Yes, it probably is just a coincidence that Steve Jobs' illness recurs just when Apple profits are this high.
But on a more fine grained timescale a few days more or less in the timing of the announcements might make quite a difference in how people perceive this. If they had announced a month ago instead of right next to record quarterly earnings it might have hurt the stock more at the time than it does today.
The market is in theory 'efficient', but in the short term it tends to overreact to bad news, so combining bad news with extremely good news may be a viable method to offset the effect of the bad news on the stock.
Savvy investors will have priced Steve Jobs health in to their stock holdings already, those that are on edge might panic and sell unless there was a second factor.
> Most iOS devices being sold are, in fact, not iPhones!
Pedantic reply: half ≠ most, half = half
But that aside, I see your point: clearly iOS can still be influential regardless of the shift in the cellphone market space. What will be interesting to see is how (or if) that changes when Verizon gets the iPhone...
I see nothing wrong with that statement, it certainly wouldn’t confuse me or seem strange to me. I would take it to mean that there are slightly more woman in north america than man. The context (i.e. the ratio of men and women is usually very close to 50:50) would tell me that there are likely not many more women in north america than men.
> I would take it to mean that there are slightly more woman in north america than man.
That's a terrible assumption. The problem with 'most' is it doesn't imply an exact number, it could be 0.01% more or 99.9% more and either is 'most'.
Nothing about that statement tells you it's near a 50:50 ratio.
Since we're not dealing with exact numbers (obviously 17 and 16 million units is rounded off) and because the difference is fairly small ~5%, it's less confusing to note a more specified relationship rather than an ambiguous and potentially confusing 'most'. Esp if later it's quoted out of context: "Most iOS device sales are not iPhones," (how many is most? how close is the margin?) is less informative than, "IDevices excluding the iPhone comprise about 5% more sales, with iPhones selling 16m units and other devices selling 17m."
The usage of “most” was completely correct and not in any way wrong or ambiguous since the comment contained pretty exact estimates and didn’t only use “most” to describe the relationship. Context matters and it is prudent to assume that HN readers possess basic reading skills.
Take “most” for what it is and always has been, a simple description of plurality or relative majority.
(You are quite correct that I cannot just assume that the gender ratio in North America is close to 50:50 but you are at the same time also completely missing the point. Context matters and context makes “most” a useful word. It’s alright to make use of term “most” in the context of gender ratios of geographic areas because everyone knows that gender ratios are always close to 50:50. “Most” is a very broad term but if used in the right context perfectly appropriate and not confusing.)
When dealing with numbers, esp financial data, it is poor form to not make specific analysis in the descriptions of the data.
Saying 'Most iOS sales are not iPhones' is actually very misleading even if generalized sales numbers are included in the same context.
Ugh, this quarter you will receive 'most' of your bonus. Oops, I hope you didn't assume that 'most' meant 95% or even 75% because it actually meant 51%.
> Context matters and it is prudent to assume that HN readers possess basic reading skills.
There is absolutely no reason to not be more specific. You're just defending poor form at this point and it wasn't even your comment.
Lastly, it's always bad to assume. Why should we assume that this won't be quoted out of context? Assumptions lead to bad things in my experience. It's never prudent to assume.
Why are you inventing arbitrary rules for the usage of “most” that have never before existed? I don’t understand that at all. “Most” denotes a relative majority and always has. It’s correct to use “most” that way.
Why are you assuming that “most” must be at least a 75% or so majority? If we are talking about wrong and bad assumptions then that is the one.
You might argue that it’s not exact enough (I don’t agree at all because the comment includes specific numbers) but to say that using most in that context is wrong is just mind boggling to me.
> Why are you inventing arbitrary rules for the usage of “most” that have never before existed?
I'm not. The rules aren't arbitrary, they're based on the grammatical roots of the word.
Your consistent problem is your assumptions, which run rampant throughout your comments, as evidenced here.
The colloquial meaning and usage of a term will always trump the technical reality.
Commonly, 'most' is used to mean there is more than a simple majority, e.g. a supermajority.
"Most senators voted in favor of the bill," this isn't said when 'most' refers to 51 senators voting in favor of a bill because journalists understand that that would cause confusion. Instead something like, "The Senate was nearly split on the vote, 51 to 49," is used because it is far more descriptive of that actual situation.
The relative position of numbers matters, whether you want to accept that or not. However you sound like a fool promoting ambiguity of information and obfuscation of data.
I must say, I’m really not familiar with this usage of “most”. It’s the first time that I hear anybody telling me that you shouldn’t use “most” if the majority is slim.
Where does your assumption come from that colloquially, “most” means “large majority”?
Now for a quick lesson in grammar, sorry it had to come to this but besides the colloquial instinct, the grammatical roots of 'most' will show, unequivocally, 'most' should not be used as it was.
'Most' is the superlative form of 'many' or 'much'. 'Many' is defined as 'a large number of' and 'much' is defined as 'a large amount'. This naturally implies a statistically significant large majority, not a simple majority such as 51 to 49.
When there is a close division we don't say, 'most of the Senators' we say 'just more than half of the Senators'. This is because we haven't reached a point where we can adequately say that a 'a large number' (as compared to those voting against) 'has voted for'. For this reason we reserve 'most' for situations where more than a simple majority, e.g. often a supermajority, comprises one of the two or more things we are comparing.
Actually, grammatically speaking, superlative adjectives should _only_ be used when three or more things are being compared, however this is a commonly ignored rule.
There is one final point, when one of the two or more things you are comparing reaches a point where it is greater than two thirds of the total, we begin to say 'nearly all,' "Nearly all of the Senators voted to pass the finance reform bill today, voting 89-11."
Well, the difference between sold iPhones and other iOS devices clearly is “statistically significant” and a 51 to 49 majority in the Senate is also “statistically significant” (but the term does not make any sense if applied to Senate majorities). So yeah, most it is.
You're missing the point: The difference between 17 and 16 million, is not a statistically significant /majority/, please be careful when quoting, because you missed the most important piece of what I said above.
17 million as compared to 16 million is not 'a large quantity' more as compared to 16 million.
If you understand that 'most' is the superlative form of 'much' or 'many' you understand that you're saying there is a large amount (more) of iDevices being sold as compared to the number of iPhones being sold, when all things considered that number is less 10%! Hardly a supermajority and very much in the realm of "simple majority."
I brought this up on Freenode in ##English. They were absolutely appalled by your claim, "half plus one is most."
If you have further questions I highly recommend the erudite discussions they can provide. They are quite well versed in English grammar and usage, I think you will find their comments instructive.
Not the way it's used in day-to-day casual language.
If your boss told you that you would be getting "not all, but most of your bonus this year" and you got half plus one dollars (or whatever you're paid in), chances are you'd be feeling cheated, no matter how 'technically right' your boss was.
No, not in this context. "Most iOS devices are not iPhones" is an accurate statement. It's also accurate in attempting to get the point across: iOS doesn't necessarily mean iPhone, and that iPhone isn't the only source of iOS platform that should be considered.
It's like the idea of significant. All too often, people confuse significant with majority. This is not the case. Significant rarely means majority, except when used in the all too cumbersome "significant majority".
Finally, most people would agree even a 5% different is a significant difference. =)
Regardless, saying ~5% is more instructive than simply saying 'most', which might confuse people who aren't reading closely or if it becomes quoted out of context. It's poor form to miss an opportunity to be more specific about data rather than less.
“It's poor form to miss an opportunity to be more specific about data rather than less.”
I don’t agree at all. It’s usually a good idea to summarize data and in the process of doing so being less specific. Most is perfectly understandable in the context. It actually is very surprising and interesting that most iOS devices that were sold in the last quarter are not phones so saying just that is perfectly fine.
There is no strange and arbitrary 5% condition for the usage of most and there never has been. Most is an exact term.
I agree completely. Compared to any other public apology I've seen in Japan, well his simply doesn't compare. Unfortunately I think he comes off as the insensitive-to-cultural-norms American which actually does more damage in Japan than it does to smooth over problems, at least in my experiences living there.
> In many ways, Job's eventual departure (hopefully based on choice, rather than necessity) could be good for Apple...[i]t won't be exactly the Apple of today, but given some of the…hostile decisions over the last 3 years, it might actually be an improvement.
My guess is they have, just not publicly. (Easy to argue the merits of that.) Pancreatic cancer doesn't leave too many survivors so I am sure the management team thought a lot about life without steve, at least at that time.
I agree, but I think it's less of an internal logistics issue and much more of a public image issue: Steve Jobs is so much the face of Apple as it sits currently, every time there's news about his health the stock fluctuates and that isn't good.
> What if we make things easier for the machine? It is obvious to a rank beginner that a perfect game with a rook handicap is a win for the side with the material advantage. No, make it a queen! Surely that must be a provable win?
Hm, I'm sure it must be. Although I don't know how you go about proving it, it's a simple matter to force equal trades; black cannot avoid the exchange of pieces forever, and if white plays a perfect game he will always win, without doubt.
> Not so fast. Even against a crushing asymmetry in material, it is not too hard to avoid mate for a couple of dozen moves, which means that calculating all the way to the end of the game is beyond the reach of search-based algorithms.
Okay, just calculate the moves it would take to force the equal exchange of material from a given position. Generally as the game progresses and the board opens up it becomes inescapable.
After a certain point, when enough material has been removed from the board, looking for mate becomes trivial. Esp if you operate with such a commanding advantage as a queen...assuming you can force the equal exchange of all other material, it is possible to calculate checkmate within a couple of moves.
it's a simple matter to force equal trades; black cannot avoid the exchange of pieces forever, and if white plays a perfect game he will always win, without doubt.
Yes, that's pretty much the human intuition for why no one doubts that White will win. But it is very, very far from a mathematical proof.
Okay, just calculate the moves it would take to force the equal exchange of material from a given position.
Let me remind you that each ply has a branching factor of about 20, which means each move has a branching factor of several hundred. In most positions you'd be lucky to be able to calculate even one or two forced exchanges, let alone all the way to the end of the game.
A program attempting to prove victory operates in a very different context from a normal chess playing program — it is not allowed to prune any positions at all. We haven't even found the status of all seven piece endings yet! That is despite intense effort. See http://en.wikipedia.org/wiki/Endgame_tablebase
It is utterly, utterly inconceivable that a search-based approach will ever prove victory with Queen odds.
Not that I disagree with the overall point. A search-based approach might solve the game at queen odds, but it will not be any time soon; indeed, it would require devoting more computing resources than the planet has for a pretty significant amount of time.
No. randomwalker's point is that every branch must be proved to be a win.
>> "it would require devoting more computing resources than the planet has for a pretty significant amount of time"
This is also incorrect. As mentioned in other comments, the branching factor of chess implies that it is not brute force solvable using the entire universe as a computer--let alone just Earth.
No. randomwalker's point is that every branch must be proved to be a win.
In alpha-beta pruning, branches are only eliminated when they can only be reached through suboptimal play on one player's part. Suboptimal play on either player's part can clearly result in a loss for that player, but this is not very interesting.
This is also incorrect. As mentioned in other comments, the branching factor of chess implies that it is not brute force solvable using the entire universe as a computer--let alone just Earth.
This depends on how much time you're willing to devote to the task. My laptop (or my phone even) would be able to "solve" Chess given an infinite amount of time.
Now, the entire universe used as memory would be unable to store such a solution, but that isn't really germane to the decision problem.
> Yes, that's pretty much the human intuition for why no one doubts that White will win. But it is very, very far from a mathematical proof.
White will win if he plays perfectly: this is mathematically assured because white has an advantage of nine points. The only way for white to lose or to draw is by extreme error. The article assumes a perfect game, so I too assume that white will play a perfect game.
> Let me remind you that each ply has a branching factor of about 20, which means each move has a branching factor of several hundred. In most positions you'd be lucky to be able to calculate even one or two forced exchanges, let alone all the way to the end of the game.
As I recall, Deep Blue was calculating at a depth of more than eight moves...
> The Deep Blue chess computer which defeated Kasparov in 1997 would typically search to a depth of between six and eight moves to a maximum of twenty or even more moves in some situations. -- http://en.wikipedia.org/wiki/Deep_Blue_(chess_computer)
I find it difficult to believe that this wouldn't have improved or could not be improved upon, since this was based on the technology available in 1997. It's safe to say we've made progress since then.
While you may not be able to calculate an entire game, you could certainly calculate all the exchanges required to reach a properly winnable and calculable endgame. This is pretty damn close to being able to calculate a whole game...and you're assured victory.
As I said, once you get to a point of two kings and a queen, it's trivial.
I think one important point you're missing is that a chess engine looking to mathematically prove a victory is very different from the ones that play against grand masters. Deep Blue may have been able to calculate 6 to 8 moves in advance, but that was after it pruned all the moves that were obviously incorrect. When you're trying to prove something mathematically however, you have to assume that even something as silly as sacrificing a queen for a pawn with no obvious positional gains is a valid move and calculate all possible branches taken from that move until checkmate some 30 moves down the road.
For example, checkers is a vastly simpler game than chess. Yet, it took 18 years of constant computation to solve it [1]. Even a game as simple as tic-tac-toe has 255,168 possible games [2]. The estimated number of chess games is 10^10^50 [3].
It is important to remember that "the value of 1 Queen is equivalent the value to 9 Pawns" is only an heuristic. It helps the players to make decisions but nobody wins only because he has more points.
And the values are not written in stone. For example, Knights and Bishops usually have the same value, that is like tree pawns for each one. But if the position is open the Bishops are usually more valuable and if the position is closed the Knight are more valuable.
I'm also completely sure that in "Chess with Queen" White has a wining strategy, but being completely sure is not a proof. (I have been wrong before!)
Serious chess player here. I'm sure that Rook or Queen odds is a win but it is not a simple matter to force equal trades in chess games. It may be easier to force them with those odds (I never studied odds games much so I don't know a ton about how the dynamics change -- but neither have you), but I don't see why.
In general in chess the opening determines how easy it is to trade pieces and who has the option to trade how much. In some openings you cannot easily trade any pieces even if you want to. In others you can trade several if you want.
It's easy to give examples. White can trade his f1 bishop off easily in many e4 openings (not just against e5 but also against c5 nf3 d6/nc6). Or in d4/d5 openings with Bf4, black can play Bd6 to trade. There are also plenty of opportunities for white to play Bg5 and trade for a knight in various openings. But if you take other openings, like a closed French it's hard to trade pieces. Or a French with black trading his bishop on c3, it's hard to trade the other pieces. Or in a king's indian you don't necessarily have any good way to trade pieces without taking on some disadvantage. Or in a Najdorf obviously you can force some trading if you want to play Nd5 in some lines for example (I mean lines with e5 by black, not the piece sac lines) but when you do so it's not actually very good for white. This is one of many examples where going after a piece trade gets you some disadvantage.
Sorry for details, but really you can't comment on this stuff without knowing a hell of a lot more details than I just wrote. There are plenty of openings that do not allow convenient trading of many minor pieces, and trades of majors aren't all that common. Another good example is all IQP openings for either side which allow some trading but also get dynamic and interesting positions (and of course they are also positions where you're completely screwed down a queen). IQP positions are also interesting in that the IQP side must avoid trades b/c he has a losing endgame, and conservative players think it's bad for the IQP player but actually those positions are fine.
The strategy "just trade stuff and get to the end game" in normal chess is not easy to implement if your opponent doesn't want it.
tl;dr in chess it's usually pretty easy for white to trade one set of minor pieces but not necessarily any more, and doing so may be (slightly) bad for him.
Even in games with a huge advantage you can get in to situations where the other party can force a draw through the 3 repeated moves rule unless you offer some piece to break their possibility of forcing the draw. That can be expensive.
There are so many exceptions and pitfalls to work out that I'm sure that there is no 'shortcut' to the answer, but I do lean to support the idea that a queen advantage should be a win, in fact, I think that any piece or even just a pawn advantage should be a win, witness how many grand master games literally turned to watershed losses once that precarious balance was lost by as much as a single pawn.
Even so, the risk a of a draw is significant, the risk of an opponents win much less so.
Chess is all about the 'mate', not about who has the most points on the board anyway. Woe the player that forgets this even for a moment.
It's funny how the '3 repeating moves' rule makes chess actually much harder to reason about in terms of guaranteed endings because it means that a definite material advantage may still result in a draw, is there a factor known for how much this rule adds to the complexity of proving a win?
Another question that might be interesting is if a pawn advantage would be enough to cancel whites advantage by starting the game, I suspect that a single pawn is worth more than whites advantage but I'm not sure about this, and it might depend on the pawn (some pawns gone would allow white to deploy much faster with the price of the pawn only appearing later on in the game).
Repetition issues are rare in practice. I see why it could be important to a proof. But if I was playing a queen up against anyone I wouldn't repeat, I'd just make progress until I won. Repetition wouldn't come up, it's rare enough in close games. It's most common in (near or exactly even material) end games which queen odds games won't reach without hanging a queen.
A minor piece odds is a win for sure, any chess play can tell you that. But a pawn up .. that is unclear. You speak of GM games being decided by a pawn which of course happens but what you do not speak of is material sacrifices in GM games which are common too. And there's all those endgames which the pawn-down player wins (often by superior play, sometimes other reasons). And there's all those well known ways an extra pawn in an end game can be a draw, including for example king+pawn vs king is drawn unless it's set up so you can force promotion quickly (e.g. defending king is out of position). King+rook+pawn vs king+rook is common and this is also drawn in general (the defense is called the philidor position) unless the pawn-up side starts in an advantageous position (something equivalent to the lucena position)
One of the interesting facts about pawn odds is that you now have an open file b/c of your missing pawn. And pawns restrict moving out your pieces so having one missing can save time. The worst thing you could do is take away black's f7 pawn at the start. That is, first guess, a loss for black. Take away some other pawn and do it from white and it's a lot harder to guess.
The rule of thumb taught to beginners is that a pawn is worth 3 moves in the early game (and a knight is 3 pawns so you might think it's worth 9 moves, but that conversion doesn't work well, with 9 free moves you could set up checkmate). I think a pawn is worth somewhat less than 3 moves but it really varies and the comparison doesn't entirely make sense.
BTW there do exist well known openings leading to an unclear/unknown result that involve a piece sacrifice (lines in the King's Gambit or Najdorf for example). Losing a piece for nothing is a clear loss but various kinds of compensation are possible.
> Serious chess player here. I'm sure that Rook or Queen odds is a win but it is not a simple matter to force equal trades in chess games. It may be easier to force them with those odds (I never studied odds games much so I don't know a ton about how the dynamics change -- but neither have you), but I don't see why.
Seriously? Tournament chess player here with experience playing a range of levels up to and including grandmasters. You've got to be off your rocker if you think that it isn't easier to force trades with a QUEEN advantage.
White needs but play to the center. Play to open up the board will be inescapable. Black's only recourse (to avoid exchanges) would be to effectively give up the center or attempt some kind of closed defense like the Sicilian Dragon. Even in the case of the latter, he'll be forced into positions that will allow for the equal exchange of material or near equal material.
Assuming black moves only with the intent to avoid exchanges, he'll land himself in a terrible tactical position.
Lastly why are you comparing this kind of game to STANDARD lines that would involve balanced material? This is not a standard game. Did you seriously spend a paragraph talking about the French, King's Indian, and Najdorf? Of these, exchanges are not only highly probable, they are INEVITABLE regardless of, "...some lines for example (I mean lines with e5 by black, not the piece sac lines) but when you do so it's not actually very good for white." Virtually no standard line can be bad for white when he is up a QUEEN.
The point being: even if white is down a pawn or two pawns or even three pawns, he is STILL ahead. White has a considerable advantage in making exchanges because they are implicitly better for him even when he may lose minor material.
To anyone here saying that a queen is just a theoretical advantage: you know nothing about chess if you seriously believe that. I'd challenge you to play against a master, not even a grandmaster, with those odds and I guarantee you'll lose every game. You won't even stand a chance of drawing. Forget that. You will lose, horribly.
It isn't just a heuristic, the queen is an extremely powerful piece, with black down a queen the remainder of play is EXTREMELY unbalanced.
Whether or not you can prove this completely is another matter, but who cares? It's an obvious win in a perfect world. There's little to no mystery there. And it's hardly a reflection of the mechanics of an actual game, given how unbalanced it becomes.
I've been using the "eGTK" (Elementary GTK) theme for just about a year now. For my money, it's the cleanest, best designed GTK theme Gnome has yet seen. It still doesn't feel as smooth as I'd like, this due I suppose in part to Gnome and the way the desktop environment is integrated, but that said it's still quite nice to look at.
So are you confirming that it's a theme? I think most of us were confused from the title "Elementary OS," and the project didn't have an About page to explain what it was.
> I am assuming the worst possible situation here. I do firmly believe that OS X will eventually be just like iOS...
This will never happen. As others have already stated, Apple would lose their developers and because they don't command the market with OS X they would never risk doing that. Let's not forget the professionals that use Macs: the design industry, the film industry, the music industry: these rely in some cases on third-party applications like the Adobe Suite, these industries are another reason why OS X cannot ever become iOS; they require a fundamentally open, extensible environment. Add that to the fact Apple has stated that the App Store will never be the only method through which developers may publish their applications to the OS X platform, and it just will not happen.
