Same stupidity-driven "worms" with end users to blame (and, to lesser extend, windows is also to blame since it executes the files without asking, where is chmod +x when you need one...).
Windows does not execute downloaded binaries without prompting.
Further, having to chmod +x would add an additional step, but anyone silly enough to download and run some random binary from an IM is hardly going to be slowed down by that - just look at how many people fell victim to the trojan that arrived in a *password protected zip file*.
[...]then have blurry, confusing fight scenes which make it really hard to figure out who's fighting who and what the hell's happening?
Transformers is hardly the only culprit here. The "handycam perspective" has been destroying films for years now.
I used to think it was done just so the director could cheap out on the choreography (after all, it doesn't have to be good when the audience can barely see what's happening), but they're so prevalent now, even in big-budget films, that I can only conclude that the two or three people in the world who like them are very influential.
Huh ? How does 30 spindles of SATA involve any more of this than 30 spindles of FC, SCSI or SAS ? You *were* lamenting about the lack of spindles from defining your storage requirements by how much space and only getting the minimum number of drives to provide that, as I understood it.
I'm using some Pillar equipment that short-strokes SATA drives but makes the residue available at a lower QoS, which works well, but I can make use of the residual space. I've done what you propose in the past, but it's a lot more expensive over the whole life of the boxes than you make out.
I don't see how it could be. 30 spindles is 30 spindles, be they fully utilised 146G drives or 1/3 utilised 500G drives.
ISO is perfectly at home with political pressure, why did it take the outrage of the entire community to make them relent after they succumbed to political pressure?
2TB is easy, but 2TB with decent performance is rather harder. 2TB with decent performance and good post-power-off performance is harder still. Back in the day, 2TB would be spread over 60 or more 7200rpm spindles, whereas today it's on half a dozen 7200rpm spindles. You've got the capacity, but you've got a tenth of the ops/sec. This might well not matter to you, if you're doing data warehousing or serving home directories, or it might be very important indeed if you're doing a lot of OLTP. NetApp still have a lot of margin in their product, and if you want storage that one vendor will look after, they're pretty damn good. The thing to remember is that storage is so cheap now that in a project that needs 10TB, the price of the 10TB is in the noise floor compared to the price of the people, integration, application development, user training, etc, etc. So shaving 20% of the storage costs is perhaps 2% off the project, and that might be worth it for the fact that NetApp kit just works.
A lot of people lose sight of this, but there's no reason why if you need "only" 2TB, that you _buy_ only 2TB. There's no reason why a "2TB array" can't be 30*500G drives from which you only access the first ~130G and ignore the rest (as you say, storage is so cheap the "waste" is irrelevant). When you cut down the amount of physical disk space the drive head has to cover by 1/3, you improve IOPS quite a bit (and potentially save a lot of money).
Many people are finding they can save big bucks by buying 7.2k SATA drives and "short stroking" them instead of buying 15k RPM FC or SAS drives, while still achieving similar performance.
The reality is NetApp is in trouble. Lets face it, 2TB of storage yesterday was big bucks. Today, it is 4 Seagate drives at Best Buy and fit in one PC. Cluster 4 dual AMD x2's together on 1000GB interconnects and it has never been cheaper to spin 8TB into your own appliance. Do it with Linux or Solaris. Or like NetApps, BSD.
You can build a storage solution with 90% of the functionality of enterprise-level kit like NetApp/EMC/Sun/IBM/etc with off the shelf parts, and for substantially less money. However, that last 10% is pretty much impossible to DIY (or DIY cheaply) and if you need it, you *really* need it (ie: the cost is relatively insignificant). Things like redundant controllers, mirrored cache, multipathing, FC target (although 10Gb ethernet will make that irrelevant very shortly), etc. The other benefit you get from that "enterprise" hardware is in terms of "people time" (something techies frequently have trouble both conceptualising and calculating).
With that said, such an "DIY" solution would probably meet the needs of the vast majority of customers, even many of those who think it wouldn't. Lots of environments that have spent big bucks on redundancy or claim to require 24/7 availability still frequently have single points of failure strewn all over the place. Machines with dual PSUs on the same electrical circuits, single database servers, active/standby machine pairs connected to the same KVM/serial server/network switches, ostensibly redundant infrastructure that requires human intervention to "fail over", etc.
None of these have stuffed so much into the kernel. They *can't*. They're userland. Win, OTOH, was GUI from Square 1, and has all sorts of things in the kernel.
What do you think is in the kernel ? (This discussion is rapidly exceeding the scope of comparing IE on Windows to its counterparts on other platforms - unless you think IE is in the kernel.)
We were talking about the days when Microsoft first discovered TCP/IP networking, and the business rags were talking about their rapid turnaround. Neither legacy software or software support concerns apply, as there were none of either.
Perhaps you don't realise pretty much the biggest reason Windows 9x even existed *at all* was to deal with legacy hardware and software.... "None of either" ? Do the ~15 years of PC usage usage preceding Windows 95 - and the significant investment by users in hardware and software during that time - not exist in your world ?
If they'd done proper multiuser/privilege separation back in the Win95 time frame, they'd have saved themselves and their users a lot of grief.
They did. Windows NT. Users weren't really interested at the time because most of their software and hardware was unsupported and it needed a relatively powerful machine to run (another price you pay).
a) Microsoft's 'embrace, extend, extinguish' approach to open standards would argue against it, as in what they did with Kerberos.
You mean extending it in a way the protocol and RFCs allowed for ? How would that argue against Windows being meaningfully more complex than its functionally-equivalent contemporaries ?
b) Much anecdotal evidence on mailing lists, etc, about subtly different APIs.
You'll need to be more specific.
c) My personal experience when faced with attempting to secure a Win2K system, and finding an NSA doc with 20+ pages of registry edits *alone*. I doubt it's become less complex in the meantime.
What's in the equivalent Linux documentation ? (Although given that SELinux-capable distros didn't start showing up until 4 - 5 years later, the comparison is hardly going to be fair.)
The point which was in between the lines is that most people who post in blogs aren't going to copyright their material Automatic copyright grants them some form of protection. They have a recourse if they want to fight someone who is using their content without permission. Corporation's latest bubble-gum pop idol soundtrack is going to be explicitly copyrighted anyway (they don't rely on automatic copyright, you know.) The only people that automatic copyright helps are those people who wouldn't have the time or money to copyright their work. Amateur bloggers fit this category perfectly. Filing for copyright on every post they make would simply take up too much time, and it would slow down the blog (because you can't post it until you copyright it, or you risk losing it.)
The obvious compromise here is that the only "automatic" copyright protection a work gets is of attribution. That is, someone else cannot claim to have created another person's work.
If you want to prevent someone from *profiting* from your work, well, I can't see how requiring you to register it first is either unreasonable or burdensome.
I haven't called anyone stupid. And "every other platforms" hasn't done the same thing. Without exception, they have remained far less monolithic than Windows.
GNOME, KDE and OS X have all implemented the same component architecture as Windows [and IE].
I'm well aware that 9x were single-user. That's what I'd just *said*. That's what I was *complaining* about. You don't point a machine with no concept of privilege separation at public networks, unless you don't mind your users paying the price.
Or you don't have a choice because there's no other way to deliver the users' other [more] important requirements. Like, say, legacy hardware and software support.
Complexity can be minimized. In secure systems, complexity *must* be minimized.
Other platforms with equivalent functionality have equivalent levels of complexity.
No one lives in the CBD. Unless you sleep in your office this shouldn't effect you because it is going to be a public holiday and you should have no reason to be in the CBD.
There's quite a few apartment buildings in, or very near, the lockdown zone. The Toaster, for example, is right next to the Opera House. Not to mention things like the ferry terminal being smack bang in the middle of it all, the pubs, hotels, general tourist attractions, etc. The Sydney CBD isn't like, say, central London - it's not a deserted wasteland on a non-work day.
Whoever thought Sydney was an appropriate place to hold something like APEC should be fired for incompetence. An island resort or relatively isolated hotel(s), would have been a far saner place to have all the meetings. Much less disruption to local residents, businesses and tourists, much easier and cheaper to secure.
Which presumably explains why every other major platform went on to do exactly the same thing, I assume, because all those developers are stupid as well ?
It lead to an awful lot of remote code execution exploits. It was originally done in Win95 days--before Windows had any notion of being a multiuser system. There was no privilege separation, so if IE was exploited, no system file, etc., was safe. I don't believe for a moment that Microsoft wasn't aware of this. I think they simply didn't care about their users.
The nature of Windows 9x means that whether or not IE was "integrated" has no bearing on the security principles you're talking about.
It's *still* a bad idea. In fact, it's a fundamentally poor practice in secure systems design. The resulting system is too monolithic--it's a maze of interdependencies. That leads to security, patching, etc., issues. You may want to Google around a bit.
Every complex system is a "maze of interdependencies". That's the price you pay for a system based on modular, reusable components.
What really disturbs me though is that we've gone from a race of creators, creating goods with agriculture or manufacturing, to a world wide economy of McJobs that pay minimum wage and create NOTHING.
That's because the amount of human labour it takes to create the requisite amount of stuff has dropped dramatically over the last few thousand (heck, mostly the last few hundred) years.
The "problem" you see is the increase in efficiency that improving technology delivers. "Making stuff" just doesn't require as many people as it used to and is more efficiently done elsewhere.
One of the great things about Solaris is its backward compatibility. Sun is not going to be able to get rid of the tools everybody is complaining about because there are a bazillion of scripts out there that just have to work. Period. Why break them? Just to please the Linux crowd? If you need GNU tools, hey most of them are bundled with Solaris 10. Others can be downloaded and installed a la carte.
(I speak as someone who cut his sysadmin teeth on Solaris and - to a lesser extent - FreeBSD boxes, although I've since moved to a primarily RH Linux environment.)
The "Linux crowd" is rapidly become a very non-trivial proportion of the sysadmin community. In less than a decade, I expect a significant majority of practicing "unix sysadmins" will have only a passing acquaintance with unix platforms other than "Linux". This is a very important issue for Sun to consider if they want those people to be using, promoting and *preferring* their platform.
The other - more important - thing is that the Solaris userland sucks. Personally I prefer BSD, but even GNU with its inconsistencies and poor documentation is preferable overall to a system that either requires years of experience or lots of tweaking to be usable. "Why break it", you ask ? They'd be *fixing* it.
Of course, the legacy support issue is equally as important and cannot be ignored. Which is why the next version of Solaris would ideally come with an install-time option for a "GNU userland" or a "Solaris userland", defaulting to the latter (and ideally changeable after install relatively easily).
No, you can't. Not nearly as well, anyway. Show me how you add a third disk to a system with two mirrored disks and convert it to a RAID-5 array so that you get some benefit from the additional disk -- all quickly, easily and automatically. RAID-Z does it just fine.
I'm about 99.9% sure ZFS cannot do what you are describing (although high-end RAID implementations often can). There's certainly nothing in the man pages about converting from one RAID level to another, and changing the number of disks in a RAIDZ[2] zpool is definitely impossible (although one of the most frequently requested features).
Show me how you build a RAID-5 array with disks of many different sizes. RAID-Z does it just fine.
Huh ? I've never seen a RAID implementation (software _or_ hardware) that *can't* do this. Your array size is simple constrainted by the smallest device (as it is in ZFS).
RAID-Z also has the ability to apply greater redundancy for more important files. At present this is only used for metadata, but it will be available for user files as well.
No idea what you're talking about here, but given you're currently at two strikes, I'd guess you don't either.
Microsoft continues to use its o/s monopoly to skew the browser market by bundling IE with every copy of Windows.
You mean like every other OS vendor ?
Are you saying Microsoft should be denied including industry standard functionality in their products ? Doesn't seem like that would be particularly fair to their customers...
In that case, why stop with the browser ? Why should they be allowed to bundle a GUI ? Or a network stack ? Or even an API ? Why are you not insisting Microsoft only be allowed to sell a kernel and nothing else ?
In another 10 years, standards compliant browsers may have matured enough to support Web 3.0 applications that make the Windows platform unnecessary or optional for application developers who want to reach a large audience. At that point, you would be right in saying that Microsoft's interference in the browser market has been overcome. But calculate the price: 25 years to fulfill the promise that Netscape unleashed in the 1990s.
The "dream that Netscape unleashed" in the 1990s was proprietry HTML extensions tied to their server platform. That was their business plan. Microsoft just grabbed the ball and ran with it.
Prices for Microsoft operating systems have actually gone UP, not down [...]
False.
Retail US$ prices at release (and adjusted for inflation to 2006):
Windows 3.1 + DOS 6.2: $150 + $50 = $200 ($214 + $71 = $285)
Windows 95: $209 ($274)
Windows 98: $209 ($255)
Windows XP Home: $199 ($226)
Windows Vista Home Basic: $199
Windows Vista Home Premium: $239
Windows *is* getting cheaper. Further, it has also vastly improved its functionality and reliability over the same timeframe, so you're getting more.
(Of course, this comparison completely ignores that almost no-one pays retail price for Windows.)
A WYSIWYG word processor can be just as stable as a non-WYSIWYG one (and vice versa). If a file from ten years ago doesn't render exactly the same today, that's because either the rendering algorithms have substantially changed, or the format was too much tied to the platform. Both are completely unrelated to WYSIWYG and are only due to bad decisions made by the program writers.
Or something else which affects how it looks has changed in the intervening 10 years. Like, say, your printer, screen resolution, paper size, etc.
The point of WYSIWYG is not that a document looks the same on different machines - that would be What You See Is What They Get - the point of WYSIWYG is that what comes out of *your* printer looks the same as what's on *your* screen.
I fail to see the fuss, both formats suck and really have no place as a desktop publishing format. They are crappy WYSIWYG data dumps that are heavily tied to rendering algorithms of their respective editor and really are not archival safe.
Indeed. You'll also find that hammers are poor at undoing screws and cars aren't so good at taking you overseas.
What is it going to be like 50 years from now when you try to pull up an old manuscript? You know how Popular Science likes to pull up magazine issues from 40+ years ago, I wonder how they are going to manage that 40 years from now when the proprietary and open file formats are unsupported and "obsolete".
They'll use a format that actually meant for that sort of thing like, say, PDF.
The point of "WYSIWYG" is not - despite what a lot of people (including those that should know better) think - that a document looks the same on computer B as it does on computer A. It's that the document that comes out of the printer looks the same as it does on the screen.
In the case of Windows, the GUI is not merely a graphical shell. It is an integral part of the OS. You cannot (except in the case of newer and much older versions) install JUST the kernel and userspace sans the GUI.
I don't entirely disagree with you, but on technicality it's more complicated. Unless I'm mistaken, a Linux system starts the GUI "from the command-line" through a script that is run when the OS is ready to start such an application. As such, the GUI is running on top of the command-line (i.e. it's just like any other program being run by the OS, just a lot more complicated).
The semantics of how X starts depends entirely upon how you have X (amongst other things) setup.
Without the command-line it's the "if a tree falls in the forest and no one is around to hear it, does it make a sound?" situation with the OS: if the command-line is non-functional, can the operating system really be said to be working?
Depends on the OS. MacOS "Classic", for example - or any number of embedded OS implementation like Symbian-using Nokia's - have no commandline at all.
On the other hand, if the GUI fails but the operating system can still be used from the command-line, it is difficult to say that the machine is now non-functional. In the former situation, no one can operate the system but in the latter, there are at least some people who can.
You're missing the point. If you consider a command line shell to be a valid component of an OS, there is no logical argument that can be made that a graphical shell is not an equally valid component in principle.
Yes, in a roundabout and counterintuitive, yet valid, way. GPL code stays "free" no matter what happens to it, whereas BSD code can always be rendered un-free.
No, it cannot. Someone else's code they added may not be BSD-licensed if they don't want it to be, but that has no bearing on the BSD-licensed code their work was based from.
Remember, the BSDL is about what you want to happen to your code. The GPL is about what you want to happen to other people's code.
So if I ask Microsoft for changes to the BSD code they used, I can get it and continue that line of development (not that I'd want to. I've seen Microsoft example code, and it makes me shudder)? If so, then the code is still available and is Free. If not, then that line of code development is dead and therefore neither Free nor free.
No, because you want *Microsoft's* code, not the BSD code. Very different situation.
Please learn what an operating system is. You don't know what you are talking about. The machine is functional without a GUI. You can do nearly everything without a GUI that you can do with a GUI. Some things are even substantially faster at the command line.
If you believe a command-line shell can be considered part of the "operating system", you have zero grounds for saying a graphical shell is not. They are merely different implementations of the same concept.
Slashdot Mirror
Same stupidity-driven "worms" with end users to blame (and, to lesser extend, windows is also to blame since it executes the files without asking, where is chmod +x when you need one...).
Windows does not execute downloaded binaries without prompting.
Further, having to chmod +x would add an additional step, but anyone silly enough to download and run some random binary from an IM is hardly going to be slowed down by that - just look at how many people fell victim to the trojan that arrived in a *password protected zip file*.
Interesting that Microsoft is, yet again, directly or indirectly, responsible for their misfortune.
Indeed. Just as interesting as how oxygen is, yet again, directorly or indirectly, responsible for their misfortune.
[...]then have blurry, confusing fight scenes which make it really hard to figure out who's fighting who and what the hell's happening?
Transformers is hardly the only culprit here. The "handycam perspective" has been destroying films for years now.
I used to think it was done just so the director could cheap out on the choreography (after all, it doesn't have to be good when the audience can barely see what's happening), but they're so prevalent now, even in big-budget films, that I can only conclude that the two or three people in the world who like them are very influential.
Power, space, heat, MTBF, complexity.
Huh ? How does 30 spindles of SATA involve any more of this than 30 spindles of FC, SCSI or SAS ? You *were* lamenting about the lack of spindles from defining your storage requirements by how much space and only getting the minimum number of drives to provide that, as I understood it.
I'm using some Pillar equipment that short-strokes SATA drives but makes the residue available at a lower QoS, which works well, but I can make use of the residual space. I've done what you propose in the past, but it's a lot more expensive over the whole life of the boxes than you make out.
I don't see how it could be. 30 spindles is 30 spindles, be they fully utilised 146G drives or 1/3 utilised 500G drives.
ISO is perfectly at home with political pressure, why did it take the outrage of the entire community to make them relent after they succumbed to political pressure?
They did ? When ? In what way ?
2TB is easy, but 2TB with decent performance is rather harder. 2TB with decent performance and good post-power-off performance is harder still. Back in the day, 2TB would be spread over 60 or more 7200rpm spindles, whereas today it's on half a dozen 7200rpm spindles. You've got the capacity, but you've got a tenth of the ops/sec. This might well not matter to you, if you're doing data warehousing or serving home directories, or it might be very important indeed if you're doing a lot of OLTP. NetApp still have a lot of margin in their product, and if you want storage that one vendor will look after, they're pretty damn good. The thing to remember is that storage is so cheap now that in a project that needs 10TB, the price of the 10TB is in the noise floor compared to the price of the people, integration, application development, user training, etc, etc. So shaving 20% of the storage costs is perhaps 2% off the project, and that might be worth it for the fact that NetApp kit just works.
A lot of people lose sight of this, but there's no reason why if you need "only" 2TB, that you _buy_ only 2TB. There's no reason why a "2TB array" can't be 30*500G drives from which you only access the first ~130G and ignore the rest (as you say, storage is so cheap the "waste" is irrelevant). When you cut down the amount of physical disk space the drive head has to cover by 1/3, you improve IOPS quite a bit (and potentially save a lot of money).
Many people are finding they can save big bucks by buying 7.2k SATA drives and "short stroking" them instead of buying 15k RPM FC or SAS drives, while still achieving similar performance.
The reality is NetApp is in trouble. Lets face it, 2TB of storage yesterday was big bucks. Today, it is 4 Seagate drives at Best Buy and fit in one PC. Cluster 4 dual AMD x2's together on 1000GB interconnects and it has never been cheaper to spin 8TB into your own appliance. Do it with Linux or Solaris. Or like NetApps, BSD.
You can build a storage solution with 90% of the functionality of enterprise-level kit like NetApp/EMC/Sun/IBM/etc with off the shelf parts, and for substantially less money. However, that last 10% is pretty much impossible to DIY (or DIY cheaply) and if you need it, you *really* need it (ie: the cost is relatively insignificant). Things like redundant controllers, mirrored cache, multipathing, FC target (although 10Gb ethernet will make that irrelevant very shortly), etc. The other benefit you get from that "enterprise" hardware is in terms of "people time" (something techies frequently have trouble both conceptualising and calculating).
With that said, such an "DIY" solution would probably meet the needs of the vast majority of customers, even many of those who think it wouldn't. Lots of environments that have spent big bucks on redundancy or claim to require 24/7 availability still frequently have single points of failure strewn all over the place. Machines with dual PSUs on the same electrical circuits, single database servers, active/standby machine pairs connected to the same KVM/serial server/network switches, ostensibly redundant infrastructure that requires human intervention to "fail over", etc.
None of these have stuffed so much into the kernel. They *can't*. They're userland. Win, OTOH, was GUI from Square 1, and has all sorts of things in the kernel.
What do you think is in the kernel ? (This discussion is rapidly exceeding the scope of comparing IE on Windows to its counterparts on other platforms - unless you think IE is in the kernel.)
We were talking about the days when Microsoft first discovered TCP/IP networking, and the business rags were talking about their rapid turnaround. Neither legacy software or software support concerns apply, as there were none of either.
Perhaps you don't realise pretty much the biggest reason Windows 9x even existed *at all* was to deal with legacy hardware and software.... "None of either" ? Do the ~15 years of PC usage usage preceding Windows 95 - and the significant investment by users in hardware and software during that time - not exist in your world ?
If they'd done proper multiuser/privilege separation back in the Win95 time frame, they'd have saved themselves and their users a lot of grief.
They did. Windows NT. Users weren't really interested at the time because most of their software and hardware was unsupported and it needed a relatively powerful machine to run (another price you pay).
a) Microsoft's 'embrace, extend, extinguish' approach to open standards would argue against it, as in what they did with Kerberos.
You mean extending it in a way the protocol and RFCs allowed for ? How would that argue against Windows being meaningfully more complex than its functionally-equivalent contemporaries ?
b) Much anecdotal evidence on mailing lists, etc, about subtly different APIs.
You'll need to be more specific.
c) My personal experience when faced with attempting to secure a Win2K system, and finding an NSA doc with 20+ pages of registry edits *alone*. I doubt it's become less complex in the meantime.
What's in the equivalent Linux documentation ? (Although given that SELinux-capable distros didn't start showing up until 4 - 5 years later, the comparison is hardly going to be fair.)
The point which was in between the lines is that most people who post in blogs aren't going to copyright their material Automatic copyright grants them some form of protection. They have a recourse if they want to fight someone who is using their content without permission. Corporation's latest bubble-gum pop idol soundtrack is going to be explicitly copyrighted anyway (they don't rely on automatic copyright, you know.) The only people that automatic copyright helps are those people who wouldn't have the time or money to copyright their work. Amateur bloggers fit this category perfectly. Filing for copyright on every post they make would simply take up too much time, and it would slow down the blog (because you can't post it until you copyright it, or you risk losing it.)
The obvious compromise here is that the only "automatic" copyright protection a work gets is of attribution. That is, someone else cannot claim to have created another person's work.
If you want to prevent someone from *profiting* from your work, well, I can't see how requiring you to register it first is either unreasonable or burdensome.
I haven't called anyone stupid. And "every other platforms" hasn't done the same thing. Without exception, they have remained far less monolithic than Windows.
GNOME, KDE and OS X have all implemented the same component architecture as Windows [and IE].
I'm well aware that 9x were single-user. That's what I'd just *said*. That's what I was *complaining* about. You don't point a machine with no concept of privilege separation at public networks, unless you don't mind your users paying the price.
Or you don't have a choice because there's no other way to deliver the users' other [more] important requirements. Like, say, legacy hardware and software support.
Complexity can be minimized. In secure systems, complexity *must* be minimized.
Other platforms with equivalent functionality have equivalent levels of complexity.
No one lives in the CBD. Unless you sleep in your office this shouldn't effect you because it is going to be a public holiday and you should have no reason to be in the CBD.
There's quite a few apartment buildings in, or very near, the lockdown zone. The Toaster, for example, is right next to the Opera House. Not to mention things like the ferry terminal being smack bang in the middle of it all, the pubs, hotels, general tourist attractions, etc. The Sydney CBD isn't like, say, central London - it's not a deserted wasteland on a non-work day.
Whoever thought Sydney was an appropriate place to hold something like APEC should be fired for incompetence. An island resort or relatively isolated hotel(s), would have been a far saner place to have all the meetings. Much less disruption to local residents, businesses and tourists, much easier and cheaper to secure.
It's a *fundamentally* bad thing to do.
Which presumably explains why every other major platform went on to do exactly the same thing, I assume, because all those developers are stupid as well ?
It lead to an awful lot of remote code execution exploits. It was originally done in Win95 days--before Windows had any notion of being a multiuser system. There was no privilege separation, so if IE was exploited, no system file, etc., was safe. I don't believe for a moment that Microsoft wasn't aware of this. I think they simply didn't care about their users.
The nature of Windows 9x means that whether or not IE was "integrated" has no bearing on the security principles you're talking about.
It's *still* a bad idea. In fact, it's a fundamentally poor practice in secure systems design. The resulting system is too monolithic--it's a maze of interdependencies. That leads to security, patching, etc., issues. You may want to Google around a bit.
Every complex system is a "maze of interdependencies". That's the price you pay for a system based on modular, reusable components.
What really disturbs me though is that we've gone from a race of creators, creating goods with agriculture or manufacturing, to a world wide economy of McJobs that pay minimum wage and create NOTHING.
That's because the amount of human labour it takes to create the requisite amount of stuff has dropped dramatically over the last few thousand (heck, mostly the last few hundred) years.
The "problem" you see is the increase in efficiency that improving technology delivers. "Making stuff" just doesn't require as many people as it used to and is more efficiently done elsewhere.
So why aren't starving people knocking down buildings to plant crops now?
Which starving people are you thinking of whose potential farmland is occupied by buildings ?
One of the great things about Solaris is its backward compatibility. Sun is not going to be able to get rid of the tools everybody is complaining about because there are a bazillion of scripts out there that just have to work. Period. Why break them? Just to please the Linux crowd? If you need GNU tools, hey most of them are bundled with Solaris 10. Others can be downloaded and installed a la carte.
(I speak as someone who cut his sysadmin teeth on Solaris and - to a lesser extent - FreeBSD boxes, although I've since moved to a primarily RH Linux environment.)
The "Linux crowd" is rapidly become a very non-trivial proportion of the sysadmin community. In less than a decade, I expect a significant majority of practicing "unix sysadmins" will have only a passing acquaintance with unix platforms other than "Linux". This is a very important issue for Sun to consider if they want those people to be using, promoting and *preferring* their platform.
The other - more important - thing is that the Solaris userland sucks. Personally I prefer BSD, but even GNU with its inconsistencies and poor documentation is preferable overall to a system that either requires years of experience or lots of tweaking to be usable. "Why break it", you ask ? They'd be *fixing* it.
Of course, the legacy support issue is equally as important and cannot be ignored. Which is why the next version of Solaris would ideally come with an install-time option for a "GNU userland" or a "Solaris userland", defaulting to the latter (and ideally changeable after install relatively easily).
No, you can't. Not nearly as well, anyway. Show me how you add a third disk to a system with two mirrored disks and convert it to a RAID-5 array so that you get some benefit from the additional disk -- all quickly, easily and automatically. RAID-Z does it just fine.
I'm about 99.9% sure ZFS cannot do what you are describing (although high-end RAID implementations often can). There's certainly nothing in the man pages about converting from one RAID level to another, and changing the number of disks in a RAIDZ[2] zpool is definitely impossible (although one of the most frequently requested features).
Show me how you build a RAID-5 array with disks of many different sizes. RAID-Z does it just fine.
Huh ? I've never seen a RAID implementation (software _or_ hardware) that *can't* do this. Your array size is simple constrainted by the smallest device (as it is in ZFS).
RAID-Z also has the ability to apply greater redundancy for more important files. At present this is only used for metadata, but it will be available for user files as well.
No idea what you're talking about here, but given you're currently at two strikes, I'd guess you don't either.
ZFS is very good, but it's not god's gift yet.
Microsoft continues to use its o/s monopoly to skew the browser market by bundling IE with every copy of Windows.
You mean like every other OS vendor ?
Are you saying Microsoft should be denied including industry standard functionality in their products ? Doesn't seem like that would be particularly fair to their customers...
In that case, why stop with the browser ? Why should they be allowed to bundle a GUI ? Or a network stack ? Or even an API ? Why are you not insisting Microsoft only be allowed to sell a kernel and nothing else ?
In another 10 years, standards compliant browsers may have matured enough to support Web 3.0 applications that make the Windows platform unnecessary or optional for application developers who want to reach a large audience. At that point, you would be right in saying that Microsoft's interference in the browser market has been overcome. But calculate the price: 25 years to fulfill the promise that Netscape unleashed in the 1990s.
The "dream that Netscape unleashed" in the 1990s was proprietry HTML extensions tied to their server platform. That was their business plan. Microsoft just grabbed the ball and ran with it.
Prices for Microsoft operating systems have actually gone UP, not down [...]
False.
Retail US$ prices at release (and adjusted for inflation to 2006):
Windows 3.1 + DOS 6.2: $150 + $50 = $200 ($214 + $71 = $285)
Windows 95: $209 ($274)
Windows 98: $209 ($255)
Windows XP Home: $199 ($226)
Windows Vista Home Basic: $199
Windows Vista Home Premium: $239
Windows *is* getting cheaper. Further, it has also vastly improved its functionality and reliability over the same timeframe, so you're getting more.
(Of course, this comparison completely ignores that almost no-one pays retail price for Windows.)
A WYSIWYG word processor can be just as stable as a non-WYSIWYG one (and vice versa). If a file from ten years ago doesn't render exactly the same today, that's because either the rendering algorithms have substantially changed, or the format was too much tied to the platform. Both are completely unrelated to WYSIWYG and are only due to bad decisions made by the program writers.
Or something else which affects how it looks has changed in the intervening 10 years. Like, say, your printer, screen resolution, paper size, etc.
The point of WYSIWYG is not that a document looks the same on different machines - that would be What You See Is What They Get - the point of WYSIWYG is that what comes out of *your* printer looks the same as what's on *your* screen.
I fail to see the fuss, both formats suck and really have no place as a desktop publishing format. They are crappy WYSIWYG data dumps that are heavily tied to rendering algorithms of their respective editor and really are not archival safe.
Indeed. You'll also find that hammers are poor at undoing screws and cars aren't so good at taking you overseas.
What is it going to be like 50 years from now when you try to pull up an old manuscript? You know how Popular Science likes to pull up magazine issues from 40+ years ago, I wonder how they are going to manage that 40 years from now when the proprietary and open file formats are unsupported and "obsolete".
They'll use a format that actually meant for that sort of thing like, say, PDF.
The point of "WYSIWYG" is not - despite what a lot of people (including those that should know better) think - that a document looks the same on computer B as it does on computer A. It's that the document that comes out of the printer looks the same as it does on the screen.
Word processing != desktop publishing.
In the case of Windows, the GUI is not merely a graphical shell. It is an integral part of the OS. You cannot (except in the case of newer and much older versions) install JUST the kernel and userspace sans the GUI.
*You* can't. Microsoft, however, can.
I don't entirely disagree with you, but on technicality it's more complicated. Unless I'm mistaken, a Linux system starts the GUI "from the command-line" through a script that is run when the OS is ready to start such an application. As such, the GUI is running on top of the command-line (i.e. it's just like any other program being run by the OS, just a lot more complicated).
The semantics of how X starts depends entirely upon how you have X (amongst other things) setup.
Without the command-line it's the "if a tree falls in the forest and no one is around to hear it, does it make a sound?" situation with the OS: if the command-line is non-functional, can the operating system really be said to be working?
Depends on the OS. MacOS "Classic", for example - or any number of embedded OS implementation like Symbian-using Nokia's - have no commandline at all.
On the other hand, if the GUI fails but the operating system can still be used from the command-line, it is difficult to say that the machine is now non-functional. In the former situation, no one can operate the system but in the latter, there are at least some people who can.
You're missing the point. If you consider a command line shell to be a valid component of an OS, there is no logical argument that can be made that a graphical shell is not an equally valid component in principle.
Yes, in a roundabout and counterintuitive, yet valid, way. GPL code stays "free" no matter what happens to it, whereas BSD code can always be rendered un-free.
No, it cannot. Someone else's code they added may not be BSD-licensed if they don't want it to be, but that has no bearing on the BSD-licensed code their work was based from.
Remember, the BSDL is about what you want to happen to your code. The GPL is about what you want to happen to other people's code.
So if I ask Microsoft for changes to the BSD code they used, I can get it and continue that line of development (not that I'd want to. I've seen Microsoft example code, and it makes me shudder)? If so, then the code is still available and is Free. If not, then that line of code development is dead and therefore neither Free nor free.
No, because you want *Microsoft's* code, not the BSD code. Very different situation.
Please learn what an operating system is. You don't know what you are talking about. The machine is functional without a GUI. You can do nearly everything without a GUI that you can do with a GUI. Some things are even substantially faster at the command line.
If you believe a command-line shell can be considered part of the "operating system", you have zero grounds for saying a graphical shell is not. They are merely different implementations of the same concept.