and locking engine. I am no operating system expert, but with the data that we have today, I think that a PowerPC optimised Linux such as Yellow Dog is a better idea for the Xserve than Mac OS X server.
I think they used 10.3 for testing. I wonder how the server results would look with 10.4? Apple made major challenges to the threading model.

Neandertal shall rise again

BBC NEWS | Science/Nature | Extinct cave bear DNA sequenced

Sequencing cave bear DNA was the test case. The real focus is on Neandertal DNA. Can anyone doubt that within 40 years a Neandertal infant will be born?

Revenge of the Bell Curve: selection for IQ in European Jews

Researchers Say Intelligence and Diseases May Be Linked in Ashkenazic Genes - New York Times
...Ashkenazi Jews occupied a different social niche from their European hosts, and that is where any selective effect must have operated, the Utah researchers say. From A.D. 800, when the Ashkenazi presence in Europe is first recorded, to about 1700, Ashkenazi Jews held a restricted range of occupations, which required considerable intellectual acumen. In France, most were moneylenders by A.D. 1100. Expelled from France in 1394, and from parts of Germany in the 15th century, they moved eastward and were employed by Polish rulers first as moneylenders and then as agents who paid a large tax to a noble and then tried to collect the amount, at a profit, from the peasantry. After 1700, the occupational restrictions on Jews were eased.

As to how the disease mutations might affect intelligence, the Utah researchers cite evidence that the sphingolipid disorders promote the growth and interconnection of brain cells. Mutations in the DNA repair genes, involved in second cluster of Ashkenazic diseases, may also unleash growth of neurons.

In describing what they see as the result of the Ashkenazic mutations, the researchers cite the fact that Ashkenazi Jews make up 3 percent of the American population but won 27 percent of its Nobel prizes, and account for more than half of world chess champions. They say that the reason for this unusual record may be that differences in Ashkenazic and northern European I.Q. are not large at the average, where most people fall, but become more noticeable at the extremes; for people with an I.Q. over 140, the proportion is 4 per 1,000 among northern Europeans but 23 per 1,000 with Ashkenazim.
A 600% relative increase in near-genius. If the averages are in fact similar, however, this may imply a roughly 600% increase in the rate of retardation (autism?). (Note if 1/250 humans are almost-genius, imagine how clever the 6,500 1/1 million minds on earth are.)

This article appears in the Times next to one about the genetic determinants of gender preference in fruit flies. A useful coincidence. I'm amazed this article was written without mention of the infamous Bell Curve book. Similar comments about ethnicity and IQ, by the way, have been made about Scots and South Koreans.

If I were to bet I'd guess most of the diseases the authors refer to are in fact 'founder effect' diseases and unrelated to a selection for extreme IQ variability (both high and low); in other words, the main point of the paper is incorrect. I would further guess that the IQ variability is real and is genetically, not environmentally, determined.

I'd also guess that the IQ variability is primarily among males, and that the "genius" outcome operates through the same gene family variation produces autism in some people.

Of course the extremely incorrect aspect of this article is that if an ethnic group has a selection for higher IQ, the converse is also likely. Let's assume this were found to be so. Consider now my recent comments on the genetics of behavior. IF we have still have an advanced technological human culture in 2025, I suspect "meritocracy" may carry little more moral approval than plutocracy. Both represent the outcomes of chance.

Omelettes not in America

Fafblog! the whole worlds only source for Fafblog.
... you may have accidentally happened upon a few bodies halfway across the world (Afwhatsistan? Bagrawho?), which may or may not have pricked whatever remains of a long-dormant and desensitized National Conscience. And you may be asking yourself what the point of all this has been, what has driven Americans halfway around the globe to sieze innocent men, beat their legs to pulp, and chain them to ceilings until they die.

Regrettable, yes, but let us remember that these two eggs, like the dozens before them, and the tens of thousands before them, were broken to make the greatest and worthiest of omelettes, the most succulent of breakfasttime generational commitments, the proudest and most visionary of truck stop slop.
And there's more.

One gene, one gender preference

For Fruit Flies, Gene Shift Tilts Sex Orientation - New York Times

Gender behavior in the fruit fly is specified with a single gene.
"We have shown that a single gene in the fruit fly is sufficient to determine all aspects of the flies' sexual orientation and behavior," said the paper's lead author, Dr. Barry Dickson, senior scientist at the Institute of Molecular Biotechnology at the Austrian Academy of Sciences in Vienna. "It's very surprising.

"What it tells us is that instinctive behaviors can be specified by genetic programs, just like the morphologic development of an organ or a nose."...

...The finding supports scientific evidence accumulating over the past decade that sexual orientation may be innately programmed into the brains of men and women. Equally intriguing, the researchers say, is the possibility that a number of behaviors - hitting back when feeling threatened, fleeing when scared or laughing when amused - may also be programmed into human brains, a product of genetic heritage.
One might think a fly is quite different from a human, but genes that code for things as fundamental as gender preference tend to be highly conserved by evolution. If there's a similar gene in humans it may well have a significant effect on male gender behavior. (Gender preference in female humans is thought to more fluid than in male humans; we may not be the same as the fruit fly.)

The more we learn, the more programmed and machine-like humans seem. The research has been pretty consistent over the past 20 years. When popular books strive to preserve a role for the environment they provide examples that are frail and tend to fall quickly to further research. In contrast genetic control of behavior has held up well. Identical twin studies seemed to preserve more room for the environment, but since then we've learned that there's a large amount of variability in gene expression even among identical twins (esp. female twins).

What will we do with this knowledge? How will it alter our thinking on self-determination, on "merit", on responsibility, on punishment? If modern Republicanism rewards the luck of parental wealth, does not the meritocratic alternative reward the luck of parental genes? (Parental wealth and parental genes, of course, are generally correlated in any event.)

If we come to see all fortune, goodness and badness as merely the expressions of random chance, will we look differently at winners and losers alike? Will we one day return to the marxist doctrine of 'from each according to their means, to each according to their needs'?

Come back in twenty years and let's see.


Coalition casualties in Iraq: tally, details and news

I was thinking this morning of how little I know about what's going on in Iraq. It's a reasonably large, populous, and complex country. A bigger Yugoslavia. With Oil. Occupied too.

I wondered again why nobody seems to have a web site that that provides an english language digest of the editorial and news pages of 6 or so Kurdish, Sunni and Shiite daily newspapers.

The New York Times could do this. Forget about investing in journalists in Baghdad. They're expensive and prone to be kidnapped and murdered. Translate and summarize some Iraqi newspapers instead. Safer and more effective.

Odd.

While looking half-heartedly for such a thing, I came across this: Iraq Coalition Casualties. They have a news summary from the usual suspects (BBC, Guardian, etc) but also a tally of coalition (not Iraqi) casualties. The totals link to details and thus to death notices. They also provide links to similar pages in the "about" page. Somewhere to go when your child next Memorial day asks about soldiers.