That's misrepresentative. The modern trend is for countries to get smaller, and for countries to form closer alliances via supra-national entities. Czechoslovakia split in two; Yugoslavia fragmented bit by bit; India was partitioned into India and Pakistan, then East Pakistan became independent as Bangladesh.

Iceland gained its independence from Denmark in the 40s, and Greenland and the Faroe Islands have gradually increased in autonomy, on a slow move that seems likely to end in full independence. It must be remembered also that Norway was a part of Denmark until 1814.

The Renaissance model of large countries was based on principles of colonialism, imperialism and restricted trade. Scotland entered the Union because the imperial powers had set up trade barriers and Scotland was effectively frozen out of all international commerce. These conditions no longer apply, so we should instead look at state size in more practical terms. As it turns out, the governance of large countries seems to be far less effective than the governance of large countries. If you look at the world's countries in terms of GDP per capita, you'll see a heck of a lot of small countries at the top. The USA is an exception, but then it (like Germany) is a federal system and its states have a high degree of autonomy. Notably, as the federal government has taken more and more power away from the state legislatures, the US has decreased in wealth relative to the rest of the world....

Fraoch (note spelling) Scottish. Please do not confuse "England" with "the UK".

The BBC weathermap is not a true planar projection, though. It's a 3D projection rendered onto 2D, emulating the view of the UK from orbit. Neither ground-level angles, nor latitudinal distance, nor longitudinal distance, nor surface area are preserved. All distances and all areas are reduced. This is because the projection is seemingly taken from somewhere above France, so the south of England is close to the camera, and the north of Scotland is not only further away (hence smaller) but also reduced in height due to the curvature of the Earth included in the projection.

When this first came in (years ago now), the justification of this was that it is more "natural" to look at, and easier to understand... but only a handful of people have ever had the opportunity to see the UK from such an angle, so I can't see what's so natural about it. Furthermore, the BBC initially refused to allow any regional opt-out from the standard projection, so the Scottish weather was on a zoomed subsection of the map, which had practically zero north/south resolution compared to exaggerated east/west.

After a lot of complaints, the BBC tweaked the angles slightly, but the problem still remains. It is particularly irritating that the Gaelic weather forecasts, half of whose target audience are in the Highlands and Islands, is forced to use the same map, where their part of Scotland is so drastically shrunken that a single weather symbol blocks out over a hundred miles on the map.

The English are free to have their own referendum on the matter. The United Kingdom is a union of equals, and any party to a treaty can choose to leave it unilaterally.

Wrong, wrong and wrong. I am not a member of the SNP. I do not particularly like the SNP. I want Scottish independence. I do not want Scotland to be "just as it was before". I want some things to remain the same and some things to change.

It's not Scotland's fault that there is no devolution settlement for England. Personally, I have always thought it ridiculous that devolution is so asymmetrical, with different devolved powers in each of Scotland, Wales and Northern Ireland, and none whatsoever for England. But that isn't Scotland's fault.

Yes, but the problem is that imperative coding abstracts the mathematics to the point where the programmer isn't thinking mathematically any more.

Ok, but most coders won't use such features, because they don't understand them.

Which calls into question the whole point of imperative programming. Why should I spend so much time telling the computer exactly what to do in what order if it's only going to attempt to extract the task logic and refactor everything anyway? Surely it's more efficient to just describe the problem and let the computer fix it?

Not "more correct", just "more computation". I'm not a functional programming hardliner, but I do see where they're coming from on stateless code. Why should determinism only apply at the system level? Do we not conceptualise procedures and functions as programs within programs? Determinism helps guarantee correctness.

We learned Standard ML in university, and my main complaint about it wasn't the weirdness of it or the impurity of the functional model, it was the fact that we were asked to do the complete task within the language. I never understood why they expected a full interactive program in FP, when instead they could have had a lightweight application layer in non-functional code that called stateless functions as required in order to manipulate and update its stateful data. Since the early days of computing, the ideal of programming was to be able to just tie prewritten blocks together to perform arbitrary functions, and every generation someone touts this idea as the next big thing, coming this year, but it never takes off. One of the reasons libraries can be such a pain to work with is because the internal stateful processing is unpredictable. Ban state, and the problem goes away, no?

Welcome to democracy.

Are you suggesting the best measure of a language is to see how easy it is to reimplement something that has been done a gazillion times before, rather than how easy it makes it to do something more complicated built on top of all the solved problems and preinvented wheels...?

The issue is more than merely about number of lines, it's about modes of thought. In most languages, coding is an exercise that basically involves a lot of mundane data shuffling that is easily expressed, but each individual line achieves nothing of consequence. When it comes to code maintenance, reading the lines is easy, but reconstructing the program logic is very difficult indeed. Wolfram's goal appears to be to strip away all the housekeeping in order to spotlight the code logic, which strikes me as the right way around.

Your use of the term "mathematician" is interesting too, because it does suggest that you don't feel mathematicians should be interfering in programming. Well, I'm of the opposite opinion. Imperative programming is the result of non enough mathematics in computer programming. It took me a long time to get my head round my maths lecturers at university telling me a matrix was "a computer", but they were right. Imperative programming is inefficient because it doesn't allow us to take f(x) and g(x) and create g(f(x)) for all x, so we have to evaluate every step. But a matrix encodes a linear transform, and multiplying two transformation matrices gives you exactly that -- g(f(x)) for all vectors x. Typically, using matrices for a single step is less efficient than just running the normal code, but when you're going to carry out the same calculation j(i(h(g(f(x))))) on hundreds or thousands of xs, calculating a single matrix and using that every time is the most efficient way.

This thinking doesn't start and end at matrices though.. most computation can be conceptualised as a data transform.

Sorry that's life, the_universe AND everything. It's a bitwise AND, of course, so that's 101010. Life is 111011, everything is (as you might expect) 111111. Finding all possible values of the universe is an exercise left to the reader.

Does the trade agreement state that every product must always be packaged in all three languages? I think you'll find that it doesn't, which means it's the producer's choice to package that way. It is therefore presumably in their commercial interest to do so. If it was cheaper to have 3 different packages, they would have three different packages.

Seriously dude? Is that what this debate has descended to? "I know you are, but what am I?"?

The rental model is supposed to increase the benefit for all parties. If the rental model is not increasing the profits of the producers, it's a bad model. The consumer gets a broader choice and easier access, and typically pays extra for the privilege in the long term, which is how the producer profits. Spotify's model relies on less, less and less. The consumer pays less than elsewhere. The producer gets less than elsewhere. Spotify takes less commission than a retail outlet. The problem with this business model is that the producers' costs haven't decreased.

Secondly, Spotify have two very different products, as you yourself pointed out: music on demand and automatically generated playlists. These two products are of different values, yet Spotify pay the same for them. In essence, they're using the non-paying customers to keep the buy price down for the guys who bring in the cash.

Now imagine if there was a different royalty for on-demand vs algorithmically selected -- what difference does that make?

Well, for now, 100 plays at 0.6 cents gets 60 cents, total. If we double that rate for the premium users and half it for the free users, that's 52.5 cents generated -- a reduction. Double/half again, premium plays give a royalty of 2.4 cents and it's now generating 71.25 cents -- a profit for the producers. Even then, that's still an effective price of 3.4 cents per track to the buyer, which is a very low rental payment. If you double/half again, the royalty becomes 4.8 cents, and the effective price is just under 6.9 cents, a tenth of a low-cost iTunes song. Now you're starting to get into reasonable rental territory, and those same 100 plays, 25 from a premium user and 75 from a free user, generate $1.25, more than double the initial take for the music producers.

Of course, that would mean doubling the subscription price.

So, back to the problem of Spotify's business model... Spotify wants to be "freemium", but the freemium model relies on you owning what you are selling, because you can't give away something that you don't own. They've twisted and contorted the business model multiple times to try to preserve something of the original idea, but it's left them with two different products that they're trying to call one thing, and the lack of profit in the freebie product is interfering with the pricing and profitability of the premium product.

Ah, the old "there aren't as many of you, so you're less important" argument. When this is about skin colour, religion or sexual orientation, it's mostly accepted as being bigotry; but when it's about language, it's fair game, right? Cos "language is a choice". Yeah. The first thing I did upon being born was select which language my mother was to speak to me, and communicate that to her via telepathic link. By a lucky coincidence, I managed to pick the one and only language she was capable of talking to me in.

And as for manuals, mostly they just put a "quickstart guide" in the box, and a CD-RoM with the main manuals in PDF format. No-one reads the manuals anyway. Everyone's a winner.

When I say "versions of English" I mean "versions of Modern English". Scots is a language that diverged from English at the very beginning of the Middle period, if not before. "English" as we know it didn't exist then.

Sorry, I missed a vitally important word there: "areas of majority white native-speaking descent". My point is that in areas where colonists were in the majority, English was preserved more or less intact, and the imported slaves, the free immigrants and the locals learned from their pattern. The children of the non-natives then still had exposure to the native pattern, and spoke better than their parents. In such a situation, the non-native errors disappear within a couple of generations (you can see this in any immigrant community in any country and any language).

I was using white as a shorthand, because when you go to India or Oceania, you can see immediately that the population isn't descended from the colonists by the colour of their skin, and their English is markedly different, because it has turned what began as non-native errors into a part of their language variety.

There are even areas in the US where true "communities" persist based on colour, and there -- guess what? Strong dialectal differences persist, and black speech even has its own name: ebonics. These modes of speech are in no way inferior to standard English, but sadly society as a whole judges them to be so.

So no, I'm not saying that only white English speakers contribute to English-language literature. I am not racist (although my poor wording may have suggested otherwise). I was simply pointing to the fact that the ratio of native speakers to non-natives is what determines whether the language changes or remains the same.

"Dozens of different 'r's"... but they are all still Rs. If I write R, I don't care which R you use, I happily recognise it as an R. That's what a phoneme is -- it's a meaningful unit of sound that can be pronounced in multiple different ways. When the R reduces to the point that it turns the vowel before it into an "eugh" sound, it's still an R. I could write it as R, as , as , as &*, as anything -- it would still be one phoneme.

The mismatch between the written form and spoken is pretty much identical everywhere. The OU in "you" preserves an old form, but it's different from almost every other OU -- out, shout, etc. Rewriting "you" as "yoo" would not favour any dialectal variation over any other.

I did not specify written Classical Latin. I was talking about Vulgar Latin. You assume that the differences in Vulgar Latin had "nothing to do with errors or failure to learn correctly", but that's a very unlikely situation. The Roman Empire grew at a phenomenal rate and conquered peoples with radically different languages. Then they threw those people together in legions and slave gangs, without taking the time to send them all to university to obtain a degree in the language. The Roman Empire at its height had a population in the hundreds of millions, but the number of actual Romans would never have been more than one million. Most of the subjects of Rome would never have had day-to-day contact with even one Roman citizen, so there was no convergent force to bring them back to the native model. The spoken vernacular can't have been "Latin" at all -- it would have been a series of Latin-based creoles.

Languages do not have rights -- correct. People do -- correct. It was not the language that asked for itself to be defended, it was the people that asked for it to be defended.