... because the article gives no evidence or reasoning to explain how Watson could scale up to become a general search engine. Watson as seen on Jeopardy used Jeopardy-specific analysis rules and an info base of a few hundred gigabytes. It's just not the same thing as a general-purpose internet search engine.

Here's a big hint for Tim: on iOS, you can't write a custom keyboard. On Android you can. This is a really big deal in Hong Kong, because iOS has no support for Cantonese-based Chinese input. The best you can do is a kludgy app where you have to copy and paste the result (see https://itunes.apple.com/us/app/canton-guang-dong-pin-yin/id385519764?mt=8).

Therefore, the Cantonese user is hamstrung by Apple's lack of support for the Cantonese-speaking market, together with their locked-down approach which prevents third party developers from filling the hole.

Compare this with the situation on Android, where there are at least five Cantonese-based keyboard input methods, together with Cantonese voice recognition. Why is it surprising if Hong Kongers find iOS seriously deficient?

Sorry, I tried to hold it in but it just burst out

This seems to be an interesting example of a court case being fought for the publicity it generates. It's surely cheaper to file this suit than to advertise his film in conventional ways. However unlike cases such as SCO v. IBM, the litigant probably believes he would win the case if it came to trial. The newsworthiness of the suit lies in the audacity of the defendants in aggressively asserting the copyright in the first place.

The essence of Tor is that your message passes through multiple nodes (say 3), none of which knows your message's origin and destination (and indeed content). But this breaks down if all the nodes are controlled by the same sysadmin.

Surely if we end up with a high proportion of nodes on Amazon, then some communications will be routed entirely between Amazon nodes. Then this breaks the anonymity model, allowing the secret policeman to log (or subpoena) the user's traffic.

On a not unrelated note, what's the general view of the current state of Perl 6? I can look at http://planet6.perl.org/ for the view of those close to the project, but what's the word on the street? I think "word on the street" is a really important metric as to how well a project is doing. Trends are a major determiner of which product potential new users will find. Rather like bank runs: it can be irrational to trigger one but nevertheless rational to follow one.

"You're an idiot. You signed something under threat of prison / arrest without bothering to consult a lawyer. No amount of mention of poverty, trust, or even just plain intimidation should have made you do such a thing without first consulting a lawyer." Hmmm, threatening to go to the police if someone doesn't sign a contract. I'm pretty sure that would constitute blackmail in the UK under the Theft Act (making aan unwarranted demand with menaces), which is a serious offence with a maximum sentence of 14 years imprisonment; see http://www.legislation.gov.uk/ukpga/1968/60/section/21. Anyone know what the law on blackmail is in his jurisdiction?

1. Spark a hunting spree that has no hope of eradicating them.
2. Apply selective pressure to breed unhuntable pythons.
3. ???
4. Profit!

Or am I missing something?

You'll find Python very easy to learn if you're already experienced in Perl. By experienced, I mean you understand the kind of Perlish programming patterns involving lists, hashes and complex data structures, and you understand object orientation in Perl, and you have a good feel about when to code something yourself versus when to start looking for a third-party module. All these things are very similar in the two languages, and different from other popular non-scripting languages such as Java. Indeed, if you understand that a Perl object is really just a hashref "bless"-ed with a class name, then you'll have a deeper understanding than most Python programmers of Python objects (which are essentially the same thing underneath, but with more "classy" syntax when you're defining them).

One major difference is reference types: Whereas Perl has both @a = (1, 2, 3) and $a = [1, 2, 3], Python effectively only has the latter. Similarly, Python does not have something like %a = (one => 'un', two => 'deux'), only $a = {one => 'un', two => 'deux'} . Also, strings and numbers don't magically behave like each other: you need to do str(123) or int("456") or float("7.89"). Since you appear to be in the USA, differences in Unicode handling probably won't matter too much.

Don't worry about the syntactical superficialities regarding semicolons, dollar sigils, whitespace etc; if you can already program productively in some language then it won't take you long to adjust. Get a good book on Python and spend a few days working through it solidly from cover to cover, or at least until you feel you don't need to continue. That way you'll crack all those minor surface-level differences in one maximally productive chunk of time.

Finally, don't waste time worrying about whether Python or Perl (or any other language) is "better" or "worse" overall -- too many lifetimes have been wasted that way :-)

Is this just the case where members can sometimes cast a vote both for and against something, effectively cancelling their vote?

This happens all the time in the UK Parliament - the MP just walks through the Aye lobby and then the Nay lobby (or the other way round), then the two votes cancel each other. It's a long-winded way of abstaining. Given how fast MEPs have to vote, maybe they do it when they've made a voting mistake. In which case, you could certainly question the value of such rapid voting, but it wouldn't be an example of fraud.

I work on Welsh-English machine translation, and have looked at doing this in the browser with a bookmarklet so that comments in Welsh can be read by non-Welsh speakers. The trouble is that people tend to use non-standard spelling etc in informal postings such as facebook, whereas the majority of parallel text available to train a statistical machine translation model tends to be formal language (government documents, press releases from business, etc).

Possibly this could be solved with two stages of translation, i.e. (Lang1 with informal spelling -> Lang1 with standard spelling -> Lang2). If this mapping is relatively straightforward (e.g. common spelling substitutions such as 'ough' -> 'u' or 'uff') then a statistical model might work quite well, if you could first split words into syllables with some rule-based algorithm.

Just a totally trivial, obvious thought about the application of statistical machine translation (sorry if I've pissed on some patent troll's livelihood, heh)

Firefox is an open-source platform which is independent of any significant content provider. Chrome, like IE, is a project controlled by one company with a vested interest in directing users to particular content. I think we should find it concerning if Chrome is succeeding at the expense of Firefox.

Now I understand that many people really, really like Google, for important reasons such as their track record of being pro-standards and pro-freedom. But we should always support or oppose individual actions on their own grounds -- wherever possible, we should avoid depending on long-term trust of particular individuals or organisations, because there is no guarantee that we will still support their actions at some point in the future. We believe in political systems which have checks and balances. The same principle should apply here. A situation where the dominant search provider is also the dominant browser provider is one where we miss out on important checks and balances.

The situation is different from the phone market, where Android is squeezing a variety of closed platforms, thereby giving manufacturers and individuals more choice. In this case, there was already a viable and independent open platform, Firefox, and Google's offering is preventing it from becoming dominant.

These people could be paying £30 to let a salesman into his house to try and fleece them for all he can.

Yeah, I'm sure you're right -- my girlfriend got a £10 Groupon with some con artist photographers called Fusion Studios. They insisted on a "deposit" of £75 which took weeks of legal threats to get back ("the manager's in America this week and nobody else can sign cheques", etc). Some of her friends ended up paying £300 due to the high pressure sales tactics.

The moral: don't buy Groupons unless you're happy to experience con artists from time to time.

I wonder WTF their contingency plan is if a big tsunami hits now ...

I strongly believe we know how to set up technical systems for safe nuclear power. However I'm extremely sceptical of the idea that we know how to set up social / administrative systems for safe nuclear power. It's too easy to hide systemic weakness behind secrecy, or too embarrassing to identify and fix present failings, or the debate gets too polarised and ideological so people, politicians and regulatory systems lose sight of the actual safety issues because of the headline effect etc.

I wouldn't be quick to blame money or corruption or unscrupulous people, either. The key problem is secrecy -- even without malice, familiarity makes you blind to system flaws -- we software people know this very well. Only total transparency can ensure that flaws do not get hidden. On the other hand I don't know how this can be reconciled with security against sabotage.

There's a need for a sober, measured debate about all this and it's a pity that a few fundamentalists (on both sides) are making this impossible.

Florian contradicts himself in the article:

Google's own patent portfolio is [...] far too weak for what's undertaken in connection with Android.
[...]
Just today, Google was granted a typical "troll" patent ... A company that seeks to monopolize such basic ideas -- behind which there really isn't any serious technology -- apparenty loves patents

He can't have it both ways -- either he wants Google to file defensive patents or he doesn't. A patent isn't a troll, a patent aggressor is a troll. If you just use a patent defensively you're not a patent troll.

He also seeks to draw an arbitrary distinction between "trivial" and "non-trivial" patents. The problem with patents is not triviality. The problem is that they can hit you even if you thought of the idea independently, or if you're just trying to conform to a de-facto standard.