Journal de Jacques

Firstly, they don’t have Stephen Conroy.

Secondly, they’re not going to introduce software patents.

NZ: 2. Australia: Sux.

Google’s announced that they were the subject of a precise and sophisticated attack, apparently aimed at getting access to the GMail accounts of pro-democracy critics of the Chinese Communist regime, both living in China and abroad.

Google don’t think that the accounts were compromised but can’t be sure.

In response Google have said that they are considering pulling out of China entirely — shutting down the self-censored Google.cn website and closing their China office.

It’s about bloody time they realised they’re dealing with gangsters and thugs.

Update: Google arch-rival Microsoft have said that the attack may have exploited a hitherto unknown flaw in Internet Explorer. They’ve been working with Google on the whole situation. It’s heartening.

Based on current feedback, I’d say paying a lawyer to talk about software patents at this point would be like setting money on fire.

– Ryan Gordon

Jon Skeet explains that human complexity is one of the causes of software complexity. Everything you might think is simple — numbers, letters and dates — is actually devilishly tricky.

A lot of progress (and sometimes, regress) in computer science and software engineering seems to come from rejecting, modifying or otherwise modifying the “natural order”. By natural order I refer to the generally accepted, industrial paradigm of how development “is done”.

A lazy student can easily find ‘revolutionary’ projects by simply fiddling with this accepted order.

For example, take the modern imperative/object-oriented paradigm for software languages. Subtract some feature, and explore the consequences. What happens when, for instance, you cannot use getters/setters/properties? I’m not sure, but it’d be interesting to know. What happens if you subtract assignment? And so on. In a way some functional languages changed thinking by subtracting changes in state.

Another thing you can try is to move something from one phase of program life to another, or to collapse phases. Lisp, for instance, allows the programmer to have instructions run at compile-time (ie, macros), rather than simply at runtime. What else can you move out of its natural phase? Take CPU scheduling. Currently this always happens during the runtime phase, but could it be moved? Can there be load time scheduling? Compile time scheduling?

I’m sure that you can think of many other such examples.

Seen on Slashdot’s front page today:

Seems that for a crime to count, it first has to hit the high and mighty.

The Electoral Reform Greenpaper has mostly received coverage for the two particularly stupid proposals that are raised: lowering the age of franchise and replacing the paper ballot with electronic or — much worse — internet voting.

Robert Merkel points out at LP that there is an almost universal condemnation of the latter idea by IT security professionals. Probably because creating a system that satisfies security, integrity and the secrecy of the ballot is impossible.

By “impossible” we don’t mean “expensive” or “highly impractical”. We mean impossible. It simply cannot be done. The requirements are mutually exclusive.

Of course the green paper covers much more than wishful thinking from non-security professionals, including voting systems, legal arrangements, AEC structure, enrolment arrangements and so forth. They’re asking for submissions up until 27 November. Here’s mine (PDF).

Some of you might know that I’ve harboured a small enthusiasm for aggressively using IT to drive down costs of health care, as well as improving quality, safety and providing data for researchers. In the future it may even be possible to couple these databases to expert systems that provide secondary diagnoses to help doctors do their work and keep an eye out for the bad apples.

In the USA US$20 billion dollars of Obama’s stimulus package has been allocated to encouraging practitioners and hospitals to introduce electronic medical record systems. Each recipient will source and install their own systems.

Much as it pains me to say it, this is a job that cries out for consolidation. Medical records work best when they are universal and portable. The approach taken in the US stimulus bill won’t achieve that. And even if it did aim at a central records service, the history of large IT projects is one of nearly universal failure.

One software package that might be a candidate for such a system is the USA Veteran Health Administration’s system VistA (not to be confused with Windows). The VHA has the lowest cost and the highest safety of any part of the US health system; the stable, mature, universal and user-friendly nature of the VistA system seems to have had a very large role in providing these outcomes. However it looks as though VistA will not even get a look-in as part of the US$20 billion allocated for healthcare IT reform.

It is very common in IT to see “Service Level Agreements” specifying a certain amount of uptime. This is usually considered in “nines”: when someone talks about five nines, they’re referring to 99.999% uptime.

Very few services actually attain it, or even come close. All it takes is one bad day and the “downtime budget” for the next century is cooked.

But what’s the real problem? Is it the mean time between failures? Or the total amount time offline? Or is it how long business is disrupted for? I believe the latter.

So here’s my thought bubble: SLAs could be specified as Maximum Expected Downtime. We could even do it with nines, if we liked, with some help from the actuaries. “30 seconds, five nines” would mean that in 99.999% of downtime events, the system is available again in less than 30 seconds.

For those who are worried about interruptions to service and who want to keep the old measure: does it really work for you? And how seriously do you need that old-style uptime SLA? If you really mean it, consider buying a NonStop or z-Series.

A cat named Hartigan has apparently put himself amongst the blogging pigeons. A generous amount of fur and feathers has flown as a result.

For example, Hartigan has defended traditional media reporting and newsroom methods; bloggers say that News Ltd don’t “get it”, or are already giving in, or their content sucks, or some combination of these.

As I pointed out about a month ago, what the content-producers for News Ltd and bloggers do is a total sideshow. This purely tribal confrontation between hacks, flacks and new jacks is just that: tribal. The internet is strangling News Ltd’s money supply, which is what counts.

It’s a mistake so common that even wise economists are missing the money.

News Ltd — ignoring the movies, music and games parts — is not a media company. They manufacture and sell advertising and classifieds inventory. That’s their core, actual business. That inventory is manufactured by journalists churning out the stuff which fills the gaps between the inventory and convinces the general public to pick it up. The problem for News Ltd isn’t that bloggers are somehow magically better at doing that, it’s that due to the internet there are no more expensive barriers to entering the advertising inventory market.

A similar mistake people make is thinking that Google is a technology company. They are not. Google’s real business is manufacturing and selling advertising inventory. Sound familiar? It should, because that’s the business News Ltd are in. On the surface these companies are chalk & cheese: one is a traditional media conglomerate with roots going back generations; the other is an upstart firm that sprung to world prominence with a vastly superior search engine offering.

But the thing that makes them money — the thing that people actually pay them to do — is to manufacture and sell advertising inventory.

Companies are usually categorised by what they market or what they spend R&D money on. This is just silly: just as economists care more about people actually do rather than what they say they will do, we should categorise companies by what people pay them for, not what they started off doing or what turns up in their marketing material.

If the term’s not taken, we could call this Revealed Industry, analogously to the concept of Revealed Preference. The revealed industry of both News Ltd and Google is selling advertising inventory. All else is, I regret to say, mere fluff. And right now Google is cleaning News Ltd’s clock. The horse has bolted from the barn, and all News Ltd and its critics seem to care about is arguing about the colour of the barn walls.

Disclosure: I was a classifieds department employee at a small News Ltd newspaper for 3 years.