* Posts by Donald Becker

138 publicly visible posts • joined 3 Dec 2007

Page:

Pilfering sysadmin gets four years and $2.3m fine for kit theft

Donald Becker

Isn't this what component serial numbers and associated bar codes are for?

And, to keep it on-topic for Friday, is the proper phrase "rounds of plastic surgery". Sure, a round might be a pair of hemispheres, but it could be more conical shaped.

Liquefied-air silos touted as enormo green 'leccy batteries

Donald Becker

Re: No idea how well this may work.

Yes, putting some of the waste heat into a gravel bed would work.

But it's back to the same problem: for efficiency you need either a massive plant or lots of time. In this case a massive gravel bed, which negates the energy density of liquified air.

Donald Becker

Compressed air has horrible efficiency because of the dynamic range of pressure.

Using liquified air reduces the problem with varying pressure, but the thermodynamic efficiency is still very low.

If you compress the air to liquify it, you need to get rid of lots of low-grade heat. It's hard to extract the energy from that heat without making the compressor work harder.

You have the problem in reverse when you let the liquid boil -- you have to keep putting in heat. You might try to recover energy from the temperature difference, but you need either a very large plant or lots of time.

Besides the giant heat exchangers, you also need massive pressure tanks for this scheme. A failure will result in a cryogenic liquid spill. The cleanup will be easy, but most equipment it touches will need to be scrapped and the humans buried. Not everything that goes wrong can be fixed with a wrench.

Tesla drops veil on top secret solar Superchargers

Donald Becker

"Or you could wait a week or two between charges"

Only a week or two? Yours must be quite small.

You can only put about 100 watts of PV panels on a car. You might be able to get a bit more power out if you park in the corner of the parking lot that has no shade, tilts to the south, manages to point all PV surfaces towards the sun, is always at midday, during the summer.

Estimate that you get that power output for 4 hours a day. 6 to be generous. So we'll call it 500 watt-hours a day.

The smallest battery you can get in the Tesla Model S is 40KWh, or 80 sunny days of charging. The large battery is 85KWh. Counting seasonal variations in solar radiation, weather, and a bit of self discharge we'll just round up to a year.

It takes about 400 watt-hours to go a mile at highway speed. Some have managed 250 watt-hours in efficient vehicles. The Tesla Model S is custom designed to be quite efficient, and they only claim 250 watt-hours with the smallest, lightest battery configuration. So you could get a 1-2 miles per day from PV panels mounted on a car.

Donald Becker

Re: Hmmm...

"Is this just a thinly disguised way of farming subsidies from some Californian feed in tariff system?"

Yes, and it's not just California.

Home-scale producers are can inject power into the grid. They are paid for it at the full retail price. It's a major, high-value hidden subsidy.

The retail rate pays for the electric distribution system, maintenance, capital risk, pricing risk, etc. It's approximately double the wholesale rate. And even the wholesale rate is the wrong reference, since the distribution company can choose to buy at that rate, while being forced to buy home solar/wind excess.

Right now it's such as minor part of the grid that the subsidy isn't distorting the electric market. But the subsidy is distorting the solar PV cell market. It's encouraging the installation of solar panels in sub-optimal locations (cloudy locales, incorrectly angled roofs that don't face south, etc) instead of high-value locations (remote locations with long high-loss connections to the grid). In California you'll find PV installations in foggy San Francisco are more common than in sunny high-altitude rural areas.

'Call Corporate Of Apple and tell them to stop there persuit!!'

Donald Becker

A bit misleading to pick a badly written comment on the discussion part of a site, and make it look as if the site itself is run by illiterate fools.

Hmmm, pretty much completely wrong about the Google v. Oracle case. The jury didn't decide that Google infringed on Oracle's rights. The question posed to them was a conditional 'if APIs are a protected right, did Goggle use them'. The jury could only answer "yes".

But the judge still has to rule if APIs are protected by copyright. Long practice and precedent says that they are not. If they are not, the question and jury verdict is moot.

Java jury finds Google guilty of infringement: Now what?

Donald Becker

A minor factual correction.

The jury did find that Android did include copied code. Nine lines. Out of 15 million.

There were other claims, such as the similarity of simple comments describing functions. Pretty much "Takes X and returns an integer Y". There are many ways to word this idea, but only a few simple ways, thus the comments look similar. The jury agreed with Google that this was not evidence of literal copying.

The 9 lines of code aren't worth "up to $150,000" as a copyright violation. That's only if there was knowing, willful infringement. Google removed those lines as soon as it was pointed out to them. Given that Sun's copyright registration was flawed, the statutory damages would typically be $200.

In a different setting there would be a defense stated that the usual threshold for infringement starts somewhere at 12 to 20 non-trivial lines of code. But the time to say even a single sentence in this trial is costing far more than $200 in legal fees. Boise reportedly got one sentence into trying to make it an issue before the court, perhaps trying to claim a partial win. The judge made it clear that Oracle had completely lost and wasn't going to get an accounting for 9 lines out of 15 million.

'Intelligent systems' poised to outsell PCs, smartphones

Donald Becker
Boffin

Re: x86 will continue to shrink in embedded.

I've read that a Cortex-M license brings in just under 25c. That probably means well under 25c for the bottom-end parts, and noticeably more for the latest M4 . Those don't run Linux, but are easier to write code for than simpler parts. The code density is pretty close to the alternatives, and the power usage is usually much below the 8 bit parts.

Running Ubuntu, I didn't have to modify the Canonical-packaged GCC-based ARM toolchain to compile. (But I didn't need to for the AVR either.) I did end up writing my own linker definition file and many header files to match my preferences, but they now can be readily downloaded. Of course "readily" means extracting and installing myself, rather than the checkbox install for the compiler.

Researchers propose ‘overclock’ scheme for mobiles

Donald Becker

Amusingly, this story appeared adjacent to an ad for the Tegra 3 based Transformer Prime.

For those that don't know, the Tegra 3 has "a knob that goes to 5". It runs on a slow, efficient low-voltage core while mostly idle, and turns on up to four much faster cores as needed. It already uses every technique suggested in this summary, and many others.

US entertainment lawyer casts doubt on Megaupload case

Donald Becker

Who protects the unpopular?

Kim Dotcom isn't someone I can get behind personally. He is pretty much at the opposite end of that scale. But the U.S. government has really overstepped. Legal protection and due process apply to everyone, not just to those we like.

I see the possibility of civil suits. Perhaps a bunch of them. But not a vast criminal conspiracy.

Scientists snap amazing technicolour dreamtoad

Donald Becker

What are you talking about?

What's this "hynotoad" thing you talking abo.... ALL HAIL THE HYPNOTOAD

Google+ disk space cockup creates notification spam-storm

Donald Becker

New backup server

Hmmm, my first thought: "Has someone found a way to use G+ as a backup server, and spread the word using a social network?"

Wouldn't you have to consume 10-100GB at a time for the service to run out of space?

LightSquared admits it will knock out 200,000 sat-navs

Donald Becker

Billions and billions and "billions"

Hmmm, "already spent billions" on reducing interference? Perhaps they have spent money developing the business plan and preliminary hardware. But that's not quite the same thing. More accurate would be "spent money trying to create a business that could be worth billions", but that's hardly likely to garner sympathy.

The "5 cents" would buy a filter that would reduce the signal from other frequencies, but it would also make the receiver less sensitive in the target band and change the phase response. Receiving a GPS signal isn't like building a AM radio with a pencil lead. The data is far below the noise floor. Thermal noise alone creates significant positional uncertainty. Throwing away information by reducing the signal strength is many times worse.

Military GPS receivers do use a different band. But that's in addition to the "civilian" band, not instead of. (They use both the L1 and L2 signal propagation delays combined with the current atmospheric delay model to compensate for the atmospheric lens effect.)

Google whacks link farms

Donald Becker

Not only long overdue, just in time

A high enough percentage of my searches have been returning non-results that I've been ready to switch to a different search engine.

There are still many, many areas that they have not fixed.

For instance almost every search for an electronic component datasheet has been blanketed with "buyers" pages. They claim to have the datasheet pdf in the page synopsis, some even have ".pdf" in the URL, but when you go there it's just a generated list of distributors and links to do a search for the datasheet.

The results have gotten a bit better in the past few days, but Google still far to go. They ignored search result quality long enough that SEO grew into a big business. Now there are tens of thousands of people that spend all of their time trying to game the system. It might already be an entrenched industry that forever degrades search results.

Feds crack phone clone scam that cost Sprint $15m

Donald Becker

"if credit card companies can do it why not Cell phone Networks"

>> if credit card companies can do it why not Cell phone Networks.

Easy answer: you sign or otherwise authenticate credit card transactions. You are also likely on camera for bigger transactions. So you can avoid paying for fraudulent charges.

With a phone, they just assert you made the call. You have to prove that you didn't, which is very nearly impossible. My guess that well over 90% of those falsely billed either didn't notice, or just gave up and paid rather than spend hours trying to get the charges reversed. For Sprint, 90% payment for 'business' that they wouldn't otherwise have is very profitable. There is little motivation to stop it until the problems become too public.

Consumer Reports: 'We were wrong about the iPhone 4'

Donald Becker

Not entirely off-topic: it really is 'duck' tape

The name comes from the original structure of the tape. It was made of cotton duck, a type of cotton canvas, with a gum adhesive. It became wildly popular with the military in WWII.

It's actually really bad for sealing ducts -- not the right product at all.

Secrecy fetish drove Steve Jobs' analytics bust

Donald Becker
FAIL

Errrmm, what about the *users*

The discussion seems to be only about what Apple wants to keep secret. And Flurry is talking about how they shouldn't have done what they did -- reveal information about devices inside of Apple.

No one in the story had the slightest concern about what information was being gathered about the users of the devices. Are they just the suckers ^W users?

Firefox's oldest friend dumps it for Google Chromium

Donald Becker
FAIL

Hear of it passing, didn't think that it was still around

On a similar topic:

I checked on favorite band's Myspace page today to find their summer schedule. It hadn't been updated in about a year. The same was true of most other pages I checked. I can only attribute the sudden crash to a worldwide shortage of pink glitter Sparkle Pony GIFs.

It's too early to predict that all social media will meet the same fate. After all, AOL is still around. I think.

Microsoft defends death of free video in IE 9

Donald Becker

Isn't it ironic

The ironic thing is that Microsoft reportedly put a great deal of money into developing a video encoding standard that they could own, expecting that they would have a lock on basically every consumer electronic device. They pushed it as a standard, and started to succeed. But that brought patent trolls (and one or two legitimate patent holders). Now, instead of getting a cut of every computer, media player, cell phone, etc. sold, they have a net outflow of payments.

But even with a net ongoing cost, it's more important to Microsoft that the market be costly and controllable.

Google Street View logs WiFi networks, Mac addresses

Donald Becker

MAC address doesn't leave the local network

A detail mentioned by a previous poster (but gotten wrong by many others) is that the MAC address is a local identifier. It doesn't leave the local network. While it's possible for a protocol to retrieve and report it, it wouldn't serve a purpose so no common ones do that.

The only straight-forward use of this information is approximate location when you don't have a GPS receiver. It would need to be combined with some simple criteria for throwing out bogus results to handle moved access points.

I can see some secondary uses in understanding the local popularity of different brands of equipment, and overall communication use by access point density. But it's hard to consider any of that an intrusion of privacy, or having any impact on anonymous access to information.

Asus' Eee keyboard out next month - official

Donald Becker

Delayed because of consumer rejection?

Wow! This looks great!

Do I plug in a USB keyboard, or do I use a BlueTooth one?

Because there is no way I want to drag a thick video cable plus power cable plus whatever else around with the keyboard I use.

Microsoft opened Linux-driver code after 'violating' GPL

Donald Becker

The non-complexity of the GPL, the complexity of trying to cheat

The GPL isn't complex.

The frequent arguments over it come from the attitude "how can I use this and pass it off as my own code?", followed by "but what if I do this...". Often followed by "... but what if I really want to?"

Silicon Graphics goes titsup (again)

Donald Becker

Just the assest, not the debt

Hmmm, my reading of the filing is that RACK is buying only the domestics assets, likely including the brand name, and the intact international operations. It doesn't appear that they are taking over the obligations.

The senior debt holders have first dibs on the $25M and remaining cash. Typically a little is set aside for other creditors -- common so that they accept a few cents on the dollar rather than cost legal fees all around, leaving everyone with zero.

Creditors have 60 days to file with the court. At the 60 day point the already-written plan will likely be rubber-stamped by the court. SGI has already been shopped around, and the filing attorneys probably have fully documented that this is the best deal they could get. A "white knight" could still show up, but they would have to ride in from an alternate financial universe.

It's not in anyones interest that this drags on. SGI has needed frequent cash infusions to continue operating. Even if Morgan Stanley had the desire to continue this, they no longer have the cash to do so.

Brussels: Old-school lightbulbs to be gone by 2012

Donald Becker

The wrong target for regulation

At first blush, this sounds like a good idea. But it's really just a bad idea with a green coat of paint.

There are many lighting uses where simple, cheap and mostly reliable incandescent lamps are still the right choice.

Consider a rarely used attic light. The one in my parents house is probably used well under an hour a year. It would still have the original bulb if I hadn't replaced it with a fancy version that turned itself off after 30 minutes. A CFL lamp in the same location would save a trivial amount of energy, would probably die quickly in the summer attic heat, and would probably not light on that cold winter day we take down the Christmas decorations.

A far better approach is one that's already common. Most regions in the U.S. have building codes that require at least one lighting fixture at key locations that meets minimum specifications -- usually implemented as florescent fixtures for kitchen and bath lights. That could be expanded to cover more living areas. Even better, allow lower cost cable (14 or 16 gauge) for circuits that serve only high efficiency lighting.

There is one product that I could see some regulation around: "long life" light bulb that are rated at 130V instead of 120V. At 120V they are very inefficient and will put out much less light than their rating. Worse, they keep on taking -- at 120V they will never burn out and be replaced with reasonable lamps. These are almost a fraud, since no standard system in the world runs at 130V.

Microsoft claims Firefox- and Chrome-whopping IE8 speeds

Donald Becker

Is "Visual clues" the clue?

I suspect that the other posters have identified the key phrase -- "visual clues".

Microsoft has optimized the engine to give priority to reading on-screen images from cache and running scripts that impact the layout of the part of the page being rendered.

There are many ways that this might be tied into special OS APIs. One way is using asynchronous I/O with priorities. Another is giving storage layout hints when initially caching e.g. grouping objects together on the same disk cylinders (hmmm, which would give the best results with a fresh install.. coincidence?)

If that's what Microsoft is doing, they are making a traditional Microsoft move. Tying an application tightly to an optimized (but private and obscure) OS interface looks great for a few years. Then third-party applications figure out how to use it, just as it's time to abandon the interface to let the OS design move forward.

It's possible that the IE development team isn't doing that. They might just be optimizing the display engine to prioritize rendering the visible part of script-based web pages faster, without using special OS APIs. Pretty much as interlaced GIF images could be displayed with low resolution before finishing a page load over a modem 15 years ago.

But my bet is on funky OS APIs.

Pirate Bay prosecutor tosses infringement charges overboard

Donald Becker

Who has really won? Certainly not the defense.

The defense has already spent the time and money to prepare for the trial. While trial lawyers are paid a higher rate for time in court, it's still only a few days compared to the larger set people involved in the months of preparation.

Not to speak of the stress on the people involved. I've been in that situation -- a meritless civil case where the other side delayed and did everything to raise our cost. The personal stress was immense. And even a meritless case is a lottery ticket.

Sanctioning the prosecution doesn't do anything positive here. They didn't have anything at risk. The only effective balance would be prosecuting the people or companies that put them up to the original prosecution, but those people are out of reach.

This isn't a victory. It's another lesson in how an industry can use the legal system to harass, thus create de facto rights even beyond copyright laws. And a lesson that it happens all over, not just in the U.S.

NEC, Nissan lob $1.1bn at electric car battery biz

Donald Becker

Valuable contribution to the future

This has much more potential to make the world better than Tesla Motorcars. Tesla was the trendy poster child for electric car, and people talked about them as if they were making a difference. But it was more of a stunt than progress. They were connecting together thousands of lower-end off-the-shelf laptop batteries and making the battery pack fit into an existing chassis. There was a lot of design and engineering work to make it function, but none of that work carries forward to a cost competitive electric car.

Why am I being so harsh? Because almost every main-stream press story about Tesla has been "it goes really fast on only pennies of electricity! It requires no gas or service! It's the future!" The stories hasn't evolved to the point where they talk about if it's progress.

My experience with laptop batteries tells me that in less than a year the owners will notice a significant performance decrease, and in two years of daily use the battery will be at 30% capacity with a major loss in performance. Most of the buyers of a high-performance car won't want a medium-performance car next year, and a low performance one the following year. After one or two $20-30K replacement batteries the car won't be worth running. They will turn into garage queens, kept in refrigerated rooms to keep the batteries from aging. ("Pennies in electricity to drive, kilobucks of refrigeration to park.")

Accused Scareware mongers held in contempt of court

Donald Becker

Belize doesn't have jurisdiction

I'm normally sensitive to jurisdiction issues, but in this case it appears that all of the players are in U.S. They are targeting U.S. consumers, taking orders in the U.S., and cashing the check in the U.S. The only thing in some hard-to-reach country is the corporate registration of a front. The money probably gets stuffed into an account in a different hard-to-reach country, in a way that's difficult to track.

Tesla takes Top Gear test to task

Donald Becker

"theroretically possible"

The claim that it's "theoretically possible to charge the Tesla's batteries in a few seconds" is wrong. The Tesla is using standard 18650 lithium cells -- low end laptop batteries from existing production lines. (High end batteries are now custom shapes, rather than standard cylinders.) They take at least two hours to charge, and produce a lot of heat, especially while finishing their charge. They will briefly accept a high charge rate, but not for very long before overheating.

The idea that you can just drain and refill the electrolyte is similarly flawed. For a lead-acid battery, it's not the acid that's consumed. The lead plates are chemically changed.

And finally, swapping batteries isn't (yet) practical. Even if a battery pack is standardized, and the engineering is worked out to mount it safely and quickly (very difficult), working out the economics will be difficult. Would you want to swap your factory-fresh battery pack for a thrashed pack with much reduced capacity on the first long trip?

Daft list names Firefox, Adobe and VMWare as top threats

Donald Becker

Yes! It should be riskier doing press releases

I don't agree that The Register should avoid giving more press to sleazy vendors like this.

How many IT publications take a press release like this, shuffle a few words around, and print it as a news story? Almost all of them.

The Register is one of the few that looks at the claims and says "maybe you shouldn't take this at face value". If they didn't do this, only the positive echo-the-press-release stories would be out there.

Electronic votes mysteriously vanish in Ohio election

Donald Becker

Even more interesting with the history

The back story here -- the part that makes these stories gratifying -- is the many years that Diebold has resisted any review. There was a small group of transparency-in-voting advocates (w/ more than a little overlap with Open Source advocates) pushing for publishing the source code for these machines. Or at least making it readily available for expert external review. For years Diebold insisted that their code had already been throughly reviewed, and that they shouldn't have to reveal their highly proprietary method of incrementing counters.

It turns out that those advocates were right. Not just "not wrong". But Completely Right. After the first confirmed errors, no apologies or changes. Nor after the second. Or after the continuing series of subtle errors. These were all "rare" errors, yet almost every careful, time-consuming review finds new ones.

This would never happen, but the original people protesting the Diebold closed system should be generously funded to be advocates for new, carefully designed, verifiable and open voting systems.

FCC boss gets knuckle-rapped

Donald Becker

A Whitewater-like investigation that found nothing?

Shouldn't this story be titled "After a year of a broad, politically-driven, investigation, no illegal or unethical actions were uncovered." With text describing how the staff didn't like having to listen to the appointed commissioners instead of following their individual agendas.

Israeli Linux fan squeezes Windows refund out of Dell

Donald Becker
Go

Rejecting the EULA gives you default right-to-use

It seems that most of the people are missing the point: the EULA is purporting to limit their normal rights.

The purchaser bought, at retail (e.g. with no power of negotiating a contract), a tool. After purchase they were presented with limitations on how they are allowed to use the tool -- limitations that go beyond the "first sale doctrine". (Yes, that's a U.S.-centric phrase, but many countries have similar principles.)

According to a plain reading of the laws, the purchaser has an option to reject the limitations. The vendor has to provide an option for a refund, or accept that the EULA limitations have been rejected and the user has their default legal rights. What might those rights be? Resale without restrictions. Moving the single copy of the OS to another machine. Using the fonts and images with other programs (on a single machine). Running the programs and libraries under a different OS kernel.

What the users are really asking for is either the right to use the installed system without the limitations of the EULA, or a refund because they don't agree with limitations that were not evident when they purchased. If the vendor says "no refund", they are Microsoft's agent in giving the user their default rights -- something Microsoft probably doesn't want to happen.

Google launching its own navy?

Donald Becker

IDS? I thought that vapor has dissipated

They announced long ago, and were to have a fully operational ship by the ballpark in San Francisco in April. Even at the time the scheme sounded... not well grounded. The only way it would make is if there was a tax or zoning dodge involved.

This concept is similarly ill-conceived -- a mishmash of concepts that sound good but are not practical in real life. But if anything similar happens they can take credit for having though of it "first".

For example, generating power from wave action. There is a lot of power in waves, but it's very diffuse. Too diffuse to effectively harvest. Even if it's concentrated to the surface by a gently sloping bottom, there are only a few sites where it's worth (marginally) building energy conversion facilities.

And using water-based heat exchangers? It's done all over, and you don't have to be aboard a ship to do it.

How to destroy 60 hard drives an hour

Donald Becker

Work safety prevents most of these methods

Most of the suggested approaches aren't usable in an office environment. Pretty much anything involving high heat or strong chemicals can't be used. Either OSHA or insurance policy rules prevent most things that we consider fun.

Even a drill press is actively dangerous, as is swinging a hammer.

Transporting drives somewhere to destroy them increases the risk of loss and cost.

The safest, quickest approach is a hand-operated hydraulic press. For under $200 you can buy one that will take care of even ancient 14" drives. A hand-pumped one avoids most safety issues, such as guards and safety switches, and you can still crush 2 or 3 drives a minute. Even a sledge takes more time to line up the drives and clean up after.

We use a 20 ton press, which far stronger than required. Anything above 4 tons should be strong enough.

SCO bags $100m to fight another day

Donald Becker

Yes! The writer's strike is over!

For years we in the tech industry have been following this as a real story.

But we should come clean and admit to ourselves that it's not.

The evidence: very little has happened since the writer's strike has a happened. Groklaw has been re-hashing old story lines. We have many more guest interviews.

But the day after the writer's strike ends, our soap opera suddenly sprouts a new twist. Veronica, who we learned died in a plane crash, is back after sudden recovery from amnesia triggered by a blood transfusion from a eskimo. But it's complicated because Brad has remarried, but is having an affair with Marsha, who is secretly plotting to...

Oh, wait, wrong channel. But the day after the writer's strike ends, SCOX rises from its deathbed, stronger than ever. The doctor pronounces them fit as a fiddle, with "up to 80 more years to live".

HP to reduce PC energy consumption by a quarter

Donald Becker

You can reverse the computation as you as you don't look at the result

So-called "reversible computing" promises zero-power computations by having the calculations done as a wave and recovering all of the energy. This requires careful attention to design details. All N-input logic elements have N outputs, and all of the input energy is directed through to the outputs. Every output must be used, so that the energy can be "reflected" back and recovered.

One of the problems here is that sampling the 'result' output removes that energy from the system, making the computation irreversible and therefore consume energy. Doh!

Rove investigator erases his PCs - to kill computer virus

Donald Becker

Nuke it from orbit. It's the only way to be sure.

Very old drives (and floppys) used stepper motors to position the heads. These drives moved the heads in tiny steps and just counted the number of steps to position the head. These ended up in different locations depending on which direction the head had approached from, as well as thermal expansion and wear. To keep this from being an operational problem the tracks were very wide, and the write heads were wider than the read heads. The read head would thus end up in a cleanly-written region, while old data might still be readable a quarter track over.

Drives have gone through much evolution since that time. Today there is basically no room on the disk surface for traces of old data. Clever head positioning designs actually have to cancel out the traces of valid data on either side as part of keeping the head centered over the current track. Pretty much like putting your face really close to the pavement and steering by making slight adjustments to keep both eyes seeing an equal amount of black pavement and white centerline. The clever part is that you can stay on your broken centerline even when the road is densely covered with other broken centerlines, as long as the other centerlines are mostly randomly broken.

Anyway, the whole point is that a single write on a modern drive gets rid of the old data on that track. Writing more than once is a hold-over from very old technology. If that's what you are focused on you'll probably miss some of the other vulnerabilities, such as bad sector and bad track remapping.

OK. That's the technology point of view. We still haven't gotten close to the answer of why there was a paranoid-level of erasing done due to a "virus".

Page: