"Aren't we talking about the high level here? We aren't talking about which one has a better core but the user experience."
User experience is all about getting the details right. In terms of high level architecture you could argue that there's no difference between XP or MacOS X (or KDE or Gnome), because they're all a WIMP (windows, icons, mouse) GUI over a multi-tasking OS. But the details are very different between them, and those details matter to users. For example, look at the OS install/registration process. Installing MacOS X requires few clicks, and setup and registration is optional. Installing XP is quite involved, and registration is not only required, it's enforced by a central server, and re-registration is required if you make any substantial system configuration changes.
I think that he's right that MacOS X is too complex to be a simple appliance. But I think that general purpose computers are by definition complex, because they can be used for *anything*, and his vision holds more true for specialized devices. For example, the iPod is elegant and transparent to use.
That being said, I'm sure that usability could always be improved. But I don't agree that there's not much difference between XP and MacOS X -- while they're similar at a very high level (mouse/windows/icons over multi-tasking OS, etc.), MacOS X is better in almost every detail. But it's best not to get into a religious war here. I can only guess that Jeff has such a radical vision for how computers could be that from his perspective XP and MacOS X aren't too different.
Actually, MacOS X has always (since it was OPENSTEP, then Rhapsody) run on x86. Apple shipped (to developers) Rhapsody for x86, and it worked fine (if you had hardware that they had drivers for). And I've been assured by Apple systems software people that they're still maintaining the x86 build (and the other CPU's that OPENSTEP ran on) in order to make sure that Apple doesn't accidentally break portability. This applies to Cocoa app's, but not Carbon (i.e. old MacOS).
"Maybe I'm missing something, but what about the legality of that site (allofmp3.com)? "
It's not legal. Despite the claim on their web site, they're: - Selling music digitally that the artists have never agreed to sell digitally (e.g. Led Zeppelin, The Beatles). - They've never (AFAIK) actually paid any royalties to any musicians, composers or music companies.
So while the technology of the site is impressive, as is often the case, this deal that looks "too good to be true" is (IMO) a scam, in this case directly ripping off artists, composers and music companies. That's (IMO) why PayPal stopped processing their payments.
"They used off the shelf parts to give XBox short time to market in exchange for their own fab facilities and eventually being able to break even on hardware"
IMO, the real mistakes that MS made with the Xbox are the same mistakes they've made every time they launch (and fail with) any product outside of the PC desktop. Those mistakes are:
1) They design the product that they want to sell, not what people want to buy. That is, they start with the assumption that they're going to apply Windows into new markets and extract more revenue, instead of objectively determining what customers want and providing the best possible solution. This is why WinCE has done terribly against PalmOS, MSN has lost a fortune as an ISP, XP's adoption rate has been miserable, their several generations of pen-based products have failed, their Windows-based set-top box failed, etc.
2) They think that they can win with with more "features", so their products are overdesigned, overly complex, and require too much hardware. This works when selling desktop software, because reviewers love making checklists of features, and inefficiency and instability are relatively accepted, but fails when other factors are more important. For example, NT has done badly (compared to projections) as a server OS because it's loaded with "features" that are useless in a server farm, but which increase software complexity, reducing efficiency and security. It's typical of this issue that when MS launches each new version of NT, they proudly announce how many hundreds of thousands of lines of new code it contains, and how many nifty new features, which is the _opposite_ of what any experienced system administrator wants to hear about an OS update. New lines = more bugs, and more weird things to learn how to deal with. All someone running a server farm wants to hear about an OS release is that it's exactly the same as the previous revision but has higher performance and fixes bugs. Ditto embedded OS's (where MS has fallen flat). Similarly, WinCE was far too complex when it launched, so that it had more "features" than PalmOS, but was less usable, and was only supported on extremely expensive, cumbersome hardware with short battery life.
3) They overvalued Windows API's and integration. Since those are the key to their desktop success, they think that they're important elsewhere. But in a PDA, you have to fundamentally rewrite your app's already (to use the touchscreen, to make sense on a much smaller screen, etc.) so it's not that hard to use a new app framework, particularly if the new framework is (like PalmOS) simpler to use. And on servers, you don't want to run Windows desktop app's, but specialized server app's, so again there's less of an overlap than you would think. And if you compare WinCE and native games written for the Sega Dreamcast, you'll notice that the non-Windows games ran better, and sold better, than the PC ports. So while in theory the shared API with the PC might be an advantage, in practice (on the only platform to offer both options) the Windows API's had negative value.
4) They're far too aggressive in their business dealings. Everyone who partners with MS is attracted by the potential of making a fortune through MS's distribution channels, and loses a fortune on the details of the contracts. And since MS designs the technology to lock partners into them, nobody wants to even use their technology. For example, look at MS's networked game architecture -- it requires you to write servers in NT that run in MS' infrastructure, and to use MS services for everything. Some of these are conveient for users, but all of it forces game companies to be dependent on MS to do business, which means that MS can prevent games from shipping, or even shut systems down, with nothing that the game company can do about it but ask politely (or, I suppose, sue). So when MS says "NO XBOX SOFTWARE PRODUCT(S) MAY BE PUBLISHED, OR DISTRIBUTED TO END USERS, EXCEPT BY A LICENSED PUBLISHER PURSUANT TO AN XBOX DEVELOPMENT KIT AND X
"Israel is building walled and ever shrinking ghettos in to which they are shoving all the Arab's in their borders"
Dramatic, but exactly wrong. Israel is walling itself _in_ to keep out the non-Israeli Arabs that keep trying to sneak into Israel to kill them.
If you're really concerned about the plight of the Arabs, you should probably take that up with the Arab governments. It's astonishing to me that Arab Israeli's are so much better off than the Arab non-Israeli's -- they're better off economically, live in a democracy where they have a meaningful vote, have free speech, etc., which is more than you can say for them in, say, Saudi Arabia, Egypt, or Pakistan. If the Arab world really cared about the Palestinians they probably wouldn't have kicked them out of every Arab country and forced them into UN camps.
"If you look at all of his positions on Iraq he has all bases covered. He voted for the war, but now he is against it..."
If you read what he said, rather than absurd snippets extracted to mislead rather than inform, you'll see that Kerry has had a consistent position on the Iraq war. He voted to give the President the authority to use military force as leverage in negotiations, and to use it only as a last resort. So despite the fact that inspections were accurately determining that Iraq had no weapons of mass destruction, Bush insisted on going to war. Kerry supported having Iraq pay for their own reconstruction (i.e. from oil revenue), while Bush insisted that the US taxpayers cover the entire expense. Bush threatened to veto Kerry's bill, and then Kerry voted against Bush's bill. If Bush had been forced to actually cast the veto, would you have argued that Bush opposed funding the soldiers? Yes, back when Kerry believed Bush, Rice and Powell's lies about the evidence supporting their claims that Iraq was behind 9/11, Iraq had WMD's and Iraq being an immanent threat, he supported invading. Once it became clear that in fact Iraq had nothing to do with 9/11, had no WMD's, and was not a threat to the US, he opposed the war.
You should be asking yourself why Bush still supports the war when all of the justifications that he presented before invading turned out to be erroneous.
"If you look at all of his positions on Iraq he has all bases covered. He voted for the war, but now he is against it..."
If you read what he said, rather than absurd snippets extracted to mislead rather than inform, you'll see that Kerry has had a consistent position on the Iraq war. He voted to give the President the authority to use military force as leverage in negotiations, and to use it only as a last resort. So despite the fact that inspections were accurately determining that Iraq had no weapons of mass destruction, Bush insisted on going to war, which Kerry opposed. Kerry supported having Iraq pay for their own reconstruction (i.e. from oil revenur), while Bush insisted that the US taxpayers cover the entire expense. Bush threatened to veto Kerry's bill, and then Kerry voted against Bush's bill. If Bush had been forced to actually cast the veto, would you have argued that Bush opposed funding the soldiers?
"Gartner's point is that Linux on the server hasn't brought Linux to the desktop in anything like the numbers enthusiasts have claimed."
Yes, and my point was that while in theory some people are probably buying Linux desktops and installing Windows, every single desktop Linux machine I've ever seen has been running on a PC that was sold with a copy of Windows.
So, if Gartner is going to issue a report attempting to compensate for post-purchase OS installs in order to determine the true market share for Windows and Linux, they should cover all factors, or they're providing an unbalanced report.
"There are two main reasons for the recovery disk both aimed at newbies:..."
Actually, the explanation provided to me by both eMachines and Compaq as to why they don't ship an installable version of Windows on their consumer computers is that MS has priced XP so high compared to the "restore CD" (I've been told $30 vs. $90), so in order to compete in the $500 PC market, they simply can't provide an installable copy of XP, because it would make the PC cost $1-200 more, which would wipe out their sales, because consumers don't realize that they can't reinstall the OS until long after the sale, when they can't do anything about it. It's clever of MS, because they make a lot more money selling a $30 restore CD license + a copy at full retail price. Of course, they alienate their customers, but since their customers have no alternative, it doesn't matter if they're pissed off.
"Gartner's making a bold prediction that the number of machines sold as Linux desktops may eclipse the number of machines actually running Linux."
While I'm would expect that somewhere there are plenty PC's being sold with Linux pre-installed that get wiped and have a pirated copy of Windows installed, my personal experience is the opposite -- I have run hundreds of Linux machines (server farms, at home,at work, etc.), and aside from rack-mounted servers the only practical option is to purchase a PC with Windows, then wipe it and install Linux. In theory you can buy a PC in the US with Linux installed, but in practice, nobody stocks them, and it's easier to get a Windows PC now than to special order a Linux PC to arrive eventually, and do the install yourself.
So, while some percentage of the small number of PC's sold with Linux on them may be converted to run Windows, certainly a percentage of the very large number of PC's sold with Windows on them are converted to run Linux, and in my experience the numbers lean strongly towards the latter case.
On top of this, I would argue that the number of copies of Windows sold (irrespective of Linux) is artificially inflated by the pre-installed copies in other ways:
With consumer PC's you almost always need to buy a "real" copy of Windows, because the pre-installed copies don't come with install CD's, or even the right to make your install CD's. So if you buy a cheap PC and _anything_ happens to it that would cause you to need to reinstall (like, say, owning the PC for six months), the only (legal) option is to run a "restore" that wipes your hard drive and restores it to factory state.
On corporate desktops, if you by PC's with Windows installed, and then wipe the drive and install a standard disk image (which most companies do, to simplify management) MS insists that you need to buy a new Windows license, because the copy in the disk image is a new copy.
If you donate a used Windows PC to a school or church, MS tells them that it's illegal to use the copy of Windows on the PC unless it's accompanies by the original certificate of authenticity, and that otherwise they must by a new copy of Windows (which would often cost more than the PC itself is worth, and wouldn't run on older PC's in any case), and that without that, they must trash the PC's.
So if Gartner is trying to correct for artificial distortions on the sales numbers to determine true numbers of users, I think that they have some more work to do.
"We install our JRE such that it's only used by our application.
This works fine and dandy if you are actually shipping an app to run stand-alone, most are however expected to function within a web browser"
Not true. It's easy to write your web pages with object tags for the applets that will cause the appropriate JVM to be used, or trigger the install (as an ActiveX control, etc.) if necessary. This solves both the versioning problem and the deployment problem, which is mighty cool.
But you're right -- if you want to write software that runs perfectly across any JVM in any browser in any OS, it's a lot of work. If it's any consolation, it's less work than writing software that runs properly in any OS without the JVM.:-) And don't get me started on what it takes to get any complex JavaScript to work properly across all versions of all browsers on all operating systems. Creating a runtime environment that looks great on all OS's is very hard, and no truly cross-platform app will look as good as a native app, but (IMO) Java is closer than any other truly cross-platform approach.
The more I use Java app's, the less I think it's Java that makes most of them suck -- I think it's Swing and ATW. SWT works great -- look at Eclipse! It gives you nice, native widgets used naturally from Java app's.
About half of C#'s "advantages" are capabilities that the designers of Java considered and rejected as being a bad idea (by making the runtime less stable and secure, unnecessarily complex, etc. -- I knew Guy Steele at the time, and he certainly was aware of operator overloading, generics, etc., and made the (smart, IMO) decision that it was more important to make the language,compiler and JVM (and security model) simpler, smaller and more predictable). You might as well argue that multiple inheritence is "better" than single inheritence; it's certainly useful at times, but it makes it much harder to read code and know what it does, which in the long run is more important than solving some fairly obscure modeling scenarios. Certainly the way that C# deals with native code is nicer than Java, and the Java GUI frameworks aren't great (though there are multiple, competing GUI frameworks to choose from, and SWT is great). But C#'s advantages are, as you point out, not a matter of any novel ideas, just making some incremental improvements and different mistakes, so overall (IMO) it's not clear to me that it's really "better" than Java, just very similar with some different, debatable design decisions.
"As for "industry support", you have got to be joking."
By industry support, I mean that there are many more Java programmers than C# programmers, many more companies and open source projects providing Java resources (libraries, app servers, etc.) than C#, etc. Yes, MS is big, but are they bigger than IBM, Sun, BEA, etc., not to mention the entire Java development community? Yes, MS can through sheer muscle get their VM distributed onto PC's, so there's no doubt that eventually they'll be able to drive a chunk of the world to use C# (i.e. it'll be easy to get some people who use only MS tools to use MS's new tool), but Sun's got JVM's made nearly universal on cell phones, preinstalled on millions of PC's, etc.,... so while MS does have powerful marketing muscle, pretty much everyone else in big the tech space committed to Java years ago, and has been investing $billions in making sure that they're ahead of MS.
And we can look at SourceForge as a rough measure of Developer support. Of the roughly 90K projects, 12,784 are in Java, and 1,824 are in C#. So after a few years of a massive PR push, C# is at 14% of Java's adoption. Not to be sneezed at, but that puts it as a contender to Delphi (1,557) and JavaScript (1,932), which are relatively obscure as application languages. At least it's more popular than TCL (816) or Prolog (89).
Or let's look at the number of app servers available. How many companies provide J2EE implementations are available? See http://www.theserverside.com/reviews/matrix.tss -- they list 30+ companies. How many companies provide.Net app servers? I can't find a list, but I only know of one.:-)
So I guess what you're arguing is that with MS behind it, C# is going to get lots of developer support, and will mature, eventually. That's probably true, since MS is unlikely to cancel C# (more embarassing to kill C# than Bob...). But it hardly shows that C# has more industry support than Java.
"When they make the CLR come pre-installed on Longhorn, it will instantly become an industry standard"
If the CLR's penetration is driven by Longhorn, that means that it won't be something developers can assume until when, 2009? Look at XP -- completely anemic takeup rates, essentially zero consumer upgrades, so it's only being deployed as new system purchases, which means something like a five year replacement rate. In the corporate world, Windows 2000 is the standard platform -- upgrading to XP is messy enough to deal with that most large companies are still forcing vendors to ship new PC's with Win2000 to keep XP out of their environment to maintain a standard, supported platform. So
I'm sure the book is nicely written and all, but I'm still mystified by why anyone would care about C# or Mono. It all reminds me of when Intel announced the Itanium and scared half of the competing CPU manufacturers to kill off their own products. As a business strategy, it hardly matters that the Itanium has failed completely in the marketplace, since the competition largely committed suicide years ago, so Intel wins by default. Imaging how cool the Alpha, MIPS, etc., might have been by now if they'd continued active development...
So, back to Mono -- why on earth should anyone care that there's something that's almost Java, only without anywhere near as much industry support, and many years less maturity? Sure, MS is scary and all, but so was Intel... now, if C# were substantially better than Java, it would at least be technically interesting, but so far the best argument I've heard for using C# instead of Java is that MS is promoting it. I can see why that would be important in an MS-only shop, but why on earth would anyone interested in cross-platform or open software care about, much less promote, Mono? Is it some twisted attempt to take control of C# from MS?
"I love the Daily Show, but if that is what passes for requiring thought then I feel really bad about the mental state of our society."
I disagree. The Daily Show's material is consistently more challenging (and, IMO, rewarding) than Letterman or Leno. If you read the report (cheating, I know) you'll see that the Daily Show's jokes were much longer, and with more substantial topics, than the others.
"Plenty of people have come along and said "Unix is a stupid design. Here is a radical alternative" and then the users have said "But this radical alternative is a broken pile of crap", and the designer says "Bah, you're just all part of the Unix conspiracy".
You've got the sequence wrong. Multics was designed and written _before_ UNIX. And I am certainly not saying that UNIX is a stupid design, just that there are fundamental design decisions in UNIX (to lose most of Multics' functionality) that made sense given the hardware costs of the day (which is why UNIX is everywhere, and Multics is dead), but which aren't as appropriate now that hardware is astoundingly fast and cheap.
This all reminds me of (for example) the "evolution" of server OS's (mainframes, UNIX) to desktops. DOS lacked almost everything (virtual memory, multi-user support, etc.) but it fit on (relatively) cheap PC's, so it became immensely popular. And as DOS has grown up, the DOS community has gradually rediscovered many of the things that were initially carved out, but it's been a long, painful process -- every time you change a fundamental design decision in the OS, tons of app's break or you have to waste near-infinite resources emulating the stupid model in parallel with the good model (how long did it take NT to actually run Windows app's properly...). So while you can certainly say that DOS was more successful than, say, VMS (way more installs...), it'd be hard to argue that it was a better OS.
"it would be pretty simple, actually. The RAW format just needs a header which describes the data it contains."
The problem is that the fundamental data isn't standardized -- the RAW data is the signal straight from the CCD, and that's very different depending on the specifics of the CCD. For example, some high-end cameras have three separate CCD's (much, much clearer image, particularly in low light), Fuji CCD's are just weird, etc. So the RAW formats are all very proprietary, not only per manufacturer, but even for each specific camera model.
So TIFF is a good starting point (it's very flexible), but all of the details would have to be defined.
" My parents, in a vain attempt to rid the basement of its malodorous "twang" purchased a dehumidifier which, because there was no electrical outlet anywhere near the floor drain, required emptying on a daily basis.
While it's a good idea to have a firewall block malformed packets, etc., this is pretty close to useless in defending from disgruntled ex-employees. Most security violations take place from inside the firewall, because the disgruntled ex-employees know passwords, have friends back in the company, etc. And most attacks that do come from outside the company aren't at the level of malformed packets -- they're application-level vulnerabilities, or simply people that try default passwords, etc.
It's not a bad idea to have a Firewall, because it's easy, but it certainly doesn't make you secure.
"And not to make light of your accomplishments, but how fucking tough could this be?"
It's actually pretty hard to design a good voting system, because there are many tricky constraints. For example, you have to make a system with a physical audit trail, but which doesn't allow someone to determine the order that the votes were cast (because someone could easily record the order in which people voted, and match the two lists up to know how people voted).
"Seems like they want big holes in their security, doesn't it?"
Yes, it's hard to believe some of the mistakes that the voting system vendors have made. My favorite is the one that ADVERTISED AS A FEATURE that they had WiFi in their voting stations in order to make vote collection easier. Think of how convenient it would be to vote from the parking lot...
"hard disks are fundamentally thousands of times slower than RAM - you don't want programmers treating it the same way."
So, you don't use virtual memory? You're a stud!
Personally, I like it when the OS manages details so that I don't have to. And if I didn't have to waste time writing code to write my data structures out to a sequential text file and read it back, I'd be a lot happer. As I mentioned before, the NewtonOS handled this issue properly, and the result was that you could write an app for the Newton in 1/2 the time of more traditional (file-based) operating systems, because you'd only have to write 1/2 as much code.
As a programmer, I have to say that it's way easier to interact with your data as native data structures, and let the OS manage persistence. SmallTalk and NewtonOS both managed this properly, and are still the best development environments around.
And yes, I've also worked with people who insisted that virtual memory was a terrible idea because the programmer lost too much control over system performance by allowing the OS to manage swapping things in and out of RAM. Of course, given modern hardware, they were wrong -- the OS can provide a finely-tuned virtual memory subsystem that will outperform any swapping code any engineer is likely to write for a single application.:-)
"You make the system double-blind accountable, so that you don't *have* to trust anyone. If trust of a single individual makes or breaks your security model, it is broken, period."
Exactly. In a well designed election system, you don't have to trust _any_ of the participants. That's why most manual recounts take place with an observer from each relevant party, so that _any_ of them can challenge anything that looks suspicious. So, you don't have to trust anyone to be honest -- you just have to trust them to hate each other enough to keep the other guy from getting away with anything.
"I think the system the grandparent was promoting was using public funds to create a public solution, which still requires buying/paying for tools from the private sector"
Exactly. Please visit http://www.openvotingconsortion.org/. We're a consortium dedicated to creating an open source voting system. The idea, exactly as you propose, is that many commercial vendors can take the open source platform and package it with hardware, training, and so on. Or a particularly motivated (or cheap) organization could run their own election system using internal technical resources.:-) The project has been under active development for several years, and has produced a system that's been publicly demonstrated.
Having worked with people who used Multics (I'm not quite that old, myself) I'd have to disagree. UNIX isn't a "better" OS than MULTICS -- in fact, it was intentionally designed to be a "worse" OS (in being far less sophisticated, and providing far less functionality), but had the pragmatic advantage of running on more widely available, less expensive hardware.
The shame is that the tradeoffs that made UNIX a success in the 70's are probably not valid any more, and modern operating systems and programming models are still (IMO) largely trapped by design decisions made by UNIX. For example, the artificial separation between how you access RAM and disk costs developers untold amounts of wasted effort. And the idea that you don't have to turn a computer off while adding or removing CPU's, RAM, disks, controllers, etc., is still quite uncommon. And the way all linking was dynamic is way better than the mess we have to deal with now. And it had a very nice multi-processor architecture, where all CPU's had access to all RAM, and through that to all secondary storage, etc.
Of course, some of these ideas have reemerged in high-end computing (the multicians, see http://www.multicians.org) are still all hard at work making things better.:-) But it's hard to correct for fundamental mistakes (e.g. having to flatten data out into files to make it persistent) without forcing all software to be rewritten. As far as I know, the closest OS to this clean model was NewtonOS...
"Having coded on Windows and Unix boxes for many years, I recently had to do some development work on a Mac. Trying to figure out how do do things via the Mac's GUI was brutal."
Yes, but we weren't talking about developers being forced to use MacOS X, but having the option to. Right now, as I read it, MacOS X developers who entered the contest would have to use either Linux or NT, which, as you point out, would be a somewhat painful transition (no Cocoa, no Carbon, no Objective C (OK, gcc does compile objective C, but no IDE support, etc.), etc.
Actually, for a UNIX developer, MacOS X development is quite painless -- a few minor path differences, comparable to the differences between, say, a Sun and an SGI, and you're there. But I wouldn't suggest that anyone should switch from their preferred platform for a programming contest, since they'd waste time learning the new platform instead of programming, which sounds like a great way to lose.
Slashdot Mirror
"Aren't we talking about the high level here? We aren't talking about which one has a better core but the user experience."
User experience is all about getting the details right. In terms of high level architecture you could argue that there's no difference between XP or MacOS X (or KDE or Gnome), because they're all a WIMP (windows, icons, mouse) GUI over a multi-tasking OS. But the details are very different between them, and those details matter to users. For example, look at the OS install/registration process. Installing MacOS X requires few clicks, and setup and registration is optional. Installing XP is quite involved, and registration is not only required, it's enforced by a central server, and re-registration is required if you make any substantial system configuration changes.
That is a nice blade sharpener.
:-)
I think that he's right that MacOS X is too complex to be a simple appliance. But I think that general purpose computers are by definition complex, because they can be used for *anything*, and his vision holds more true for specialized devices. For example, the iPod is elegant and transparent to use.
That being said, I'm sure that usability could always be improved. But I don't agree that there's not much difference between XP and MacOS X -- while they're similar at a very high level (mouse/windows/icons over multi-tasking OS, etc.), MacOS X is better in almost every detail. But it's best not to get into a religious war here. I can only guess that Jeff has such a radical vision for how computers could be that from his perspective XP and MacOS X aren't too different.
Hmm, kinda like Nader!
Actually, MacOS X has always (since it was OPENSTEP, then Rhapsody) run on x86. Apple shipped (to developers) Rhapsody for x86, and it worked fine (if you had hardware that they had drivers for). And I've been assured by Apple systems software people that they're still maintaining the x86 build (and the other CPU's that OPENSTEP ran on) in order to make sure that Apple doesn't accidentally break portability. This applies to Cocoa app's, but not Carbon (i.e. old MacOS).
"Maybe I'm missing something, but what about the legality of that site (allofmp3.com)? "
It's not legal. Despite the claim on their web site, they're:
- Selling music digitally that the artists have never agreed to sell digitally (e.g. Led Zeppelin, The Beatles).
- They've never (AFAIK) actually paid any royalties to any musicians, composers or music companies.
So while the technology of the site is impressive, as is often the case, this deal that looks "too good to be true" is (IMO) a scam, in this case directly ripping off artists, composers and music companies. That's (IMO) why PayPal stopped processing their payments.
"They used off the shelf parts to give XBox short time to market in exchange for their own fab facilities and eventually being able to break even on hardware"
IMO, the real mistakes that MS made with the Xbox are the same mistakes they've made every time they launch (and fail with) any product outside of the PC desktop. Those mistakes are:
1) They design the product that they want to sell, not what people want to buy. That is, they start with the assumption that they're going to apply Windows into new markets and extract more revenue, instead of objectively determining what customers want and providing the best possible solution. This is why WinCE has done terribly against PalmOS, MSN has lost a fortune as an ISP, XP's adoption rate has been miserable, their several generations of pen-based products have failed, their Windows-based set-top box failed, etc.
2) They think that they can win with with more "features", so their products are overdesigned, overly complex, and require too much hardware. This works when selling desktop software, because reviewers love making checklists of features, and inefficiency and instability are relatively accepted, but fails when other factors are more important. For example, NT has done badly (compared to projections) as a server OS because it's loaded with "features" that are useless in a server farm, but which increase software complexity, reducing efficiency and security. It's typical of this issue that when MS launches each new version of NT, they proudly announce how many hundreds of thousands of lines of new code it contains, and how many nifty new features, which is the _opposite_ of what any experienced system administrator wants to hear about an OS update. New lines = more bugs, and more weird things to learn how to deal with. All someone running a server farm wants to hear about an OS release is that it's exactly the same as the previous revision but has higher performance and fixes bugs. Ditto embedded OS's (where MS has fallen flat). Similarly, WinCE was far too complex when it launched, so that it had more "features" than PalmOS, but was less usable, and was only supported on extremely expensive, cumbersome hardware with short battery life.
3) They overvalued Windows API's and integration. Since those are the key to their desktop success, they think that they're important elsewhere. But in a PDA, you have to fundamentally rewrite your app's already (to use the touchscreen, to make sense on a much smaller screen, etc.) so it's not that hard to use a new app framework, particularly if the new framework is (like PalmOS) simpler to use. And on servers, you don't want to run Windows desktop app's, but specialized server app's, so again there's less of an overlap than you would think. And if you compare WinCE and native games written for the Sega Dreamcast, you'll notice that the non-Windows games ran better, and sold better, than the PC ports. So while in theory the shared API with the PC might be an advantage, in practice (on the only platform to offer both options) the Windows API's had negative value.
4) They're far too aggressive in their business dealings. Everyone who partners with MS is attracted by the potential of making a fortune through MS's distribution channels, and loses a fortune on the details of the contracts. And since MS designs the technology to lock partners into them, nobody wants to even use their technology. For example, look at MS's networked game architecture -- it requires you to write servers in NT that run in MS' infrastructure, and to use MS services for everything. Some of these are conveient for users, but all of it forces game companies to be dependent on MS to do business, which means that MS can prevent games from shipping, or even shut systems down, with nothing that the game company can do about it but ask politely (or, I suppose, sue). So when MS says "NO XBOX SOFTWARE PRODUCT(S) MAY BE PUBLISHED, OR DISTRIBUTED TO END USERS, EXCEPT BY A LICENSED PUBLISHER PURSUANT TO AN XBOX DEVELOPMENT KIT AND X
"Israel is building walled and ever shrinking ghettos in to which they are shoving all the Arab's in their borders"
Dramatic, but exactly wrong. Israel is walling itself _in_ to keep out the non-Israeli Arabs that keep trying to sneak into Israel to kill them.
If you're really concerned about the plight of the Arabs, you should probably take that up with the Arab governments. It's astonishing to me that Arab Israeli's are so much better off than the Arab non-Israeli's -- they're better off economically, live in a democracy where they have a meaningful vote, have free speech, etc., which is more than you can say for them in, say, Saudi Arabia, Egypt, or Pakistan. If the Arab world really cared about the Palestinians they probably wouldn't have kicked them out of every Arab country and forced them into UN camps.
"If you look at all of his positions on Iraq he has all bases covered. He voted for the war, but now he is against it..."
If you read what he said, rather than absurd snippets extracted to mislead rather than inform, you'll see that Kerry has had a consistent position on the Iraq war. He voted to give the President the authority to use military force as leverage in negotiations, and to use it only as a last resort. So despite the fact that inspections were accurately determining that Iraq had no weapons of mass destruction, Bush insisted on going to war. Kerry supported having Iraq pay for their own reconstruction (i.e. from oil revenue), while Bush insisted that the US taxpayers cover the entire expense. Bush threatened to veto Kerry's bill, and then Kerry voted against Bush's bill. If Bush had been forced to actually cast the veto, would you have argued that Bush opposed funding the soldiers? Yes, back when Kerry believed Bush, Rice and Powell's lies about the evidence supporting their claims that Iraq was behind 9/11, Iraq had WMD's and Iraq being an immanent threat, he supported invading. Once it became clear that in fact Iraq had nothing to do with 9/11, had no WMD's, and was not a threat to the US, he opposed the war.
You should be asking yourself why Bush still supports the war when all of the justifications that he presented before invading turned out to be erroneous.
"If you look at all of his positions on Iraq he has all bases covered. He voted for the war, but now he is against it..."
If you read what he said, rather than absurd snippets extracted to mislead rather than inform, you'll see that Kerry has had a consistent position on the Iraq war. He voted to give the President the authority to use military force as leverage in negotiations, and to use it only as a last resort. So despite the fact that inspections were accurately determining that Iraq had no weapons of mass destruction, Bush insisted on going to war, which Kerry opposed. Kerry supported having Iraq pay for their own reconstruction (i.e. from oil revenur), while Bush insisted that the US taxpayers cover the entire expense. Bush threatened to veto Kerry's bill, and then Kerry voted against Bush's bill. If Bush had been forced to actually cast the veto, would you have argued that Bush opposed funding the soldiers?
"Gartner's point is that Linux on the server hasn't brought Linux to the desktop in anything like the numbers enthusiasts have claimed."
Yes, and my point was that while in theory some people are probably buying Linux desktops and installing Windows, every single desktop Linux machine I've ever seen has been running on a PC that was sold with a copy of Windows.
So, if Gartner is going to issue a report attempting to compensate for post-purchase OS installs in order to determine the true market share for Windows and Linux, they should cover all factors, or they're providing an unbalanced report.
"There are two main reasons for the recovery disk both aimed at newbies: ..."
Actually, the explanation provided to me by both eMachines and Compaq as to why they don't ship an installable version of Windows on their consumer computers is that MS has priced XP so high compared to the "restore CD" (I've been told $30 vs. $90), so in order to compete in the $500 PC market, they simply can't provide an installable copy of XP, because it would make the PC cost $1-200 more, which would wipe out their sales, because consumers don't realize that they can't reinstall the OS until long after the sale, when they can't do anything about it. It's clever of MS, because they make a lot more money selling a $30 restore CD license + a copy at full retail price. Of course, they alienate their customers, but since their customers have no alternative, it doesn't matter if they're pissed off.
"Gartner's making a bold prediction that the number of machines sold as Linux desktops may eclipse the number of machines actually running Linux."
While I'm would expect that somewhere there are plenty PC's being sold with Linux pre-installed that get wiped and have a pirated copy of Windows installed, my personal experience is the opposite -- I have run hundreds of Linux machines (server farms, at home,at work, etc.), and aside from rack-mounted servers the only practical option is to purchase a PC with Windows, then wipe it and install Linux. In theory you can buy a PC in the US with Linux installed, but in practice, nobody stocks them, and it's easier to get a Windows PC now than to special order a Linux PC to arrive eventually, and do the install yourself.
So, while some percentage of the small number of PC's sold with Linux on them may be converted to run Windows, certainly a percentage of the very large number of PC's sold with Windows on them are converted to run Linux, and in my experience the numbers lean strongly towards the latter case.
On top of this, I would argue that the number of copies of Windows sold (irrespective of Linux) is artificially inflated by the pre-installed copies in other ways:
With consumer PC's you almost always need to buy a "real" copy of Windows, because the pre-installed copies don't come with install CD's, or even the right to make your install CD's. So if you buy a cheap PC and _anything_ happens to it that would cause you to need to reinstall (like, say, owning the PC for six months), the only (legal) option is to run a "restore" that wipes your hard drive and restores it to factory state.
On corporate desktops, if you by PC's with Windows installed, and then wipe the drive and install a standard disk image (which most companies do, to simplify management) MS insists that you need to buy a new Windows license, because the copy in the disk image is a new copy.
If you donate a used Windows PC to a school or church, MS tells them that it's illegal to use the copy of Windows on the PC unless it's accompanies by the original certificate of authenticity, and that otherwise they must by a new copy of Windows (which would often cost more than the PC itself is worth, and wouldn't run on older PC's in any case), and that without that, they must trash the PC's.
So if Gartner is trying to correct for artificial distortions on the sales numbers to determine true numbers of users, I think that they have some more work to do.
"We install our JRE such that it's only used by our application.
:-) And don't get me started on what it takes to get any complex JavaScript to work properly across all versions of all browsers on all operating systems. Creating a runtime environment that looks great on all OS's is very hard, and no truly cross-platform app will look as good as a native app, but (IMO) Java is closer than any other truly cross-platform approach.
This works fine and dandy if you are actually shipping an app to run stand-alone, most are however expected to function within a web browser"
Not true. It's easy to write your web pages with object tags for the applets that will cause the appropriate JVM to be used, or trigger the install (as an ActiveX control, etc.) if necessary. This solves both the versioning problem and the deployment problem, which is mighty cool.
But you're right -- if you want to write software that runs perfectly across any JVM in any browser in any OS, it's a lot of work. If it's any consolation, it's less work than writing software that runs properly in any OS without the JVM.
The more I use Java app's, the less I think it's Java that makes most of them suck -- I think it's Swing and ATW. SWT works great -- look at Eclipse! It gives you nice, native widgets used naturally from Java app's.
"C# is better than Java in almost every way:"
... so while MS does have powerful marketing muscle, pretty much everyone else in big the tech space committed to Java years ago, and has been investing $billions in making sure that they're ahead of MS.
.Net app servers? I can't find a list, but I only know of one. :-)
About half of C#'s "advantages" are capabilities that the designers of Java considered and rejected as being a bad idea (by making the runtime less stable and secure, unnecessarily complex, etc. -- I knew Guy Steele at the time, and he certainly was aware of operator overloading, generics, etc., and made the (smart, IMO) decision that it was more important to make the language,compiler and JVM (and security model) simpler, smaller and more predictable). You might as well argue that multiple inheritence is "better" than single inheritence; it's certainly useful at times, but it makes it much harder to read code and know what it does, which in the long run is more important than solving some fairly obscure modeling scenarios. Certainly the way that C# deals with native code is nicer than Java, and the Java GUI frameworks aren't great (though there are multiple, competing GUI frameworks to choose from, and SWT is great). But C#'s advantages are, as you point out, not a matter of any novel ideas, just making some incremental improvements and different mistakes, so overall (IMO) it's not clear to me that it's really "better" than Java, just very similar with some different, debatable design decisions.
"As for "industry support", you have got to be joking."
By industry support, I mean that there are many more Java programmers than C# programmers, many more companies and open source projects providing Java resources (libraries, app servers, etc.) than C#, etc. Yes, MS is big, but are they bigger than IBM, Sun, BEA, etc., not to mention the entire Java development community? Yes, MS can through sheer muscle get their VM distributed onto PC's, so there's no doubt that eventually they'll be able to drive a chunk of the world to use C# (i.e. it'll be easy to get some people who use only MS tools to use MS's new tool), but Sun's got JVM's made nearly universal on cell phones, preinstalled on millions of PC's, etc.,
And we can look at SourceForge as a rough measure of Developer support. Of the roughly 90K projects, 12,784 are in Java, and 1,824 are in C#. So after a few years of a massive PR push, C# is at 14% of Java's adoption. Not to be sneezed at, but that puts it as a contender to Delphi (1,557) and JavaScript (1,932), which are relatively obscure as application languages. At least it's more popular than TCL (816) or Prolog (89).
Or let's look at the number of app servers available. How many companies provide J2EE implementations are available? See http://www.theserverside.com/reviews/matrix.tss -- they list 30+ companies. How many companies provide
So I guess what you're arguing is that with MS behind it, C# is going to get lots of developer support, and will mature, eventually. That's probably true, since MS is unlikely to cancel C# (more embarassing to kill C# than Bob...). But it hardly shows that C# has more industry support than Java.
"When they make the CLR come pre-installed on Longhorn, it will instantly become an industry standard"
If the CLR's penetration is driven by Longhorn, that means that it won't be something developers can assume until when, 2009? Look at XP -- completely anemic takeup rates, essentially zero consumer upgrades, so it's only being deployed as new system purchases, which means something like a five year replacement rate. In the corporate world, Windows 2000 is the standard platform -- upgrading to XP is messy enough to deal with that most large companies are still forcing vendors to ship new PC's with Win2000 to keep XP out of their environment to maintain a standard, supported platform. So
I'm sure the book is nicely written and all, but I'm still mystified by why anyone would care about C# or Mono. It all reminds me of when Intel announced the Itanium and scared half of the competing CPU manufacturers to kill off their own products. As a business strategy, it hardly matters that the Itanium has failed completely in the marketplace, since the competition largely committed suicide years ago, so Intel wins by default. Imaging how cool the Alpha, MIPS, etc., might have been by now if they'd continued active development...
So, back to Mono -- why on earth should anyone care that there's something that's almost Java, only without anywhere near as much industry support, and many years less maturity? Sure, MS is scary and all, but so was Intel... now, if C# were substantially better than Java, it would at least be technically interesting, but so far the best argument I've heard for using C# instead of Java is that MS is promoting it. I can see why that would be important in an MS-only shop, but why on earth would anyone interested in cross-platform or open software care about, much less promote, Mono? Is it some twisted attempt to take control of C# from MS?
"I love the Daily Show, but if that is what passes for requiring thought then I feel really bad about the mental state of our society."
I disagree. The Daily Show's material is consistently more challenging (and, IMO, rewarding) than Letterman or Leno. If you read the report (cheating, I know) you'll see that the Daily Show's jokes were much longer, and with more substantial topics, than the others.
"Plenty of people have come along and said "Unix is a stupid design. Here is a radical alternative" and then the users have said "But this radical alternative is a broken pile of crap", and the designer says "Bah, you're just all part of the Unix conspiracy".
You've got the sequence wrong. Multics was designed and written _before_ UNIX. And I am certainly not saying that UNIX is a stupid design, just that there are fundamental design decisions in UNIX (to lose most of Multics' functionality) that made sense given the hardware costs of the day (which is why UNIX is everywhere, and Multics is dead), but which aren't as appropriate now that hardware is astoundingly fast and cheap.
This all reminds me of (for example) the "evolution" of server OS's (mainframes, UNIX) to desktops. DOS lacked almost everything (virtual memory, multi-user support, etc.) but it fit on (relatively) cheap PC's, so it became immensely popular. And as DOS has grown up, the DOS community has gradually rediscovered many of the things that were initially carved out, but it's been a long, painful process -- every time you change a fundamental design decision in the OS, tons of app's break or you have to waste near-infinite resources emulating the stupid model in parallel with the good model (how long did it take NT to actually run Windows app's properly...). So while you can certainly say that DOS was more successful than, say, VMS (way more installs...), it'd be hard to argue that it was a better OS.
"it would be pretty simple, actually. The RAW format just needs a header which describes the data it contains."
The problem is that the fundamental data isn't standardized -- the RAW data is the signal straight from the CCD, and that's very different depending on the specifics of the CCD. For example, some high-end cameras have three separate CCD's (much, much clearer image, particularly in low light), Fuji CCD's are just weird, etc. So the RAW formats are all very proprietary, not only per manufacturer, but even for each specific camera model.
So TIFF is a good starting point (it's very flexible), but all of the details would have to be defined.
" My parents, in a vain attempt to rid the basement of its malodorous "twang" purchased a dehumidifier which, because there was no electrical outlet anywhere near the floor drain, required emptying on a daily basis.
:-)
Uh, an extension cord perhaps?"
Or, to invert the problem, a hose?
While it's a good idea to have a firewall block malformed packets, etc., this is pretty close to useless in defending from disgruntled ex-employees. Most security violations take place from inside the firewall, because the disgruntled ex-employees know passwords, have friends back in the company, etc. And most attacks that do come from outside the company aren't at the level of malformed packets -- they're application-level vulnerabilities, or simply people that try default passwords, etc.
It's not a bad idea to have a Firewall, because it's easy, but it certainly doesn't make you secure.
"And not to make light of your accomplishments, but how fucking tough could this be?"
It's actually pretty hard to design a good voting system, because there are many tricky constraints. For example, you have to make a system with a physical audit trail, but which doesn't allow someone to determine the order that the votes were cast (because someone could easily record the order in which people voted, and match the two lists up to know how people voted).
"Seems like they want big holes in their security, doesn't it?"
Yes, it's hard to believe some of the mistakes that the voting system vendors have made. My favorite is the one that ADVERTISED AS A FEATURE that they had WiFi in their voting stations in order to make vote collection easier. Think of how convenient it would be to vote from the parking lot...
"hard disks are fundamentally thousands of times slower than RAM - you don't want programmers treating it the same way."
:-)
So, you don't use virtual memory? You're a stud!
Personally, I like it when the OS manages details so that I don't have to. And if I didn't have to waste time writing code to write my data structures out to a sequential text file and read it back, I'd be a lot happer. As I mentioned before, the NewtonOS handled this issue properly, and the result was that you could write an app for the Newton in 1/2 the time of more traditional (file-based) operating systems, because you'd only have to write 1/2 as much code.
As a programmer, I have to say that it's way easier to interact with your data as native data structures, and let the OS manage persistence. SmallTalk and NewtonOS both managed this properly, and are still the best development environments around.
And yes, I've also worked with people who insisted that virtual memory was a terrible idea because the programmer lost too much control over system performance by allowing the OS to manage swapping things in and out of RAM. Of course, given modern hardware, they were wrong -- the OS can provide a finely-tuned virtual memory subsystem that will outperform any swapping code any engineer is likely to write for a single application.
"You make the system double-blind accountable, so that you don't *have* to trust anyone. If trust of a single individual makes or breaks your security model, it is broken, period."
Exactly. In a well designed election system, you don't have to trust _any_ of the participants. That's why most manual recounts take place with an observer from each relevant party, so that _any_ of them can challenge anything that looks suspicious. So, you don't have to trust anyone to be honest -- you just have to trust them to hate each other enough to keep the other guy from getting away with anything.
"I think the system the grandparent was promoting was using public funds to create a public solution, which still requires buying/paying for tools from the private sector"
:-) The project has been under active development for several years, and has produced a system that's been publicly demonstrated.
Exactly. Please visit http://www.openvotingconsortion.org/. We're a consortium dedicated to creating an open source voting system. The idea, exactly as you propose, is that many commercial vendors can take the open source platform and package it with hardware, training, and so on. Or a particularly motivated (or cheap) organization could run their own election system using internal technical resources.
Having worked with people who used Multics (I'm not quite that old, myself) I'd have to disagree. UNIX isn't a "better" OS than MULTICS -- in fact, it was intentionally designed to be a "worse" OS (in being far less sophisticated, and providing far less functionality), but had the pragmatic advantage of running on more widely available, less expensive hardware.
:-) But it's hard to correct for fundamental mistakes (e.g. having to flatten data out into files to make it persistent) without forcing all software to be rewritten. As far as I know, the closest OS to this clean model was NewtonOS...
The shame is that the tradeoffs that made UNIX a success in the 70's are probably not valid any more, and modern operating systems and programming models are still (IMO) largely trapped by design decisions made by UNIX. For example, the artificial separation between how you access RAM and disk costs developers untold amounts of wasted effort. And the idea that you don't have to turn a computer off while adding or removing CPU's, RAM, disks, controllers, etc., is still quite uncommon. And the way all linking was dynamic is way better than the mess we have to deal with now. And it had a very nice multi-processor architecture, where all CPU's had access to all RAM, and through that to all secondary storage, etc.
Of course, some of these ideas have reemerged in high-end computing (the multicians, see http://www.multicians.org) are still all hard at work making things better.
"Having coded on Windows and Unix boxes for many years, I recently had to do some development work on a Mac. Trying to figure out how do do things via the Mac's GUI was brutal."
Yes, but we weren't talking about developers being forced to use MacOS X, but having the option to. Right now, as I read it, MacOS X developers who entered the contest would have to use either Linux or NT, which, as you point out, would be a somewhat painful transition (no Cocoa, no Carbon, no Objective C (OK, gcc does compile objective C, but no IDE support, etc.), etc.
Actually, for a UNIX developer, MacOS X development is quite painless -- a few minor path differences, comparable to the differences between, say, a Sun and an SGI, and you're there. But I wouldn't suggest that anyone should switch from their preferred platform for a programming contest, since they'd waste time learning the new platform instead of programming, which sounds like a great way to lose.