The only thing NT3.51 actually required more of was memory. I've run it with decent success on 386s.
So have I. But you're talking 4 - 6 vs 12 - 16 MB of RAM, when RAM was probably the most expensive component in the entire machine. That was a *huge* difference in hardware price ca. 1994-95. Added to that, Windows 95 was usable on 386SX machines, whereas NT 3.51 really needed a fast 386DX or slow 486.
Dell machines already cost MORE than Apple's [...]
Only in a handful of very specific configurations.
[...] so I suspect the conditions you outline are a null set.
So if you're so sure Dell can't produce machines any better than Apple's, why do you want them to license OS X ? Possibly because you know Dell would quickly move to sell machines filling the gaping holes in Apple's lineup and subsequently slaughter Apple's hardware sales ?
Kind of sounds like a history of NT, huh? But with some different processor families.
Not really. NT hasn't had the advantage of a) being initially developed without any need for legacy support and b) then having another significant working over, with little regard for legacy support, as NeXT was to get to OS X. *That* is why OS X bloody well should have brought something new to the table (which it did, with its GUI layer).
While it's certainly arguable some parts of NeXT were better than NT (display postscript, for example), its similarly arguable than at least as many parts of NT were better than NeXT (NTFS, pervasive multithreading, pervasive ACLs).
I did notice that OSX is a turd.
I think that's rather harsh. OS X is most certainly sluggish, even on high end hardware (which has brought me to the conclusion it's a low-level system problem - I used to think it was just the old G4 machines), but it has many excellent attributes, including its UI and rich APIs.
Security was an issue that had to be looked at, for sure; with sufficient review, probably not a big deal.
Well, I'd be hesistant about calling a migration from a single-user architecture to a multiuser architecture "not a big deal"...
BeOS was certainly cool. It's problem was that it wasn't better _enough_ at the things it was better at and that many of them were considered to be of less importance than the things it wasn't better at by consumers.
On the other hand, Linux puts them together perhaps most successfully.
I can't agree. From a technical perspective I find Linux to be largely a well of mediocrity (with the exception of its RAID and LVM capabilities) plagued by a seemingly unending string of rewrites of its core components, usually to bring them to the same capabilities that similar components in other OSes have had from day one (because they were actually *designed* before being implemented).
It runs on more hardware than anything but perhaps netbsd - I haven't done a comparison lately, but last time I did, netbsd was winning that particular race.
I'm not sure why people think a huge list of supported platforms is important. It's clearly a metric massively biased against commercial platforms.
And it's very stable and can be made pretty incredibly secure if you're willing to use nonstandard security frameworks.
True enough, but Windows on decent hardware is hardly _unstable_, and the vast, vast majority of security breaches don't happen due to software faults, but from user error.
I'm willing to forgive Linux and other FOSS projects for reinventing the wheel, because their reimplementations typically do things the originals can't handle.
Linux definitely handles corner cases better than Windows or OS X, but I can't say I'd agree it handles the common cases better.
On the home desktop, you could be right. I don't think you are, but you could be. It's still true that you can get a dinky low-end PC (the slowest crappiest PC you can buy new today is sufficient for 99% of the population, probably) and slap Linux on it for probably half the price of the cheapest Macintosh. Oh sure, it's a bit less powerful, but that won't even affect most people.
Certainly you can. But a hell of a lot more people use Mac Minis, MacBooks, iMacs, etc than use bottom-end PCs running Linux.
Macintosh does have something Linux doesn't, and that's marketing.
Actually, the biggest thing "Macintosh" has that "Linux" doesn't is people selling it. Plus widespread commercial support (both hardware and software), better usability and better software integration. Apple hardware also tends to be fairly solid (if somewhat more expensive) - something that can be a bit hit and miss with PCs.
If I were to think of one place money could best be spent to promote Linux, it would be advertising. Get the word out. Put out some clever TV commercials, etc.
No-one is really trying to sell Linux PCs (and make no mistake, average consumers don't buy OSes, they buy computers). Given that in the cutthroat world of PCs, the economic advantage of being able to undercut your competitors on price by ~$50 (roughly the OEM cost of Windows) is huge, I think that's pretty indicative that no-one thinks Linux is a strong competitor to either Windows or Mac in that market space.
On the corporate desktop, most companies won't even consider macintosh computers, because they can't be maintained the same way other PCs can - you just don't have access to the same commodity parts, which is mostly to say you can't just throw a new motherboard in a machine that HAS to be up today and reinstall the OS for drivers - if your OS even demands that, like Linux doesn't.
Assuming by "corporate world" you mean "big business" and not "small company", then I think any suggestions people aren't buying Macs because they can't just throw an off the shelf replacement part in is rather naive (and that's ignoring that a significant proportion of hardware in modern Macs *is* replacable with off the shelf parts). The corporate world works on 3 or 5 year warranties and *loathes* having to do such hands-on maintenance itself - particularly especially labour intensive work like changing motherboards - because it's more expensive.
(There are exceptions to this, of course, but they're few and far between.)
Macs are (relatively) unpopular in business because they're relatively expensive, because until quite recently they didn't integrate particularly well into existing infrastructure and because frighteningly large numbers of businesses rely on poorly written software, often developed in-house and/or no longer maintained, that only runs on Windows.
Obviously there are some stunning counterexamples. NeXTStep rides again. Too bad it didn't gain acceptance when nobody else had features like those.
That probably had something to do with its pricetag.
Linux can be a DC, but you're right, I wouldn't do that either.
As far as I know, Samba can't be an Active Directory Domain Controller and, to be quite frank, even if the developers said it could, it would have to demonstrate trouble-free running in such a role for a good year or two before I'd trust it as an integral piece of infrastructure (not that I have anything against the Samba developers, or question their abilities, but their entire application *is* based around reverse engineering another very large, very complex piece of software).
And that active directory shit wouldn't be necessary if it weren't a big fat windows network to begin with.
If better alternatives were widespread and cheaper, they'd be used instead. Not to mention, AD stands up quite well on its own merits.
Anyway, tt's better at being a fileserver than Windows, so I think you're on drugs t
You would still have the right to charge for software in a world without copyright.
Of course you could. You have the "right"[0] to try and charge people for anything, if you want.
It wouldn't be as practical, which is one reason why we have copyrights, but copyright is certainly not what gives you the right to sell something. Copyright gives you exclusive distribution rights. There's a difference.
Copyright - more accurately the enforcement thereof - is what gives you the *ability* (ie: makes it "practical") to charge for things that could otherwise, easily, be obtained for free (or close enough to free). Again, my only major point is that trying to equate book and software sales is disingenuous, to say the least - both with or without the existence of copyright.
[0] I really hate it when the word "right" is tossed around so casually like this, it devalues its importance.
This would make a perfect HTPC for my living room... Think of it - hang a 24" HDTV on your wall and use it 85% of the time (for news, backround tv, etc) then when it's time for a movie, drop down the projection screen and have the same 24" PC feed your HD projector... I've thought about doing this for a while with a Cinema display and an iMac, but honestly this is even slicker!
Maybe if your "living room" is only 6 feet wide...
1.66Ghz Core Duo in the low-end, 1.83Ghz Core Duo in the hi-end. No pricedrop though:(.
Personally, I consider the Mini update to be the most interesting. The bottom end Mini is now an ideal machine to buy and put Windows MCE on for a small, silent, HD-capable HTPC.
And all the people who said they would buy a machine if this was available (the specs were pretty much exactly what was asked for), suddenly clammed up, and slowly backed out the door with a myriad excuses why they suddenly had something else to do.
Your argument is disingenuous to the point of deliberate deception. There was only one major reason why people didn't buy the Cube, and that was the ridiculously high price.
Far from delivering a headless iMac (which by all rational thought, would be - at worst - no more expensive than an actual iMac), Apple delivered a machine that cost about the same (more than, IIRC) as a PowerMac of the day with less capabilities.
I think Apple learned an important lesson that day. The most vocal group of people demanding a specific product and promising to buy it will usually not actually buy what they say they want. They are just looking to get something they can't have, and when they can have it, they don't want it anymore.
Apple didn't deliver what they wanted. Nor did they ever have any intention of doing so, because Apple is well aware that a headless iMac would destroy PowerMac/Mac Pro sales.
Clearly, they don't think so, because they're an iMac buyer, not a Mac Mini buyer. What an utterly astounding revelation! It's almost like different people can have different sets of needs and values!
If the only difference between and iMac and a Mini was the screen, you might just have a point.
But it's not. The internal differences are significant.
A rather large proportion of users have been calling for a headless iMac pretty much since the first G3 iMac was released, precisely because they know it *would* be better value.
Apple won't do it, of course, because such a machine would absolutely slaughter PowerMac (/"Mac Pro") sales.
IME, the older a floppy disk is, the more reliable it will be. Certainly, the last few times I've bought (admittedly 3.5", not 5.25") floppies, it hasn't been at all unusual to have a couple not work out of the box.
For those who want reliable floppies, look for some old Windows, DOS, OS/2 or Office installation disks - practically indestructible.
I'm pretty sure all Dell would offer right now is to become another Power Computing and just drain Apple's profits. For it to be a serious offer, they would have to agree to put it on some large number of machines.
Which would just drain Apple's profits even more...
Apple make their money from hardware sales. When those sales tank because most people find they can buy a more desirable machine from Dell for less money, Apple is going to be in serious trouble, just like they were the last time they went down the OS licensing path.
Unless, of course, the OS licensing costs are high enough to cover the difference - in which case a Dell Mac and an Apple Mac are going to cost the same and no-one will buy the Dell Macs, making any such arrangement of zero interest to Dell.
Here's an idea: why don't you come up with some sort of reasonable example of how Apple might license OS X to Dell that doesn't involve either a) Apple going broken because the bulk of their hardware sales evaporate, or b) Dell machines that cost as much as Apple machines and subsequently being about as popular as a fart in an elevator.
The problem for Apple is that there are quite a few gaping holes in their lineup, precisely where Dell can put machines and large numbers of customers want to buy.
For a guy who started cursing me out for being dense, [...]
Because you *were* being dense. That, or deliberately feigning ignorance.
[...] you sure are the pot. It's called long-sleeves and slacks. As for looking not quite right, you don't know a thing about what you are talking about.
I asked *you* what it would look like, precisely because I don't know. Your answer was "if anything, too perfect". People with "perfect" skin, are not common. They will stand out. That is the only point I am making.
You imply that you are in europe, [...]
I have implied nothing about my location except that I'm not an American.
[...] go look for a tv show called "there's something about miriam" in the UK and the lawsuit it generated. Miriam's level of ability to pass was probably around the 75th percentile for that age range, and she did it in a bikini and made out with the guys.
You'll pardon me if I'm not prepared to judge the effectiveness of a trained observer based on the events in a "reality TV" show.
No, those parts are NOT. They are EXPENSIVE. Just think about it for a second, enough people to professionally evaluate every single person getting on each plane, plus back-ups for when someone is questioned multiple times. Similarly properly, or rather effectively, trained observers of xray machines are also expensive because they burn out real quick.
So we shouldn't pursue better security methods because they're not free ?
In just the US alone there are about 1.8M passengers per day. The TSA currently employes over 50,000 screeners to handle that many passengers with just cursory, 4-5 second examinations of luggage. In order to include even just 30 second interrogations of every single passenger you are going to need at least the same number of people - but you are going to have to pay them more since they will be higher skilled. Right there you've probably more than doubled the entire budget of the TSA. That isn't anyone's definition of feasible.
Pretty much every time I fly, I get *at least* a 30 second chat with an airline employee at the check-in desk, including the standard questions about who packed my bag. I'm not talking about adding (many) more people to the existing processes, I'm talking about making them better with improvements to existing resources. That whole stepping through the metal detector, being "randomly" selected and subsequently searched could use some improvements, as well.
Near as I can tell, your attitude is that we shouldn't be making any attempts to improve the current systems at all (oh, apart from making it harder to get weapons on board, because apparently where you fly it's trivial to carry an arsenal into the plane) and if anything should be making them *less* effective by not paying any attention at all to passenger attributes.
Maybe my constant quoting has gone over your head. Everytime I say "stupid, idealistic, bleeding-heart-leftists system" that's a slightly paraphrased quote from your first post.
That phrase from my first post, which you continually misquote out of context, was not referring to any profiling system at all, it was referring to the attitudes of such people, which preclude them from admitting even the slightest possibility that maybe some people are more likely to be terrorists than others.
Which is pretty obvious, when you actually read what I wrote. But you were only interested in latching on to some suitable catchphrase and beating it to death.
OBVIOUSLY if you know people would object to something then you know what they are objecting to.
Indeed. They are objecting not to a specific implementation, but to a concept - and not on rational grounds, but idealistic ones.
If, on the other hand, you think a proper profiling system would not cause such objections, then why did you bother making that statement in the first place?
Yes, companies like Dell are saying they want to ship OS X, but I doubt they would currently be willing to ship OS X as their default OS. If Vista isn't working by spring '07, they may be offering to do exactly that.
The roadblock to Dell shipping OS X machines isn't Dell, Microsoft or Vista, it's Apple.
So, I'm thinking most non-windows folks are having a good chuckle right now because I know I've had little thumbnail previews of everything in KDE for going on 3+ years at least on a sleepy old 1GHZ laptop too.
Thumbnailing in Explorer windows has been around since the first release of IE4, back in 1997. It's not a new feature to Windows.
In the Mac world, I seem to recall that working in -early- OSX versions.
And it was frighteningly buggy in them, as well. Particularly dealing with "different" codecs like DivX and friends.
All that being said, the public beta, 10.0, and 10.1 were really dog slow, and it wasn't until 10.2 that you stopped noticing the slow downs.
My 1Ghz, 768Mb RAM, 7200rpm HDD iBook has 10.4, and it's still annoyingly slow for anything non-trivial (and occasionally for trivial things).
Fortunately, I only ever do trivial things on it, so that's not a huge problem (the size, durability and battery life are more important to me), but even on fast dual-core and up Macs, I still find OS X sluggish to use under any sort of meaningful load.
I've been using OS X development and release versions regularly since it still looked like MacOS Classic, and while it has been getting better with every release, it is by no means "fast", unless you have a monstrously fast machine and relatively light application load.
BeOS didn't come to PC until like 1997 or 1998 but it at least did so. It had been around (however humbly) since 1991.
You can't really leave BeOS in if you're going to disallow NetBSD, OpenBSD and Linux for being "relatively unfinished".
I don't think a serious argument can be made for Windows 95 being superior to OS/2 3.0, except in the area of available software.
Single Input Queue.
I was an OS/2 user back in the day (migrated to NT4 Beta 2 in early 1996) and I think picking between Windows 95 and OS/2, without any historical influence, would be extrememly difficult. OS/2 technical had "better" multitasking and memory protection, but in actual usage - particularly if you were only using 32 bit applications and drivers in Windows 95 - the two of them were basically the same. Windows 95 was much nicer in terms of hardware resources and was more usable on lower end machines. I only stuck with OS/2 because I already had a significant investment in OS/2 software (NT4 - even at Beta 2 stage - was vastly superior to both in pretty much every way, which was why I switched).
This is also ignoring hardware support, something which Windows 95 had orders of magnitude more of.
Oh and, Windows 95 can't take credit for plug and play. Most of that is done by the hardware; the OS can choose to move things around to other settings, but typically it does not. The drivers are more responsible for this.
How it's done is irrelevant. Windows 95 was the first mainstream OS that supported it (and you *did* need support on the OS side as well back then).
So basically, Windows 95 was a sad, sad joke even then.
Not even close. Windows 95 did what it was designed to do exceptionally well (too well, if anything - it and Windows 98 arguably worked so well they knocked back the migration to Windows NT by several years). It allowed people to install it on reasoanbly priced hardware of the time. It gave them complete backwards compatibility with all their existing software and a substantial chunk of their existing hardware (ie: peripherals).
Nothing else even came close. OS/2 needed at least as much - more, IMHO - hardware resources and had worse legacy support. Linux wasn't even playing the same game in terms of important capabilities (software, UI, hardware support). Windows NT required *way* too much hardware grunt and had even worse legacy software capabilities than OS/2. MacOS required expensive new hardware and software. NeXT required frighteningly expensive new hardware and software.
For those who had access to the media, and had no hardware requiring Windows 95, even NT 3.51 was a better choice.
Untrue. NT 3.51 had substantially higher hardware requirements, the sucky old Windows 3.x interface and dramatically worse legacy software support.
Then along came NT4, with support for that hardware, but with less separation between memory spaces, and it was probably about half as reliable as 3.51.
But with substantially better performance. For the market space, the performance was more important. Not to mention if you stuck with the bog-standard VGA driver the stability difference was negligible.
Name one thing Linux does worse than Windows, on a totally technical basis.
Asynchronous I/O. API/ABI stability. A development process that isn't design-driven.
Personally, I'm struggling to think of anything Linux does notably *better*, on a totally technical basis. LVM and software RAID are the only things I can come up with and even the former arguably has more to do with Linux tending to be used more for lower-end hardware tha anything else (Linux's LVM, however, is excellent and unmatched in any other stock-standard OS, with the possible exception of Solaris).
Or perhaps you might convince me that OSX is not superior to Windows, again, on a technical basis.
OS X has atrocious performance. There are technical reasons, I am sure, why this is so, but Windows (and Linux) spank it in performance - *expecially* interactive GUI performance in the case of Windows (with the exception of certain eyecandy that's handled almost completely by the video card). Its scheduling isn't particularly good either - multitasking in OS X is not as good as Windows or Linux - again, especially in terms of GUI interactiveness.
OS X *is* a more advanced OS in some areas, but given how much newer it is, it would damn well want to be. Windows "XP" (ie: Windows NT) was conceived in the late '80s and first released in 1993. You'd certainly hope an OS only released in 2000 that was largely just a warmed over version of an earlier OS (NeXT) would introduce something new.
Hell, just talk me into believing that Windows is a better candidate than BeOS for the user desktop, even in its current state, and I'll be amazed.
BeOS was only single user, therefore less secure. This is before we even get into things like little hardware support, a questionable network stack and it's poor handling of higher end machines (eg: large memory hacks). BeOS's widespread use of multithreading (which Windows also has) was about its most technologically interesting feature. It's kind of hard to make any meaningful comparison, though, since BeOS was never really finished and hasn't been under worthwhile active development for a very long time.
The simple truth is that Windows is technologically retarded compared either to OSX or Linux, its two most significant competitors.
The simple truth is that all contemporary OSes are largely a wash, technologically speaking. Linux is probably the *least* interesting of the bunch, since most of its features were already present in its competitors long, long before they appeared in it (eg: the O(1) scheduler that was such a big milestone, Windows NT had since its first release) and "new" features tend to be reimplementations after they appear on other platforms.
Basically your argument is what windows users use when Linux is eating windows' lunch in the server market. which, by the way, it is doing. Only Linux and Windows are gaining market share in this space now (they seem to trade off gaining ground here and there, but that could simply be due to the release of various studies at various times.)
The difference is Windows is taking either new marketshare, or from existing non-Windows installs, whereas Linux is largely just displacing legacy unix installs.
I'm highly sceptical of any assertions Linux is "eating Windows's lunch" in the server world - the evidence simply doesn't bear it out.
I also have to agree with the GP. By and large, Windows and Linux are not competing in the same market spaces (there are some crossovers, eg: webservers, but not that many). On the Desktop it's not even worth talking about - OS X is the real competition to Windows there. On the server side, most people who are after a Windows server, are after it because of functionality Linux can't provide as well (if at all). Similarly, for people actively looking at Linux as a server OS, it's unlikely they'll be considering Windows, usually due to an existing reliance on a unix-like systems (be it software or skill sets), or the higher licensing costs (although that's almost always false economy).
We use both Windows and Linux servers, and the Linux machines vastly outnumber the Windows ones. However, where we do use Windows, Linux is simply not an option - domain controllers, fileserving, groupware, windows-only software. Similarly, where we use Linux - mostly for in-house applications - Windows is a relatively poor choice due to its lack of easy customisability.
I can't think of a single Windows server we have that I would want to replace with a Linux server (and, believe, me, I've considered and even trialled it many times). Likewise, I can't think of any of our Linux servers that would be better off running Windows.
But it is. You fly 10 times, and the chance that you're IN profile but NOT ever searched drops to near nil.
Unless, of course, one of the criteria is "more than ten flights in 3 months", or something similar, and they get checked the eleventh time.
For profiling criteria NOT to be easy to deduce, every now and then the officers would have to NOT check a profiled person. Which defies common sense even more!
From what I can gather, the current "profiling" system used in US airports is utterly worthless (for starters, it's trivial for passengers to determine if they're being searched randomly or because they were flagged). Using it as an example of why profiling in general wouldn't work is, then, not a particularly strong argument.
A woman with model quality skin is far from being a prime terrorist candidate. Perfect skin is easily concealed with light make-up.
Even on the arms and legs ?
My point here is that a "woman" who already doesn't quite "look right" (like the vast majority of cross-dressers) will generate more interest the more things about her that are "weird".
That level of profiling, as fully distinct from the "stupid, idealistic, bleeding-heart-leftists" system that you proposed, [...]
I proposed no system. I gave a throwaway example of two different types of travellers.
[...] is not feasible with the kind of limited resources available.
No, but *parts* of it are. Using properly trained staff to ask pointed questions at check-in, having a properly trained observer as part of the metal detector/security gateway, etc.
Well, excuse me for going by what you write and not having ESP.
I challenge you to quote any part of any of my posts even suggesting - let alone stating - my examples were to be considered representative of a "proper profiling system".
My position has been clearly stated, and regularly illustrated, from the start - you pick characteristics that can be altered by the people you are trying to identify and they will alter those characteristics - thus costing you resources and reducing your security. You just keep jumping from characteristic to characteristic and seem to hope that by combining them into a "system" that somehow they won't be figured out.
Your "position" is that profiling won't work, based on the obviously flawed system apparently currently being used in the US.
If I were to "advocate" a solution to the problem, I would pick things that the enemy can not change, or in changing them he neutralizes the threat himself. For example, better automated screening systems - the harder it is to take a weapon on board, the less likely a person is going to do so - regardless of race, religion, family life or anything else directly under the enemy's control.
Getting weapons onto flights is already quite difficult (or at least it is where I fly). Not to mention, a weapon on a plane isn't anywhere near as useful as it was before 9/11 - no plane full of passengers and crew is just going to let a bunch of people take their aircraft over again. The "get a few people with knives on board and take the plane over" plan is vastly less feasible than it used to be.
Most airline security is a placebo at best - even the new measures restricting liquids and gels are pretty much worthless, because if someone really wants to get some liquid explosives onto a plane, they'll just put them into a plastic cigar tube or film container and shove it up their arse. The benefit of it lies not in the probability of catching someone in the process of smuggling a bomb on board, but to force them to spend non-trivial amounts of time planning a mission, allowing regular intelligence and police forces to identify, locate and apprehend them beforehand.
It wasn't an analogy. The person I was responding to wrote a book, and I don't think anyone in this thread or even the original topic was talking specifically about software. Incidentally, since his book is in PDF, it would be just as easy to reproduce and distribute his book as it would be to reproduce and distribute software.
You said it wasn't copyright that allowed someone to "charge for stuff", and used the example of a bookshop selling books to support your argument. My point is this conclusion is wrong, because someone selling physical books is a vastly different prospect to someone trying to sell software, in a world without copyright.
No, you could still easily charge for stuff without copyright. Bookstores still charge for Shakespeare books, even though there's no copyright on that. What copyright does is prevent other publishers from taking your work and selling it without your permission or giving you any money. Shakespeare, being dead, doesn't mind, but you might.
Much as I dislike copyright, your analogy is broken. Books are quite difficult to reproduce and distribute, whereas software is trivial.
Slashdot Mirror
So have I. But you're talking 4 - 6 vs 12 - 16 MB of RAM, when RAM was probably the most expensive component in the entire machine. That was a *huge* difference in hardware price ca. 1994-95. Added to that, Windows 95 was usable on 386SX machines, whereas NT 3.51 really needed a fast 386DX or slow 486.
Only in a handful of very specific configurations.
[...] so I suspect the conditions you outline are a null set.
So if you're so sure Dell can't produce machines any better than Apple's, why do you want them to license OS X ? Possibly because you know Dell would quickly move to sell machines filling the gaping holes in Apple's lineup and subsequently slaughter Apple's hardware sales ?
Not really. NT hasn't had the advantage of a) being initially developed without any need for legacy support and b) then having another significant working over, with little regard for legacy support, as NeXT was to get to OS X. *That* is why OS X bloody well should have brought something new to the table (which it did, with its GUI layer).
While it's certainly arguable some parts of NeXT were better than NT (display postscript, for example), its similarly arguable than at least as many parts of NT were better than NeXT (NTFS, pervasive multithreading, pervasive ACLs).
I did notice that OSX is a turd.
I think that's rather harsh. OS X is most certainly sluggish, even on high end hardware (which has brought me to the conclusion it's a low-level system problem - I used to think it was just the old G4 machines), but it has many excellent attributes, including its UI and rich APIs.
Security was an issue that had to be looked at, for sure; with sufficient review, probably not a big deal.
Well, I'd be hesistant about calling a migration from a single-user architecture to a multiuser architecture "not a big deal"...
BeOS was certainly cool. It's problem was that it wasn't better _enough_ at the things it was better at and that many of them were considered to be of less importance than the things it wasn't better at by consumers.
On the other hand, Linux puts them together perhaps most successfully.
I can't agree. From a technical perspective I find Linux to be largely a well of mediocrity (with the exception of its RAID and LVM capabilities) plagued by a seemingly unending string of rewrites of its core components, usually to bring them to the same capabilities that similar components in other OSes have had from day one (because they were actually *designed* before being implemented).
It runs on more hardware than anything but perhaps netbsd - I haven't done a comparison lately, but last time I did, netbsd was winning that particular race.
I'm not sure why people think a huge list of supported platforms is important. It's clearly a metric massively biased against commercial platforms.
And it's very stable and can be made pretty incredibly secure if you're willing to use nonstandard security frameworks.
True enough, but Windows on decent hardware is hardly _unstable_, and the vast, vast majority of security breaches don't happen due to software faults, but from user error.
I'm willing to forgive Linux and other FOSS projects for reinventing the wheel, because their reimplementations typically do things the originals can't handle.
Linux definitely handles corner cases better than Windows or OS X, but I can't say I'd agree it handles the common cases better.
Certainly you can. But a hell of a lot more people use Mac Minis, MacBooks, iMacs, etc than use bottom-end PCs running Linux.
Macintosh does have something Linux doesn't, and that's marketing.
Actually, the biggest thing "Macintosh" has that "Linux" doesn't is people selling it. Plus widespread commercial support (both hardware and software), better usability and better software integration. Apple hardware also tends to be fairly solid (if somewhat more expensive) - something that can be a bit hit and miss with PCs.
If I were to think of one place money could best be spent to promote Linux, it would be advertising. Get the word out. Put out some clever TV commercials, etc.
No-one is really trying to sell Linux PCs (and make no mistake, average consumers don't buy OSes, they buy computers). Given that in the cutthroat world of PCs, the economic advantage of being able to undercut your competitors on price by ~$50 (roughly the OEM cost of Windows) is huge, I think that's pretty indicative that no-one thinks Linux is a strong competitor to either Windows or Mac in that market space.
On the corporate desktop, most companies won't even consider macintosh computers, because they can't be maintained the same way other PCs can - you just don't have access to the same commodity parts, which is mostly to say you can't just throw a new motherboard in a machine that HAS to be up today and reinstall the OS for drivers - if your OS even demands that, like Linux doesn't.
Assuming by "corporate world" you mean "big business" and not "small company", then I think any suggestions people aren't buying Macs because they can't just throw an off the shelf replacement part in is rather naive (and that's ignoring that a significant proportion of hardware in modern Macs *is* replacable with off the shelf parts). The corporate world works on 3 or 5 year warranties and *loathes* having to do such hands-on maintenance itself - particularly especially labour intensive work like changing motherboards - because it's more expensive.
(There are exceptions to this, of course, but they're few and far between.)
Macs are (relatively) unpopular in business because they're relatively expensive, because until quite recently they didn't integrate particularly well into existing infrastructure and because frighteningly large numbers of businesses rely on poorly written software, often developed in-house and/or no longer maintained, that only runs on Windows.
Obviously there are some stunning counterexamples. NeXTStep rides again. Too bad it didn't gain acceptance when nobody else had features like those.
That probably had something to do with its pricetag.
Linux can be a DC, but you're right, I wouldn't do that either.
As far as I know, Samba can't be an Active Directory Domain Controller and, to be quite frank, even if the developers said it could, it would have to demonstrate trouble-free running in such a role for a good year or two before I'd trust it as an integral piece of infrastructure (not that I have anything against the Samba developers, or question their abilities, but their entire application *is* based around reverse engineering another very large, very complex piece of software).
And that active directory shit wouldn't be necessary if it weren't a big fat windows network to begin with.
If better alternatives were widespread and cheaper, they'd be used instead. Not to mention, AD stands up quite well on its own merits.
Anyway, tt's better at being a fileserver than Windows, so I think you're on drugs t
Of course you could. You have the "right"[0] to try and charge people for anything, if you want.
It wouldn't be as practical, which is one reason why we have copyrights, but copyright is certainly not what gives you the right to sell something. Copyright gives you exclusive distribution rights. There's a difference.
Copyright - more accurately the enforcement thereof - is what gives you the *ability* (ie: makes it "practical") to charge for things that could otherwise, easily, be obtained for free (or close enough to free). Again, my only major point is that trying to equate book and software sales is disingenuous, to say the least - both with or without the existence of copyright.
[0] I really hate it when the word "right" is tossed around so casually like this, it devalues its importance.
Maybe if your "living room" is only 6 feet wide...
Personally, I consider the Mini update to be the most interesting. The bottom end Mini is now an ideal machine to buy and put Windows MCE on for a small, silent, HD-capable HTPC.
I may just have to buy one and do just that :).
A Mac Mini is an iBook/MacBook without a screen. IIRC, they're based on the same motherboard.
Your argument is disingenuous to the point of deliberate deception. There was only one major reason why people didn't buy the Cube, and that was the ridiculously high price.
Far from delivering a headless iMac (which by all rational thought, would be - at worst - no more expensive than an actual iMac), Apple delivered a machine that cost about the same (more than, IIRC) as a PowerMac of the day with less capabilities.
I think Apple learned an important lesson that day. The most vocal group of people demanding a specific product and promising to buy it will usually not actually buy what they say they want. They are just looking to get something they can't have, and when they can have it, they don't want it anymore.
Apple didn't deliver what they wanted. Nor did they ever have any intention of doing so, because Apple is well aware that a headless iMac would destroy PowerMac/Mac Pro sales.
If the only difference between and iMac and a Mini was the screen, you might just have a point.
But it's not. The internal differences are significant.
A rather large proportion of users have been calling for a headless iMac pretty much since the first G3 iMac was released, precisely because they know it *would* be better value.
Apple won't do it, of course, because such a machine would absolutely slaughter PowerMac (/"Mac Pro") sales.
IME, the older a floppy disk is, the more reliable it will be. Certainly, the last few times I've bought (admittedly 3.5", not 5.25") floppies, it hasn't been at all unusual to have a couple not work out of the box.
For those who want reliable floppies, look for some old Windows, DOS, OS/2 or Office installation disks - practically indestructible.
Which would just drain Apple's profits even more...
Apple make their money from hardware sales. When those sales tank because most people find they can buy a more desirable machine from Dell for less money, Apple is going to be in serious trouble, just like they were the last time they went down the OS licensing path.
Unless, of course, the OS licensing costs are high enough to cover the difference - in which case a Dell Mac and an Apple Mac are going to cost the same and no-one will buy the Dell Macs, making any such arrangement of zero interest to Dell.
Here's an idea: why don't you come up with some sort of reasonable example of how Apple might license OS X to Dell that doesn't involve either a) Apple going broken because the bulk of their hardware sales evaporate, or b) Dell machines that cost as much as Apple machines and subsequently being about as popular as a fart in an elevator.
The problem for Apple is that there are quite a few gaping holes in their lineup, precisely where Dell can put machines and large numbers of customers want to buy.
For a guy who started cursing me out for being dense, [...]
Because you *were* being dense. That, or deliberately feigning ignorance.
[...] you sure are the pot. It's called long-sleeves and slacks. As for looking not quite right, you don't know a thing about what you are talking about.
I asked *you* what it would look like, precisely because I don't know. Your answer was "if anything, too perfect". People with "perfect" skin, are not common. They will stand out. That is the only point I am making.
You imply that you are in europe, [...]
I have implied nothing about my location except that I'm not an American.
[...] go look for a tv show called "there's something about miriam" in the UK and the lawsuit it generated. Miriam's level of ability to pass was probably around the 75th percentile for that age range, and she did it in a bikini and made out with the guys.
You'll pardon me if I'm not prepared to judge the effectiveness of a trained observer based on the events in a "reality TV" show.
No, those parts are NOT. They are EXPENSIVE. Just think about it for a second, enough people to professionally evaluate every single person getting on each plane, plus back-ups for when someone is questioned multiple times. Similarly properly, or rather effectively, trained observers of xray machines are also expensive because they burn out real quick.
So we shouldn't pursue better security methods because they're not free ?
In just the US alone there are about 1.8M passengers per day. The TSA currently employes over 50,000 screeners to handle that many passengers with just cursory, 4-5 second examinations of luggage. In order to include even just 30 second interrogations of every single passenger you are going to need at least the same number of people - but you are going to have to pay them more since they will be higher skilled. Right there you've probably more than doubled the entire budget of the TSA. That isn't anyone's definition of feasible.
Pretty much every time I fly, I get *at least* a 30 second chat with an airline employee at the check-in desk, including the standard questions about who packed my bag. I'm not talking about adding (many) more people to the existing processes, I'm talking about making them better with improvements to existing resources. That whole stepping through the metal detector, being "randomly" selected and subsequently searched could use some improvements, as well.
Near as I can tell, your attitude is that we shouldn't be making any attempts to improve the current systems at all (oh, apart from making it harder to get weapons on board, because apparently where you fly it's trivial to carry an arsenal into the plane) and if anything should be making them *less* effective by not paying any attention at all to passenger attributes.
Maybe my constant quoting has gone over your head. Everytime I say "stupid, idealistic, bleeding-heart-leftists system" that's a slightly paraphrased quote from your first post.
That phrase from my first post, which you continually misquote out of context, was not referring to any profiling system at all, it was referring to the attitudes of such people, which preclude them from admitting even the slightest possibility that maybe some people are more likely to be terrorists than others.
Which is pretty obvious, when you actually read what I wrote. But you were only interested in latching on to some suitable catchphrase and beating it to death.
OBVIOUSLY if you know people would object to something then you know what they are objecting to.
Indeed. They are objecting not to a specific implementation, but to a concept - and not on rational grounds, but idealistic ones.
If, on the other hand, you think a proper profiling system would not cause such objections, then why did you bother making that statement in the first place?
Any suggestion of "profiling" in any form
The roadblock to Dell shipping OS X machines isn't Dell, Microsoft or Vista, it's Apple.
Thumbnailing in Explorer windows has been around since the first release of IE4, back in 1997. It's not a new feature to Windows.
In the Mac world, I seem to recall that working in -early- OSX versions.
And it was frighteningly buggy in them, as well. Particularly dealing with "different" codecs like DivX and friends.
Eh ? The vast majority of "new features" in Vista are in the low-level system and are neither easily visible to, nor easily duplicated by, end users.
My 1Ghz, 768Mb RAM, 7200rpm HDD iBook has 10.4, and it's still annoyingly slow for anything non-trivial (and occasionally for trivial things).
Fortunately, I only ever do trivial things on it, so that's not a huge problem (the size, durability and battery life are more important to me), but even on fast dual-core and up Macs, I still find OS X sluggish to use under any sort of meaningful load.
I've been using OS X development and release versions regularly since it still looked like MacOS Classic, and while it has been getting better with every release, it is by no means "fast", unless you have a monstrously fast machine and relatively light application load.
You can't really leave BeOS in if you're going to disallow NetBSD, OpenBSD and Linux for being "relatively unfinished".
I don't think a serious argument can be made for Windows 95 being superior to OS/2 3.0, except in the area of available software.
Single Input Queue.
I was an OS/2 user back in the day (migrated to NT4 Beta 2 in early 1996) and I think picking between Windows 95 and OS/2, without any historical influence, would be extrememly difficult. OS/2 technical had "better" multitasking and memory protection, but in actual usage - particularly if you were only using 32 bit applications and drivers in Windows 95 - the two of them were basically the same. Windows 95 was much nicer in terms of hardware resources and was more usable on lower end machines. I only stuck with OS/2 because I already had a significant investment in OS/2 software (NT4 - even at Beta 2 stage - was vastly superior to both in pretty much every way, which was why I switched).
This is also ignoring hardware support, something which Windows 95 had orders of magnitude more of.
Oh and, Windows 95 can't take credit for plug and play. Most of that is done by the hardware; the OS can choose to move things around to other settings, but typically it does not. The drivers are more responsible for this.
How it's done is irrelevant. Windows 95 was the first mainstream OS that supported it (and you *did* need support on the OS side as well back then).
So basically, Windows 95 was a sad, sad joke even then.
Not even close. Windows 95 did what it was designed to do exceptionally well (too well, if anything - it and Windows 98 arguably worked so well they knocked back the migration to Windows NT by several years). It allowed people to install it on reasoanbly priced hardware of the time. It gave them complete backwards compatibility with all their existing software and a substantial chunk of their existing hardware (ie: peripherals).
Nothing else even came close. OS/2 needed at least as much - more, IMHO - hardware resources and had worse legacy support. Linux wasn't even playing the same game in terms of important capabilities (software, UI, hardware support). Windows NT required *way* too much hardware grunt and had even worse legacy software capabilities than OS/2. MacOS required expensive new hardware and software. NeXT required frighteningly expensive new hardware and software.
For those who had access to the media, and had no hardware requiring Windows 95, even NT 3.51 was a better choice.
Untrue. NT 3.51 had substantially higher hardware requirements, the sucky old Windows 3.x interface and dramatically worse legacy software support.
Then along came NT4, with support for that hardware, but with less separation between memory spaces, and it was probably about half as reliable as 3.51.
But with substantially better performance. For the market space, the performance was more important. Not to mention if you stuck with the bog-standard VGA driver the stability difference was negligible.
Asynchronous I/O. API/ABI stability. A development process that isn't design-driven.
Personally, I'm struggling to think of anything Linux does notably *better*, on a totally technical basis. LVM and software RAID are the only things I can come up with and even the former arguably has more to do with Linux tending to be used more for lower-end hardware tha anything else (Linux's LVM, however, is excellent and unmatched in any other stock-standard OS, with the possible exception of Solaris).
Or perhaps you might convince me that OSX is not superior to Windows, again, on a technical basis.
OS X has atrocious performance. There are technical reasons, I am sure, why this is so, but Windows (and Linux) spank it in performance - *expecially* interactive GUI performance in the case of Windows (with the exception of certain eyecandy that's handled almost completely by the video card). Its scheduling isn't particularly good either - multitasking in OS X is not as good as Windows or Linux - again, especially in terms of GUI interactiveness.
OS X *is* a more advanced OS in some areas, but given how much newer it is, it would damn well want to be. Windows "XP" (ie: Windows NT) was conceived in the late '80s and first released in 1993. You'd certainly hope an OS only released in 2000 that was largely just a warmed over version of an earlier OS (NeXT) would introduce something new.
Hell, just talk me into believing that Windows is a better candidate than BeOS for the user desktop, even in its current state, and I'll be amazed.
BeOS was only single user, therefore less secure. This is before we even get into things like little hardware support, a questionable network stack and it's poor handling of higher end machines (eg: large memory hacks). BeOS's widespread use of multithreading (which Windows also has) was about its most technologically interesting feature. It's kind of hard to make any meaningful comparison, though, since BeOS was never really finished and hasn't been under worthwhile active development for a very long time.
The simple truth is that Windows is technologically retarded compared either to OSX or Linux, its two most significant competitors.
The simple truth is that all contemporary OSes are largely a wash, technologically speaking. Linux is probably the *least* interesting of the bunch, since most of its features were already present in its competitors long, long before they appeared in it (eg: the O(1) scheduler that was such a big milestone, Windows NT had since its first release) and "new" features tend to be reimplementations after they appear on other platforms.
The difference is Windows is taking either new marketshare, or from existing non-Windows installs, whereas Linux is largely just displacing legacy unix installs.
I'm highly sceptical of any assertions Linux is "eating Windows's lunch" in the server world - the evidence simply doesn't bear it out.
I also have to agree with the GP. By and large, Windows and Linux are not competing in the same market spaces (there are some crossovers, eg: webservers, but not that many). On the Desktop it's not even worth talking about - OS X is the real competition to Windows there. On the server side, most people who are after a Windows server, are after it because of functionality Linux can't provide as well (if at all). Similarly, for people actively looking at Linux as a server OS, it's unlikely they'll be considering Windows, usually due to an existing reliance on a unix-like systems (be it software or skill sets), or the higher licensing costs (although that's almost always false economy).
We use both Windows and Linux servers, and the Linux machines vastly outnumber the Windows ones. However, where we do use Windows, Linux is simply not an option - domain controllers, fileserving, groupware, windows-only software. Similarly, where we use Linux - mostly for in-house applications - Windows is a relatively poor choice due to its lack of easy customisability.
I can't think of a single Windows server we have that I would want to replace with a Linux server (and, believe, me, I've considered and even trialled it many times). Likewise, I can't think of any of our Linux servers that would be better off running Windows.
Unless, of course, one of the criteria is "more than ten flights in 3 months", or something similar, and they get checked the eleventh time.
For profiling criteria NOT to be easy to deduce, every now and then the officers would have to NOT check a profiled person. Which defies common sense even more!
From what I can gather, the current "profiling" system used in US airports is utterly worthless (for starters, it's trivial for passengers to determine if they're being searched randomly or because they were flagged). Using it as an example of why profiling in general wouldn't work is, then, not a particularly strong argument.
Even on the arms and legs ?
My point here is that a "woman" who already doesn't quite "look right" (like the vast majority of cross-dressers) will generate more interest the more things about her that are "weird".
That level of profiling, as fully distinct from the "stupid, idealistic, bleeding-heart-leftists" system that you proposed, [...]
I proposed no system. I gave a throwaway example of two different types of travellers.
[...] is not feasible with the kind of limited resources available.
No, but *parts* of it are. Using properly trained staff to ask pointed questions at check-in, having a properly trained observer as part of the metal detector/security gateway, etc.
Well, excuse me for going by what you write and not having ESP.
I challenge you to quote any part of any of my posts even suggesting - let alone stating - my examples were to be considered representative of a "proper profiling system".
My position has been clearly stated, and regularly illustrated, from the start - you pick characteristics that can be altered by the people you are trying to identify and they will alter those characteristics - thus costing you resources and reducing your security. You just keep jumping from characteristic to characteristic and seem to hope that by combining them into a "system" that somehow they won't be figured out.
Your "position" is that profiling won't work, based on the obviously flawed system apparently currently being used in the US.
If I were to "advocate" a solution to the problem, I would pick things that the enemy can not change, or in changing them he neutralizes the threat himself. For example, better automated screening systems - the harder it is to take a weapon on board, the less likely a person is going to do so - regardless of race, religion, family life or anything else directly under the enemy's control.
Getting weapons onto flights is already quite difficult (or at least it is where I fly). Not to mention, a weapon on a plane isn't anywhere near as useful as it was before 9/11 - no plane full of passengers and crew is just going to let a bunch of people take their aircraft over again. The "get a few people with knives on board and take the plane over" plan is vastly less feasible than it used to be.
Most airline security is a placebo at best - even the new measures restricting liquids and gels are pretty much worthless, because if someone really wants to get some liquid explosives onto a plane, they'll just put them into a plastic cigar tube or film container and shove it up their arse. The benefit of it lies not in the probability of catching someone in the process of smuggling a bomb on board, but to force them to spend non-trivial amounts of time planning a mission, allowing regular intelligence and police forces to identify, locate and apprehend them beforehand.
Presumably not, then (not being American, I wouldn't know).
However, the airline is. No-one is forcing you to fly. If you don't want to get searched, don't fly.
You said it wasn't copyright that allowed someone to "charge for stuff", and used the example of a bookshop selling books to support your argument. My point is this conclusion is wrong, because someone selling physical books is a vastly different prospect to someone trying to sell software, in a world without copyright.
Much as I dislike copyright, your analogy is broken. Books are quite difficult to reproduce and distribute, whereas software is trivial.