Wednesday, May 24, 2006

More net neutrality

This time its Kevin Drum

Please, most people agree that net neutrality is a good thing, but that does not equate to beleiving that there should be government regulation to require neutrality.

Sir Tim Berners-Lee's remarks on net neutrality were widely reported but the interpretations in the blogosphere are rather different from what I heard. He was asked about net neutrality again at the CIO dinner last night. The point Tim was making was about outcome, not the means of achieving that outcome.

Mashups are Semantic Web

People looking at Mashups usually end up talking about Web Services that are SOAP or REST based.

If we could create semantic web tools for building mashups that might become the killer app for Semantic Web.

Sunday, May 21, 2006

Internet Neutrality

Net neutrality again.

I am not opposed to the idea of government regulation to protect the net. But the point of government regulation is to correct injustice resulting from unequal distribution of power. While the local telephone operators have a certain degree of unjust leverage as a result of their local monopolies at the 'last mile', the balance of power remains with the Internet users and service companies.

Google is understandably concerned that Verizon might attempt to extort 'settlements' from them. As Ronnie Kray might say "nice little search engine you have got there, be shame if it got slow wouldn't it?"

In practice the largest Web companies have the best defense, if connections to Google search become slow customers are going to blame Verizon rather than Google. It is very unlikely that Verizon does not understand that the time consumers would accept a walled garden are long gone.

Verizon customers have already paid for their 1mb/s plus. It makes no sense for Google to pay them a second time. Attempts to force a third party to pay for what has already been paid for will inevitably set off a rash of anti-trust and class actions.

The real question is whether there should be an option for a third party to pay for a temporary performance boost above and beyond what the customer paid for. If the customer has paid for a 1 mb/s connection and a video on demand provider wants to provide a temporary boost to 8mb/s should this be allowed?

This question is rather harder to answer. One possibility of accepting this type of scheme is that we end up with a situation where control of the last mile becomes control over all value added content. Instead of Internet distribution trumping the local cable monopoly it would be a change of owner.

I don't think that happens though, provided that the content providers can pass on the upgrade charge to the consumer.

If I want to watch the latest episode of Dr Who via the Internet I might well be prepared to pay a $0.50 surcharge for a high bandwidth link if thats the type of thing I do occasionally. If I am using the higher bandwidth so often that the upgrade charge is significant I probably want to pay my local broadband provider for a faster connection.

What is not going to be acceptable or sustainable is a situation where the content distributor is prevented from passing the additional cost onto me as a consumer. Or rather as the cost is going to be passed onto me in any case, the content provider must not be prevented from offering me a discount when I bring my own bandwidth.

The tin foil hat brigade are almost certainly right in thinking that there are commercial interests plotting to do evil things. Where I disagree is I don't think that Dr Evil has the capability to make good on his plans. I suspect that there will be some unpleasant business where broadband providers attempt to block competing VOIP services but the unpleasantness will inevitably be settled in favor of the consumer and the VOIP providers.

Instead what I believe will happen is that everyone will win something, but nobody will win everything. The content distributors and broadband providers have a common interest in building out new high bandwidth infrastructure to support Video On Demand over IP (VODIP). Someone has to pay for that infrastructure and it is going to be difficult to persuade consumers to pay for higher bandwidth until they understand why they might need it. The temporary upgrade model allows gradual phased deployment of high bandwidth links, I only need to pay for the high bandwidth when I need it.

Eventually the basic Internet connection will be fast enough to support VODIP in high definition. Upgrade charges will only be required for people who need the next bandwidth upgrade quanta to support the next bleeding edge bandwidth hog applications.

I cannot get excited about the net neutrality issue in Congress because I don't think it has any difference on the outcome. The most a net neutrality bill would achieve is to create a profitable market to circumvent it.

Blogger is grumpy

Blogger is grumpy today, just had to delete a post that had half of a different person's post attached half way through.

Just as well that it wasn't from a blog on hardcore porn or folk dancing.

Scalia on foreign law

The Washington Post reports Scalia tells Congress to mind its own business.

The US is a common law country. It is undisputed that the basis of US law is English common law. The nationalist breast beating that takes place on this question is quite ridiculous.

The case in which the citation of common law is controversial is of course the Supreme court interpretation of the constitutional ban on 'cruel and unusual punishment'. It is pretty difficult to understand how the term 'unusual' could be interpreted except by reference to other jurisdictions.

What seems to have escaped notice however is Scalia's comments on the weight to be given to the legislative history of a bill which is very little. There is good reason to treat legislative history with skepticism. It is frequently contradictory and it is impossible to know quite what the legislators intended when they voted on a bill. Allowing legislators remarks to trump the stated intention of the act means that a single legislator can in effect negate the entire purpose of a bill by reading a tendentious interpretation of the measure into the record. Instead of having to fillibuster the civil rights bill Strom Thurmond could have simply stated that none of the measures in the bill were intended to actually inconvenience the predjudices of racist thugs like himself.

Which brings us neatly to the issue of the 750 Presidential signing statements which are alleged to change the meaning of virtually every Act of Congress during the Bush administration. If legislative intent is to be interpreted skeptically then a presidential signing statement that is not subject to rebuttal or correction must be interpreted with even greater skepticism.

Legislative intent and foreign authority are both tools to be avoided but nevertheless have occasional justification. One of the main justifications is in the field of technology.

Very few laws are so baddly drafted that they are ambiguous when originally drafted. Ambiguity is usually the result of social or technological change. Execution of minors was certainly not considered cruel or unusual when the constitution was drafted but the framers certainly did intend to prohibit hanging drawing and quartering, the traditional English penalty for treason. It is arguable that the framers also intended the phrase to be interpreted in the context of contemporary ethical standards.

Technological change is often the cause of thorny legal problems. If Congress enacts a bill requiring telephone providers to support lawful intercepts they probably intend the provisions to apply to voice communications established through a telephone number regardless of how the communication itself is actually achieved. When Congress passed CALEA the legislators clearly did not understand or anticipate the implications of VOIP.

I don't think it makes sense to read CALEA in a way that creates an exception for VOIP based telephone operators such as Vonage or Skype. If the courts rule otherwise Congress is certain to react by passing an even stricter set of requirements than CALEA.

Tuesday, May 16, 2006

Google to hire five interns

The big news of the day on Slashdot is that Google is hiring five interns.

This is obviously the most important and suprising technology news story we are going to see all week.

Saturday, May 13, 2006

New Fears of Security Risks in Electronic Voting Systems - New York Times

I have a policy of not commenting on specific security vulnerabilities in products of customers or competitors. Fortunately my employer does not do voting systems so I can dump on the the Diebold guys who have screwed up again.

The real problem here is not malice, its culture. I do not believe that it is likely the Republican party is fixing the vote by tampering with the voting machines. You fix the vote by stopping the voters of the other side making it to the polls at all, most notoriously the old Southern Democrats whose 'litteracy tests' were really voter suppression. Today corrupt Secretaries of state such as Katherine Harris in Florida or Ken Blackwell in Ohio use subtler tricks to do the same thing. Harris hired a political aly to purge the voter lists of alleged convicts that just happened to disenfranchise about a hundred thousand black people who had never been convicted of anything. Ken Blackwell just happened to allocate voting machines in such a way that students were waiting to vote at 3am in the morning.

The problem for Diebold is that they are used to building ATM machines where the security risks are very clear and failures are easy to detect. If the money isn't there when it should be there is a problem. Voting machines are a much more complex problem because you have conflicting confidentiality and audit issues.

Sunday, May 07, 2006

Another botch job? DisplayPort vs UDI

Give an engineer a problem to solve and they will solve it. Give a group of engineers a problem to solve and they will define the problem so narrowly that they end up solving the wrong problem.

It is bad enough that we have two competing next generation monitor interface standards, DisplayPort and UDI. Unfortunately neither solves the user's real problem which is not how to connect their display to their computer its how to connect the stuff on their desktop to their computer.

At least we can now use one cable to connect audio and video at the same time. Most monitors have built in speakers. Unfortunately the UDI spec only supports this by falling back to HDMI when a TV is detected. DisplayPort has native support for audio and video.

What both specs lack is the ability to send keyboard, mouse, USB or power. A wireless keyboard/mouse is not a substitute, the signal does not carry far enough.

So this spec is not going to provide a laptop doc replacement. Which is probably part of the point. If the connection is made too good it might threaten the bogus market for docking stations, power adaptors and all the rest of the overpriced proprietary peripherals that are effectively mandatory for laptop users.

Both specs re-use the DVI/HDMI connector so expect plenty of confusion if both make it to market. The chance that this is going to be a net benefit is not at all good.

Geeks want bigger displays. Consumers want to get rid of cable clutter. So guess who wins?

Don't expect the standards groups to solve the consumer's problem soon. In the meantime the best we can hope for is that the manufacturers of overpriced cables might partly solve the problem by selling DVI/USB combination cables. If we are really lucky someone might manage to get a standard layout for the ports at the display or computer end.

Will the PC makers botch Origami?

Will the PC makers botch the mid size form factor PC? Signs point to yes, and not just the Windows version.

The Pepper Pad is a Linux version of Microsoft's Origami. It is being touted as being 'first to market', we will see whether this is actually true if and when devices ship.

The Pepper Pad has one feature that I suspect the PC versions will quickly copy, a pair of built in thumb pads at the left and right of the screen. I suspect that the Pepper Pad has this for the same reason that Sharp's Zaurus did: good open source handwriting recognition is really hard to find. I suspect that this feature will quickly make it to the Windows machines as well, the dirty little secret of pen computing is that good proprietary handwriting recognition is also hard to find. RIM's thumbpad quickly blew away Palm's 'Graphiti'.

The problem is not just the handwriting recognition algorithms, its the whole interation pattern. Conventional Web sites have HTML forms that are optimized for keyboard input. Most people's handwriting is rather bigger than the font sizes used in form input.

Unfortunately the Pepper Pad people seem to have blown the design in the same way I predict some of the Windows versions will: No video out. The killer application for a device of this type has to be to carry into a room to give presentations with a DLP projector. Composite video does not cut it, I can't do powerpoint from composite video.

The way a product becomes a breakthrough product is that it supports a killer application for a niche group of early adopters, an application that is compelling enough to be able to build out the infrastructure necessary to support it. The biggest complaint against the mid size format is that the machines are twice the promised price ($1000 after necessary accessories instead of $500). The only way to make the product cheap is high volume.

I suspect that the reason the video port has been cut on most of the machines is an attempt to make the product cheaper. Its a stupid strategy, there was never any way that the first units were going to be cheap. So what we have is a machine that has a laptop price that isn't a laptop replacement.

I think that laptop replacement is what most people would use a midi format machine for. Without the thumbpad they are at best a replacement for passive Web browsing. You can read Web sites and read email but you can't reply or post easily. That is fine for many applications but fine does not justify a $1000 gizmo. The gadget may be pitched at the first time Web user but these are the most conservative computer users.

The person I would be trying to sell such a gadget to would be someone who travels quite a bit, has a desktop or laptop which is their 'daily drive' but has the occasional need for something smaller, lighter, for out of the office use.

Origami and Pepper are fine for the in-transit portion of the trip but they don't have an answer for what you do when you arrive. I don't know many travelling professionals who can drop $1000 on a gadget that won't be doing powerpoint on arrival. The only exception would be photographers using the device as a portable media vault. But that's the exception proving the rule, photographers may not do powerpoint but the need for a bigger screen is even more urgent. If I was using a midi in that way I would probably want the option of carrying a larger display to use in the hotel.

For the midi format to be viable it needs infrastructure. At a minimum much better mirroring software to allow people to easily keep their systems in sync. In the longer term a 21" or bigger higher definition flat screen monitors are going to break the $250 barrier. 19" has already done this. At that price offering guests the use of a large monitor in their hotel room starts to make a lot of sense.

Thursday, May 04, 2006

Why Size Matters

C-Net has an article on Digital SRL Lens sizes.

There seems to be a lot of confusion about the switch from 35mm to Digital. The sensors used in high end Digital SLRs are much larger than the ones used in consumer models but they are all quite a bit smaller than a 35mm film negative.

As a result the field of view of a Digital SLR is somewhat narower than the field of view of a 35mm camera with the same lens. Put a 50mm lens on a Nikon DSLR and the field of view will be what you see with a 75mm lens on a film camera.

With film cameras the manufacturers all agreed to use a common sensor (i.e. film) size and so a 50mm lens on a Nikon gives the same field of view as a Cannon, a Minolta or a Pentax. Digital allows each manufacturer to make their own choice. For reasons best known to itself Cannon has made two different choices. Their professional line of cameras has a 1.3 conversion factor, their consumer DSLR has a 1.6 factor.

What this means is if you do portrait photography, sports, birdwatching or anything else that requires a long focal length a DSLR gives you a 50% bonus in focal length. If you are taking landscapes and want the widest possible field of view you have to get a lens that is 33% wider than you would need for film.

Contrary to a claim I have seen in a number of other reviews the focal length extension factor does not affect the perspective or the depth of field. All you are doing is cropping the negative. That does not change the depth of field and certainly not the perspective.

Although it is true that your existing wide angle lenses are no longer quite as wide there is a compensation. When you buy a new wide angle lens that is designed for the new digital format (Nikon call them DX series lenses) you will find that a 10mm lens for the digital format is considerably cheaper than an equivalent quality 10mm lens for the 35mm format. Nikon sell a 16mm lens for $550 that works with film or digital and a 10.5mm lens (i.e. 16mm equivalent) for $550 as well. What is going on?

The answer is quite simple. The new digital lenses don't have to cover such a wide field of view. The flaws in wide angle lenses generally show up at the edge of the picture. The focus is usually a little softer, on cheaper lenses there is often noticable distortion. Designing a lens that produces a flat picture right out to the edge of the frame is very hard. The new digital formats mean that the lens designers only need to get the picture right in the middle of the frame.

The result of all this is that when you switch to digital all your existing lenses become 50% longer. You now have a super-telephoto if you didn't have one earlier. If you did you now have a super-super telephoto. While you may decide to replace all your lenses over time the only lens you are likely to find you need in the short term is a new wide angle to replace your current widest lens.

Tuesday, May 02, 2006

Bad security is worse than no security

One of the big arguments between US and UK intelligence is over the value of polygraph tests. In the US their use is ubiquitous, in the UK they are considered to be pseudoscientific quackery.

Psychology has long held the dubious distinction of being the scientific field most susceptible to psuedoscience. The Rorschach ink blot test was used for decades despite being a transparent piece of hokus pokus. Many of the crazier notions of Freud and Jung are still accepted by self described 'psychoanalysts'.

So when I can across the site which is run by a polygraph skeptic I started to think about the problem of testing the effectiveness of polygraphs in blind trials. If polygraph testing is scientific their effectiveness should be empirically verifiable. The National Academy of Sciences has found no evidence to support the claims made for polygraph test. The theoretical basis for the claims is weak, empirical studies do not support the high accuracy claims made.

All of which should be a serious cause for concern as the security services increasingly rely on polygraph tests despite ample evidence that they don't work. Aldrich Ames passed two lie detector tests while he was spying for the Soviets. Instead of throwing the junk out the CIA stepped up their polygraph testing.

Monday, May 01, 2006

Paul from the Prize claim center

I just got called by a guy purporting to be Paul from the Prize claim center. I get about two calls a week from him, all to my business cell phone. This is the first time I have had a call while sitting down at a computer. So I did a Google on the telephone number. Apparently it has been used by a timeshare company with a certain reputation. Other people report the same scam.

"Hi, this is Paul from the prize claim center. We have great news for you regarding the contest entry form you submitted to win a brand-new automobile. We pulled your ticket; you guys have actually won one of our top four major prizes, and it would be in your best interest to give us a call back as soon as possible. We can be reached at 877-256-7894. Do not hesitate to give us a call back, folks. We have great news for you."

It does not appear to be a dialing fraud scheme, it is an 877 toll free number, the dodgy SS7 setups that used to be used to forward to premium rate numbers have been shut down.

I have not entered a lottery recently and if I had a company contacting me to tell me I had won a car would not be using a pre-recorded message- how likely is it that they are genuinely giving away so many cars that they would do that? If I had genuinely won a prize there would be a live person on the call, they would begin by greating me by name, they would tell me the competition I entered and where.

Nielsen on Mastery vs Mystery

In his latest alertbox Nielsen links to an earlier article where he has a go at the flash style of user interface that tries to turn the Web into TV.

At this point the Web vs TV battle seems to be definitively tipping towards the Web. Flash infested web sites are much less common than they were even a couple of years ago and where Flash is used it is generally used for things it is actually useful for: mostly games.

People still use javascript hyperlinks in places where plain old hyperlinks would work better. I hate this because it breaks the 'open page in new window' option.

The point Nielsen does not make, possibly because he did not spend so much time interacting with the appostles of Interactive TV as I did is that in the early days of the Web many people saw the Web as 'Interactive TV' done wrong, they saw their mission as correcting that mistake.