Because most Linux users are geeks, and us geeks enjoy customization.*
Actually, I don't think most geeks prefer customisation per se. Many clearly do, but I for one find it a sometimes-necessary evil rather than a benefit.
Personally, what I want is something that works right. That does what I need, and then gets out of my way. The best way is for the software to work right to start with. Being able to customise it is a workaround for it not working properly in the first place. Sure, it's a lot better than nothing. But it's only a workaround.
I've seen both sides of this. For a decade, I was an Atari user. Yes, yes, I know, stop sniggering at the back. I spent a lot of time getting it set up, installing lots of utilities and gubbins. And the end result was worth it: a modern GUI desktop, task bar, Start button, great integration with my command shell, good use of the available sound and graphics, TrueType fonts, web browser, pre-emptive multitasking, email, printer, scanner, and lots more I've forgotten about, all nicely integrated. Nothing exciting today, but in the 1990s it was quite comfortable and standard. But it took an awful lot of time keeping up with the latest developments, downloading, installing, setting up, tweaking, reading docs, rebooting, etc.
Fast-forward a few years, and I'm now a Mac user. Oi, I said STOP sniggering at the back! And now I spend hardly any time customising it. Why? Because I don't need to. It's already comfortable, powerful, standard, integrated, etc.** All the things I used to spend time and effort setting up are there already. So instead, I spend most of my time actually doing things with it. Do I miss that time spend hacking around trying to get things to work decently? No, of course not. I spend most of it reading Slashdot...
So for me, the question is "How do I get a decent environment to work in?" And while customisation (like skinning) is an answer, it's the wrong answer. The right answer is to provide decent software to start with.
(* 'we geeks prefer customisation', not 'us geeks'. You wouldn't say 'us prefer', would you? Addition of the extra qualifier doesn't alter the noun case. And yes, I am a member of the Campaign for Real Pedantry. CaRP, for short.)
(** I am of course talking about Mac OS X. I did use Mac OS 9 for six months, and hated every minute. There was a system that did half of everything wrong, and provided absolutely no way to escape, either. For an old-school Unix hacker like me, it was purgatory.)
I know many physicians who prescribe placebo treatments and tests. I have trouble doing this
Would you have trouble prescribing a placebo if you told the patient what it was? If I recall correctly, at least one study showed benefit from taking a placebo even when it was carefully explained that there were no active ingredients in the pills.
For more information, check out a pair of radio programmes called 'Placebo' by Dr Ben Goldacre. (They were on BBC Radio 4 last year.) In fact, I'm surprised no-one's mentioned him before; his blog (Bad Science), occasional radio programmes and other media appearances are an inspiration to anyone interested in evidence-based medicine or the media's blatant misrepresentation of science in general.
It's not very sexy, but it's fascinating and readable. I remember coming across it in Dillons bookshop, not knowing the name, and flicking through. Half an hour later, when I realised the time, I knew I had to buy it! Other books go into more exhausting detail (Knuth in particular), or cover a wider range (Knuth again!), or more modern ideas or languages. But Sedgewick is a great read, and I've been through it several times.
It covers all the basics (maths, searching, sorting, strings, graphs, and touches on FFTs and hardware and optimisation), and gives enough detail that you could go off and write some programs yourself. But more importantly, it explains them: how each algorithm works, what it's trying to achieve, how it behaves, and why. And it's because it explains the ideas so well that I'd recommend it. After every section I felt I'd learned something -- not because I had to, but for the sheer pleasure of understanding something new and interesting.
Other recommendations: Effective Java (a staggering amount of insight into the language), Thinking in Java (by someone who understands that language is more than just syntax), Deep C Secrets (again a pile of insight, interspersed with anecdotes and some rather off-the-wall diversions), Programming Pearls and More Programming Pearls (problem-solving in bite-sized chunks -- a little dated but still interesting). Plus I've already mentioned Knuth. K&R is well done, though narrow in scope. I find Design Patterns useful, but more for clarifying things I've already seen than for learning new things. I've never actually read The Mythical Man-Month, but people I respect mention it, so I'm sure it's well worth reading too!
Of course, times being what they are, especially in this field, a lot of interesting stuff is on-line. Some hat should go without saying hereabouts include the latest Jargon File, some of Eric Raymond's books, and more online documentation and archives than anyone but Google can cope with.
Here in the UK, the media (or at least, a good number of newspapers, TV programmes, etc.) have been trumpeting the possibility of a link between this vaccine and autism for the last few years -- despite the only 'lab' claiming any link turning out to be a shed belonging to someone with a correspondence degree.
(It's interesting that most of these vaccine scares seem to be restricted to one country at a time, even though the vaccines themselves are used across the world...)
As others have said, Dr Goldacre is a voice well worth listening to. At least partly because he shows, time and time again, that you have to go to the actual results, the facts and figures, to see what's really going on. He's done a lot of that work. It's a shame that mainstream journalism rarely follows suit.
Your plan looks a lot more sane than some current copyright schemes, but it may need more work.
Take, for example, the Lord of the Rings movies. They involved two years of pre-production, a year of principal photography, and then three years of post-production and pickups. If copyright only applies once the finished work is registered, that leaves nearly six years for material to be stolen and used freely. On the other hand, if it applies automatically, then that takes six years off the time during which the result may be profitable. You don't suggest a timespan for movies, but even at ten years that's the majority of the time gone; and at five years (as you suggest for recordings), the work would be out of copyright before it was even finished!
Of course, few movies or albums take anywhere near as long to make, but you can see how short copyright times could promote speed, even at the expense of quality or scope.
And what about releases of older material? I know bands who have reached into their 'vaults' and released live and studio material recorded decades earlier; under short copyright terms, there would be no incentive to release such recordings at all.
Also, I believe that the title track from Jean-Michel Jarre's album 'Waiting for Cousteau' was computer-generated in some way -- though I've no firm evidence other than intuition, and other unsubstantiated claims such as this one.
Coz what with the JVM, CLR, Parrot, p-code machine, LLVM, Rubinius, SWF, Lua, Squeak, Dis, Waba, Z-machine, and a whole host of others, a new VM is just what we need!!!
...
Er, sorry about that. My sarcasm chip seems to have overloaded. But several existing VMs already let you run code written in many different languages. And issues of static (before-the-event) compilation can be rather different from dynamic (while it's running) compilation; the intermediate code may be wholly unsuitable for running in a VM efficiently. I doubt that, even if the GCC back-end were rewritten as a dynamic compiler, it would be much good without the sort of work equivalent to a rewrite from scratch. Ideally, it would have the benefit of compatibility with a lot of existing C etc. code; but that presumes no code changes would be needed; other than that, I can't think of any particular advantages.
Could the GCC folks hack a VM out of their back-end? Probably. Would it be worth doing? I can't see why.
if you don't vote for the most popular candidate, you[r] vote is worthless
Why??? I see this view -- though only rarely stated so explicitly -- rather a lot, and I can't explain how it appals me.
If they're the most popular candidate, they're going to win anyway, so why do they need your vote? If you could 'waste' a vote, surely that'd be by voting for them!
But a vote can't be worthless, whether it's for the winner or not. Winning isn't everything; a 51% majority wins just as surely as a 100% one, but will surely lead to a rather different climate afterwards. And anyway, could you be so certain which candidate was most popular, especially if every other voter was hiding their true feelings too?
Of course, things would be rather different in some other voting system. First-past-the-post, as used in my country as well as the US, does risk these issues. But even in STV or instant run-off, your argument would lead to the most popular candidate being put at the top of every vote and winning by 100%. Is that what you want? If not, then put your vote where your mouth is!
Yeah. Or they could give away free Macs to everyone. Either way, it'd be great for the few months the company would survive, but I don't see much long-term gain for anyone.
Nope, but... [fx: reaches into pocket]...this 'ere 5mx is always with me, and gets used for, well, just about everything. (I've extolled its virtues enough times already, so I'll leave it at that for now.)
Admittedly, I don't stress it on this 'ere XP box as much as I do on my G5 at home, where my norm is probably four or five windows holding anything up to 50 or 60 tabs between them, and I leave it running all the time. (Or at least, as long as I can until it gets unbearable and I have to take my life in my hands and hope that it doesn't forget all my tabs when I restart!)
What's most annoying with the Mac version, though, is the lack of responsiveness. I don't mind it taking ten seconds or so to open a page; well, I do, but it'd be very bearable if it didn't also completely freeze the GUI for the first half of that, and then react very slowly and jerkily for the rest.
I can't say I like XP, but there at least Firefox is relatively responsive when it's busy. Having the Mac OS X version beachball whenever it does anything is most annoying.
Oh, and while we're talking of Mac Firefox woes, does anyone else find that it won't show Flash unless it doesn't fit on screen? If I scroll down so that the the Flash area goes off the top, even just by a pixel, then it appears fine; but if the whole of the Flash area is visible then it won't redraw properly; often it's completely blank. Could this be a plugin issue, or am I doing something wrong?
what application or kernel area would you attempt to improve[...]?
That's the problem, right there. As long as developers continue to view each application, each kernel or kernel area, each driver, each service, as its own little kingdom, then Linux will fall short of its potential. I believe the real gains will only occur when people consider the system as a whole.
Disclaimer time: I've never run Linux myself; what I know is from a few glimpses of friends' desktops, and reading an awful lot on this site and many others. But I hope I'm not entirely ignorant about it.
Many years ago, I ran an Atari STE, and later an Atari Falcon. It became a surprisingly powerful system: pre-emptive multitasking, standard web browsers and email clients, a task bar and start button, Unix shells, gcc, perl, etc. It did most things I wanted -- many of them very well indeed. However (which is why I bring up all this history) that came at a price: discovering, installing, configuring, and updating all the many apps and utilities which were needed, and getting them all working properly together, took a lot of my time and effort. Time not spent actually using it to do things!
When I switched to (the newly-released) Mac OS X, it was as if a great weight was lifted from my shoulders. Here was another powerful system that worked really well, but which didn't need an input of time and effort to make it so! The Mac's motto 'It Just Works' is pretty close to the truth: I spend very little time searching for apps, utilities, or drivers, persuading bits of the system to work properly, upgrading, recovering after crashes, or any of the other ways I wasted time on my Ataris or still do on my work PC. And that's not due to any single feature, application, or utility: it's in the way everything works together. The whole system has been designed to get out of my way and let me do my work with as little interruption as possible, right from the moment I first turned it on. And from all I've heard about Linux, it sounds like that's where it could benefit most. Is that a valid impression?
...do you think Microsoft will throw them a biscuit, too?
Seriously, I'm sure it's an impressive achievement, but I can't help thinking that in the long term, the only people who will end up benefiting from this are in the vicinity of Redmond.
Agreed. While generics are very powerful and make for nice, clear, safe code, they only do so for simple examples. When you get to more complicated uses, it all starts to get very painful.
And that's because, as you say, it's a kludge.
Why oh why oh why oh why oh why didn't they put it into the runtime? That would have given a huge amount more safety, and probably efficiency. The main excuse I heard was that they didn't want to change the JVM or the bytecode -- but then they went and changed both in the same release (to add annotations) anyway!
So now we have something that can look neat in the code but isn't any more than syntactic sugar -- it's dead easy to subvert and end up with something completely unexpected in a collection, or passed as a parameter to a generic method. In practice, it's still executing all the same casts, so it's unlikely to be any more efficient than the old way -- which is what it pretty much boils down to once it gets to the runtime anyway.
And it has stupid limitations. For example, although a you can write a class that's comparable to Integers (by implementing Comparable) or one that's comparable to Strings (by implementing Comparable), you can't write one that's comparable to both! (The reason is that the compiler translates both to Comparable, and you can't provide two different implementations for the same interface.) Stupid stupid stupid. In your generic class, although you can define members and accept method arguments that appear to be of the parameterised type, you can't create an array of that type, nor create a new instance; you can't even use reflection to find out what that parameterised type is. In fact anything mixing arrays and generics is a world of pain, and there are many other hidden gotchas. Not to mention that you have to make it all defeatable to support existing non-generic code...
So while type erasure initially sounds like a nice way to implement generics (it has almost no overhead at runtime, is backward-compatible, and simple examples work well), I think it was the wrong decision. Java used to be a relatively simple language, with all the complexity hidden away in the runtime. Now it's the other way around; there's a huge amount of complexity foisted onto the compiler, and hence the programmer, simply to maintain this pretence to the runtime. And we all suffer for it.
Indeed. But I reckon that's a very mixed blessing for Google.
Google's ability to rank the search results is based on links from other pages. Now, when the web was young and blue-eyed and men everywhere were free, they navigated using lists of bookmarks, and by following links. In many cases they published their lists of bookmarks on their web sites, so that others could discover interesting sites. This meant that most interesting sites had lots of links to them, which Google could harvest and use.
But now, when the main route to finding sites is to use Google, Google itself has less to go on. With fewer people creating links for it to harvest, how can it work out its rankings?
Of course, it still has lots of links from blogs and automated sources. But are those as relevant, as useful from Google's PoV? Have we reached a point where even the best search algorithms are being starved of data to work from?
I personally don't think it's the OS that makes you productive, but the software and how you use it.
Perhaps that's how it should be. Certainly, the OS should get out of the way and let you get on with something productive. However, in practice, we're not always very close to that ideal.
I use XP at work every day, and every day it gets in my way in countless little ways, from snatching focus when I'm typing, to making my Explorer windows bigger each time, to taking unnecessarily long to unminimise apps, to getting confused about whether I'm pressing any of the control keys after switching away from Remote Desktop, to freezing the entire window manager when an app is slow responding, to continually popping up windows on the wrong desktop (using MS's own Virtual Desktop Manager), to presenting lots of unnecessary and annoying fake directories while pretending that my desktop is the root of the filesystem, to making it unreasonably hard to enter accented characters, to... to... Well, I could rant on for hours more!
I use Mac OS X at home, and though it's far from perfect, it annoys me far less because it keeps out of my way more. I'm doing different things, of course, so it's hard to compare; but I certainly feel more productive with it.
Now, I obviously have more emotional investment in a machine that I chose and bought than in one I'm expected to use at work. But I've been using Windows regularly for over twice as long (since 3.11) as Mac OS X (since the Public Beta), so it's not that I'm more familiar or more experienced with the Mac.
I haven't used Vista for more than a few minutes. (Not that I'm bothered.) Any Vista users care to mention whether they thing Vista has a better 'gets out of my way and lets me work' score than XP?
Slashdot Mirror
"How much polonium do you take in your tea?"
Actually, I don't think most geeks prefer customisation per se. Many clearly do, but I for one find it a sometimes-necessary evil rather than a benefit.
Personally, what I want is something that works right. That does what I need, and then gets out of my way. The best way is for the software to work right to start with. Being able to customise it is a workaround for it not working properly in the first place. Sure, it's a lot better than nothing. But it's only a workaround.
I've seen both sides of this. For a decade, I was an Atari user. Yes, yes, I know, stop sniggering at the back. I spent a lot of time getting it set up, installing lots of utilities and gubbins. And the end result was worth it: a modern GUI desktop, task bar, Start button, great integration with my command shell, good use of the available sound and graphics, TrueType fonts, web browser, pre-emptive multitasking, email, printer, scanner, and lots more I've forgotten about, all nicely integrated. Nothing exciting today, but in the 1990s it was quite comfortable and standard. But it took an awful lot of time keeping up with the latest developments, downloading, installing, setting up, tweaking, reading docs, rebooting, etc.
Fast-forward a few years, and I'm now a Mac user. Oi, I said STOP sniggering at the back! And now I spend hardly any time customising it. Why? Because I don't need to. It's already comfortable, powerful, standard, integrated, etc.** All the things I used to spend time and effort setting up are there already. So instead, I spend most of my time actually doing things with it. Do I miss that time spend hacking around trying to get things to work decently? No, of course not. I spend most of it reading Slashdot...
So for me, the question is "How do I get a decent environment to work in?" And while customisation (like skinning) is an answer, it's the wrong answer. The right answer is to provide decent software to start with.
(* 'we geeks prefer customisation', not 'us geeks'. You wouldn't say 'us prefer', would you? Addition of the extra qualifier doesn't alter the noun case. And yes, I am a member of the Campaign for Real Pedantry. CaRP, for short.)
(** I am of course talking about Mac OS X. I did use Mac OS 9 for six months, and hated every minute. There was a system that did half of everything wrong, and provided absolutely no way to escape, either. For an old-school Unix hacker like me, it was purgatory.)
Anyone who hasn't seen it already should read What Colour Are Your Bits? which discusses this in detail.
Would you have trouble prescribing a placebo if you told the patient what it was? If I recall correctly, at least one study showed benefit from taking a placebo even when it was carefully explained that there were no active ingredients in the pills.
For more information, check out a pair of radio programmes called 'Placebo' by Dr Ben Goldacre. (They were on BBC Radio 4 last year.) In fact, I'm surprised no-one's mentioned him before; his blog (Bad Science), occasional radio programmes and other media appearances are an inspiration to anyone interested in evidence-based medicine or the media's blatant misrepresentation of science in general.
It's not very sexy, but it's fascinating and readable. I remember coming across it in Dillons bookshop, not knowing the name, and flicking through. Half an hour later, when I realised the time, I knew I had to buy it! Other books go into more exhausting detail (Knuth in particular), or cover a wider range (Knuth again!), or more modern ideas or languages. But Sedgewick is a great read, and I've been through it several times.
It covers all the basics (maths, searching, sorting, strings, graphs, and touches on FFTs and hardware and optimisation), and gives enough detail that you could go off and write some programs yourself. But more importantly, it explains them: how each algorithm works, what it's trying to achieve, how it behaves, and why. And it's because it explains the ideas so well that I'd recommend it. After every section I felt I'd learned something -- not because I had to, but for the sheer pleasure of understanding something new and interesting.
Other recommendations: Effective Java (a staggering amount of insight into the language), Thinking in Java (by someone who understands that language is more than just syntax), Deep C Secrets (again a pile of insight, interspersed with anecdotes and some rather off-the-wall diversions), Programming Pearls and More Programming Pearls (problem-solving in bite-sized chunks -- a little dated but still interesting). Plus I've already mentioned Knuth. K&R is well done, though narrow in scope. I find Design Patterns useful, but more for clarifying things I've already seen than for learning new things. I've never actually read The Mythical Man-Month, but people I respect mention it, so I'm sure it's well worth reading too!
Of course, times being what they are, especially in this field, a lot of interesting stuff is on-line. Some hat should go without saying hereabouts include the latest Jargon File, some of Eric Raymond's books, and more online documentation and archives than anyone but Google can cope with.
Other interesting articles include The Programmer's Stone, a guide to writing Unmaintainable Code, The Ten Commandments for C Programmers (annotated edition), Ken Thompson's Reflections on Trusting Trust, What Colour are your Bits?, and Guy Steele's Growing a Language.
Here in the UK, the media (or at least, a good number of newspapers, TV programmes, etc.) have been trumpeting the possibility of a link between this vaccine and autism for the last few years -- despite the only 'lab' claiming any link turning out to be a shed belonging to someone with a correspondence degree.
(It's interesting that most of these vaccine scares seem to be restricted to one country at a time, even though the vaccines themselves are used across the world...)
As others have said, Dr Goldacre is a voice well worth listening to. At least partly because he shows, time and time again, that you have to go to the actual results, the facts and figures, to see what's really going on. He's done a lot of that work. It's a shame that mainstream journalism rarely follows suit.
Take, for example, the Lord of the Rings movies. They involved two years of pre-production, a year of principal photography, and then three years of post-production and pickups. If copyright only applies once the finished work is registered, that leaves nearly six years for material to be stolen and used freely. On the other hand, if it applies automatically, then that takes six years off the time during which the result may be profitable. You don't suggest a timespan for movies, but even at ten years that's the majority of the time gone; and at five years (as you suggest for recordings), the work would be out of copyright before it was even finished!
Of course, few movies or albums take anywhere near as long to make, but you can see how short copyright times could promote speed, even at the expense of quality or scope.
And what about releases of older material? I know bands who have reached into their 'vaults' and released live and studio material recorded decades earlier; under short copyright terms, there would be no incentive to release such recordings at all.
Also, I believe that the title track from Jean-Michel Jarre's album 'Waiting for Cousteau' was computer-generated in some way -- though I've no firm evidence other than intuition, and other unsubstantiated claims such as this one.
Hey, wow! What a great idea!
Coz what with the JVM, CLR, Parrot, p-code machine, LLVM, Rubinius, SWF, Lua, Squeak, Dis, Waba, Z-machine, and a whole host of others, a new VM is just what we need!!!
...
Er, sorry about that. My sarcasm chip seems to have overloaded. But several existing VMs already let you run code written in many different languages. And issues of static (before-the-event) compilation can be rather different from dynamic (while it's running) compilation; the intermediate code may be wholly unsuitable for running in a VM efficiently. I doubt that, even if the GCC back-end were rewritten as a dynamic compiler, it would be much good without the sort of work equivalent to a rewrite from scratch. Ideally, it would have the benefit of compatibility with a lot of existing C etc. code; but that presumes no code changes would be needed; other than that, I can't think of any particular advantages.
Could the GCC folks hack a VM out of their back-end? Probably. Would it be worth doing? I can't see why.
Sorry if that sounds a bit negative :)
Why??? I see this view -- though only rarely stated so explicitly -- rather a lot, and I can't explain how it appals me.
If they're the most popular candidate, they're going to win anyway, so why do they need your vote? If you could 'waste' a vote, surely that'd be by voting for them!
But a vote can't be worthless, whether it's for the winner or not. Winning isn't everything; a 51% majority wins just as surely as a 100% one, but will surely lead to a rather different climate afterwards. And anyway, could you be so certain which candidate was most popular, especially if every other voter was hiding their true feelings too?
Of course, things would be rather different in some other voting system. First-past-the-post, as used in my country as well as the US, does risk these issues. But even in STV or instant run-off, your argument would lead to the most popular candidate being put at the top of every vote and winning by 100%. Is that what you want? If not, then put your vote where your mouth is!
Yes there is: read every single patent that's been filed with your Patent Office, read every line of Linux code, and try to match them up.
Nope, but... [fx: reaches into pocket] ...this 'ere 5mx is always with me, and gets used for, well, just about everything. (I've extolled its virtues enough times already, so I'll leave it at that for now.)
Admittedly, I don't stress it on this 'ere XP box as much as I do on my G5 at home, where my norm is probably four or five windows holding anything up to 50 or 60 tabs between them, and I leave it running all the time. (Or at least, as long as I can until it gets unbearable and I have to take my life in my hands and hope that it doesn't forget all my tabs when I restart!)
What's most annoying with the Mac version, though, is the lack of responsiveness. I don't mind it taking ten seconds or so to open a page; well, I do, but it'd be very bearable if it didn't also completely freeze the GUI for the first half of that, and then react very slowly and jerkily for the rest.
I can't say I like XP, but there at least Firefox is relatively responsive when it's busy. Having the Mac OS X version beachball whenever it does anything is most annoying.
Oh, and while we're talking of Mac Firefox woes, does anyone else find that it won't show Flash unless it doesn't fit on screen? If I scroll down so that the the Flash area goes off the top, even just by a pixel, then it appears fine; but if the whole of the Flash area is visible then it won't redraw properly; often it's completely blank. Could this be a plugin issue, or am I doing something wrong?
The first quote is Hanlon's razor, from Robert J. Hanlon in 1980.
And while you may have coined the second, 13,600 hits in Google suggest you weren't the first to do so. It's been attributed to Vernon Schryver, Paul Ciszek, been dubbed Clark's Law (no 'e') and Grey's Law, and even seen here on Slashdot.
(It's a great line, nonetheless!)
That's the problem, right there. As long as developers continue to view each application, each kernel or kernel area, each driver, each service, as its own little kingdom, then Linux will fall short of its potential. I believe the real gains will only occur when people consider the system as a whole.
Disclaimer time: I've never run Linux myself; what I know is from a few glimpses of friends' desktops, and reading an awful lot on this site and many others. But I hope I'm not entirely ignorant about it.
Many years ago, I ran an Atari STE, and later an Atari Falcon. It became a surprisingly powerful system: pre-emptive multitasking, standard web browsers and email clients, a task bar and start button, Unix shells, gcc, perl, etc. It did most things I wanted -- many of them very well indeed. However (which is why I bring up all this history) that came at a price: discovering, installing, configuring, and updating all the many apps and utilities which were needed, and getting them all working properly together, took a lot of my time and effort. Time not spent actually using it to do things!
When I switched to (the newly-released) Mac OS X, it was as if a great weight was lifted from my shoulders. Here was another powerful system that worked really well, but which didn't need an input of time and effort to make it so! The Mac's motto 'It Just Works' is pretty close to the truth: I spend very little time searching for apps, utilities, or drivers, persuading bits of the system to work properly, upgrading, recovering after crashes, or any of the other ways I wasted time on my Ataris or still do on my work PC. And that's not due to any single feature, application, or utility: it's in the way everything works together. The whole system has been designed to get out of my way and let me do my work with as little interruption as possible, right from the moment I first turned it on. And from all I've heard about Linux, it sounds like that's where it could benefit most. Is that a valid impression?
Seriously, I'm sure it's an impressive achievement, but I can't help thinking that in the long term, the only people who will end up benefiting from this are in the vicinity of Redmond.
As long as we don't go straight to the Month of Stomach Bugs...
And that's because, as you say, it's a kludge.
Why oh why oh why oh why oh why didn't they put it into the runtime? That would have given a huge amount more safety, and probably efficiency. The main excuse I heard was that they didn't want to change the JVM or the bytecode -- but then they went and changed both in the same release (to add annotations) anyway!
So now we have something that can look neat in the code but isn't any more than syntactic sugar -- it's dead easy to subvert and end up with something completely unexpected in a collection, or passed as a parameter to a generic method. In practice, it's still executing all the same casts, so it's unlikely to be any more efficient than the old way -- which is what it pretty much boils down to once it gets to the runtime anyway.
And it has stupid limitations. For example, although a you can write a class that's comparable to Integers (by implementing Comparable) or one that's comparable to Strings (by implementing Comparable), you can't write one that's comparable to both! (The reason is that the compiler translates both to Comparable, and you can't provide two different implementations for the same interface.) Stupid stupid stupid. In your generic class, although you can define members and accept method arguments that appear to be of the parameterised type, you can't create an array of that type, nor create a new instance; you can't even use reflection to find out what that parameterised type is. In fact anything mixing arrays and generics is a world of pain, and there are many other hidden gotchas. Not to mention that you have to make it all defeatable to support existing non-generic code...
So while type erasure initially sounds like a nice way to implement generics (it has almost no overhead at runtime, is backward-compatible, and simple examples work well), I think it was the wrong decision. Java used to be a relatively simple language, with all the complexity hidden away in the runtime. Now it's the other way around; there's a huge amount of complexity foisted onto the compiler, and hence the programmer, simply to maintain this pretence to the runtime. And we all suffer for it.
Indeed. But I reckon that's a very mixed blessing for Google.
Google's ability to rank the search results is based on links from other pages. Now, when the web was young and blue-eyed and men everywhere were free, they navigated using lists of bookmarks, and by following links. In many cases they published their lists of bookmarks on their web sites, so that others could discover interesting sites. This meant that most interesting sites had lots of links to them, which Google could harvest and use.
But now, when the main route to finding sites is to use Google, Google itself has less to go on. With fewer people creating links for it to harvest, how can it work out its rankings?
Of course, it still has lots of links from blogs and automated sources. But are those as relevant, as useful from Google's PoV? Have we reached a point where even the best search algorithms are being starved of data to work from?
Perhaps that's how it should be. Certainly, the OS should get out of the way and let you get on with something productive. However, in practice, we're not always very close to that ideal.
I use XP at work every day, and every day it gets in my way in countless little ways, from snatching focus when I'm typing, to making my Explorer windows bigger each time, to taking unnecessarily long to unminimise apps, to getting confused about whether I'm pressing any of the control keys after switching away from Remote Desktop, to freezing the entire window manager when an app is slow responding, to continually popping up windows on the wrong desktop (using MS's own Virtual Desktop Manager), to presenting lots of unnecessary and annoying fake directories while pretending that my desktop is the root of the filesystem, to making it unreasonably hard to enter accented characters, to... to... Well, I could rant on for hours more!
I use Mac OS X at home, and though it's far from perfect, it annoys me far less because it keeps out of my way more. I'm doing different things, of course, so it's hard to compare; but I certainly feel more productive with it.
Now, I obviously have more emotional investment in a machine that I chose and bought than in one I'm expected to use at work. But I've been using Windows regularly for over twice as long (since 3.11) as Mac OS X (since the Public Beta), so it's not that I'm more familiar or more experienced with the Mac.
I haven't used Vista for more than a few minutes. (Not that I'm bothered.) Any Vista users care to mention whether they thing Vista has a better 'gets out of my way and lets me work' score than XP?
Which, er, historically, hasn't always been their first choice...