Also note that only Client (Windows and CALs), Information Worker (Office, Project, Etc) and Server and Tools (Exchange, SQL, Windows Server, and Visual Studio) are the only three groups that make money. This is an extremely successful one hit pony, they have never made money at anything else.
It's funny that they're a one hit pony when you just listed 7 different products that they're making money on across 3 different divisions. So depending on how you're counting, MS is either a 7 trick or 3 trick pony.
Not to mention you've certainly left off a large number of products within those divisions (thinks like BizTalk server and the other stuff rolled up into Small Business Server that's available seperately; MS Money, Encarta and other consumer software. Games like Flight Simulator and other PC games MS publishes but doesn't create). Certainly these businesses make money otherwise MS would have dropped them a long time ago. They're profit is probably rolled up into one of the 3 divisions you've mentioned, but again, it shows that MS has a wide selection of software available. So depending on how you're counting, MS is either a more than 7 or 3 trick pony.
Now you might argue that the "one trick" is selling software, but then by that logic MOST companies are one trick ponies. WalMart's just doing it's one retailing trick. But anyway, MS recently posted a profit on MSN for the first time. So depending on how you're counting, MS is either a more than 7, 3 trick, or 2 trick pony.
The point of using Linux is that you can find out, and the OS will deliver you an honest list, unlike Windows, where you can have a daemon not listed in the Task window.
Try:
netstat -a -n -o | findstr LISTEN
The number on the right is the process ID. That process is listening on an open port. The port number comes after the first : on each line. Was that so hard?
Re:seen the price of VS.NET?
on
Kylix in Limbo
·
· Score: 1
First off, it's the CLR, not the CRL. Next, what's your definition of using the CLR to it's full capacity?
In C# you cannot do tailcalls, and you cannot do filter statements for exceptions (essentially where you conditionally execute a catch block). You also can't blend native & managed code in a single executable. In VB.NET you can do exception filters. But you don't get signed integers. In C++ you can mix native & managed code in a single executable. You also get explicit control over boxing which isn't available to C#. In Whidbey VB.NET will have Edit and Continue support, while C# won't. Not to mention that C# builds language construcuts that the CLR doesn't provde, and other languages do the same thing.
So there are differences between the languages. There are differences between the features of the CLR they expose. And if anything they are diverging, not converging.
Other articles, such as this one mention that this is about Avalon.
Avalon, for those of you who haven't been paying attention, is Microsoft's new UI programming paradigm for Longhorn. And it's all based, and written, on ".net". I use ".net" in quotes because what that really means is that it runs in the CLR and is written in C#, managed C++, VB, or anything else that compiles to IL such as XAML. Avalon has also been beendescribed as being much like ASP.NET w/ it's declarative markup syntax. So it's got that going for it too.
So you can be sure that the ".net" technologies are living on, even if the brand may be deprecated (which doesn't necessarily appear to be happening [VS.NET is still alive and well] but who can tell what the MS marketing machine will come up with next).
Windows has finally reached a stability that Linux/Unix/Mac has been boasting for a very long time.
Minor nitpick: Mac has not been boasting about stability for very long. Infact, I'd say Microsoft has been boasting about stability since NT 3.5. Mac has been boasting about stability since OS X. Of course that also means MS has actually been boasting about stability for at least as long as Linux. That leaves Unix which these days is a small player (Windows = ~50%, Linux = ~25%, everyone else fights for the remaining scraps). If you're stable and no one wants to use your product, what's it really matter?
The problem for a cracker is that they need a user's login and password, not just the root password; and even then that user has to be allowed to use the su command.
It's called Escalation of Privledge. Hackers don't need to have an account that is designed to become root. Hackers don't need the root password.
Hackers compromise a setuid program or a service that's only available locally. It just takes 1 user account. Traditionally there's a lot less remote exploits than there are local escalation of privledge attacks.
Except the meta-data rich world you describe still requires some sort of absolute referencing system, like say, filepaths. I agree that the user shouldn't always have to know the filepath, but I don't see any way for the computer itself to avoid having an index relating item (i.e. file) X to physical location Y.
I also don't see any reasonable way to make this list non-hierarchial -- if it was you'd have to search through the index, entry by entry, every time youwanted to look up anything. Again, the user doesn't have to see this, but sorted list with hierarchial entry points seem a very reasonable way to organize an index..
The computer certainly needs to index an item to a physical location. What the computer doesn't necessarily need to do is build up one true hierarchy of files & folders. All it really needs is an inode & it knows the file. That implies that in a database it has an inode column & a whole bunch of other metadata columns. You could imagine one of those is "Directory", and the directory could have a parent directory as well. But you could have multiple such hierarchies. Anyway, I suspect WinFS will still have one true hierarchy though (as NTFS isn't going away) - but if you truly had a database filesystem it seems reasonable to NOT have one true hierachy.
So they get to break all the existing applications - yet again.
I still daily run code that I first wrote on Unix 18 years ago, it has evolved at bit but is largely unchanged in 10 years.
It's funny you mention that, because they ran VisiCalc (search the article for VisiCalc) at the PDC. That's 20 years old, and it's a completely unmodified binary. Compare that to your 18 year old code, that's evolved, and is only "largely unchanged".
I'm sure you could find some software that could do that for Unix, but the hard part would probably be finding the Unix that has been on the same hardware for 20 years. It's not Linux, it's not Solaris; SGI maybe? They've been MIPS all the way I believe.
Anyway, the point being that WinFX, like every other major Microsoft upgrade, is backwards compatible. You like OLE? DDE? Win16? DOS? You can still do it all & run it all on Windows today and apparently tommorow too.
But the thing is, and what I think the original poster was saying, that doesn't seem to be the case. Microsoft appears to have a form that you fill out saying that you want to license something, and then they "get back to you" about the fees. That doesn't sound like the behavior of an entity that intends to use the same royalty schedule for IBM and SCO. In fact, it sounds suspiciously like an entity that's trying to avoid licensing its technology at all, except to carefully-chosen partners. (As "proof" that its complying)
The Justice department (and the enforcement panel) can certainly inspect the fees that are being charged here.
Also convenient that you have to sign an NDA to look at the protocol list. This prevents you from telling the public, for example, that there are no protocols on the list. Or that Microsoft wanted to know what you'd use it for before giving you pricing information. Or any one of a dozen other violations.
First, given that there are licencees, obviously there are protocols to be licensed. And there's no doubt that the Justice department & friends have seen the list. Obviously this process is going to be looked at very closely by the Justice department. Now the Judge wants an even CLOSER look. If your allegations were true they would be violating the terms, and I think we'd see something significant. What, do you think no one's paying any attention to what Microsoft's doing these days?
I agree that platforms themselves should be free from control.
Being that your goal is to write closed source software on the open source platform you should watch out for KDE. It's GPL only, so if you link against it (e.g., write any KDE app) you have to go GPL as well. You can of course buy a Qt license to make your software non-free, but you still can't link against KDE! So you can only write Qt apps, not KDE apps.
Gnone & GTK don't have this GPL problem (it's LGPL) so you're free to control your own source code. So you can write Gnome apps that are closed source. Is it any wonder that most large companies support Gnome over KDE?
And why would no one want to license SMB or CIFS? Is it maybe because Samba offers a (nearly?) complete implementation for free? And that Microsoft has done nothing to prevent this? Nahh... that couldn't be the reason at all.
Imagine if Samba didn't exist: There'd be a lot more people clamoring at the door to license these protocols. They'd be able to sell a product for Unix/Windows interoperability and make a bunch of money. Obviously there's demand for it.
The Bush administration may have changed the DOJ's perspective on the case, but ultimately, it's up to the judge, not the DOJ, to determine what the actual penalty would be, regardless of the DOJ's opinion
Even this is debatable. It's easy to forget the finer points of the long legal battle, but Microsoft has been involved in negotiations with the Department of Justice many times before Bush was in office. A great book that covers this is World War 3.0 (unfortunately I read the original, not this newly updated copy).
Anyway, to make a long story short, there was a point when Microsoft had made an agreement with the Clinton administration. It was the STATES that blocked it; and of course it was the states that didn't sign on to the final agreement.
So while everyone constantly moans about Bush settling with Microsoft it was really the case that the states steadfastly refused to do so (and it remains that way to this day). So that change on perspective was at most "fuck the other states" rather than "fuck prosecuting Microsoft."
At least they're trying to shake things up on the software front, with their Java Enterprise System, which costs companies only $100 per user.
You left off the "per year" on that cost. 3 years is a fast upgrade cycle. 4 years is probably the most common, but many companies still take longer than that.
Uhh, let's see. They're talking about how Mac OS X does what they want (bidi) and apparently enough people use it to complain that Office doesn't support this on OS X. They are also contemplating a switch to OpenOffice.
Remind me, what product is it that Microsoft holds monopoly power for again? From this it certainly doesn't sound like it's operating systems or office productivity software.
You are both correct and incorrect about this. Opera does mimic the beginning of MSIE's user agent (eg, "Mozilla/4.0 (compatible; MSIE 6.0; MSIE 5.5; Windows 98)") but it does ALSO include an Opera specific portion of the user agent string (the full string is "Mozilla/4.0 (compatible; MSIE 6.0; MSIE 5.5; Windows 98) opera 7.0 [en]").
So most websites that are MSIE aware but not Opera aware will identify the browser as being MSIE. But a web site can still be Opera aware. I don't know if Google is doing this or not (I suspect not as they don't call out Opera in their stats) but they could be lumping it in with Other.
(I wonder if I'll get called a jack-monkey for posting raw facts again):)
I know I shouldn't feed the trolls, but I can't resist, you're just too stupid.
Having a capacity for intelligent conversation obviously excludes yourself. At least I can figure out how to create an account.
What the hell is a jack-monkey anyway? Insults don't do much good if they're not universally understood fuck head.
Uhh, check out Google's Zeitgeist. It includes browser usage stats, and just about everyone is using IE. I'm sure you're aware of this and were just trying to be pedantic, but you're just being stupid.
Not only do many of the respondents take the attitude of it beats everything else, but this is also a non-scientific self selecting survey:
From the article: The Reader Question is by no means scientific; we solicit responses on InternetWeek.com and in the newsletter and make no effort to qualify respondents. QuickPolls are the Internet journalism equivalent of man-in-the-street interviews and call-in talk radio.
And who is reading InternetWeek.com? I'm betting they don't see many AOL users, secretary's, accountants, or other "non-computer people". Rather these are people who are probable intimiately familiar with computers. For your typical user (let alone from a scientific poll) the answers may be completely different.
The first comment in the article even mentions this: Only 5% of their users would be "good candidates" of using Linux on the desktop.
SGI has remedied the *MINOR* breaches it previously had, and is free. Its own analysis proved that.
They gave the code to the entire world. Can I still go to kernel.org and get an old kernel with the code? I suspect the answer is yes, and if it isn't, I suspect the code is floating around somewhere still.
How do you remedy that?
The very fact that Unix-like OSs have a concept of a "root" account (which the Windows "equivalent", "administrator", does not even come CLOSE to matching in terms of actual separation of permissions), makes it all but invincible to virii.
Lots of other people have already corrected you on many points, but no one commented on "seperation of permission" so I'll go ahead and tell you why you're wrong about this one;)
"root" gets full control of your box. In Unix-like OSes there is no "half-root" that can only do some super powerful things, and you can't easily & efficiently break out security roles.
Certainly "everything is a file" (except for when it isn't) so you can have a group own a file, and have multiple users be in that group. For example you can't have one user have full control, another user have only read access, another user only have write access, and a fourth user have both read&write access, while everyone else has no access. It's impossible in traditional Unix (some Unix's do support ACLs, but it's certainly not a standard feature). I imagine many syscalls are probably (although I haven't investigated) root or not root.
NT, on the other hand, does actually contain more seperation of privledges. Certainly there is still an Administrators group, and anyone belonging to this group has full access to the machine. NT uses security groups to give permission sets to users. So far we sound like Unix. But under NT the groups are used for providing broad different types of access. To list a few: Debugging Users, DNS admin, DHCP admin, certificate publishers, group policy admin, and I'm sure there's many more. And with NT's system you never get into a jam where you can't assign the permission combinations you want. ACLs provide all the flexibility you want.
And you can also take steps to provide interesting degrees of flexibility. Let me give you an example: I could make a InstallsNewSoftware group that has write access to %PROGRAMFILES% & access to write to a subset of the registry. Hell, I could even NOT give them the ability to delete files in %PROGRAMFILES%.
Under those conditions most programs would install. And guess what? Windows XP ships with a power users group which is like this, and even better. It also allows power users to create new user accounts & modify the accounts they've created (interestingly NOT the accounts they didn't create). They also get the ability to update settings in the control panel such as printers, date/time, & power options.
But they can't mess up the operating system. That's a pretty heavy set of privledge seperation that ships out of the box in a 2 year old Microsoft product. It beats anything the major Linux distributors ship today. Hopefully you can now affort to go and buy a clue:)
As much as I detest Microsoft, Motorola has made the terrible decision of putting out a single MS powered phone (for Orange SA in Europe). Motorola then proceeded to restate that they're firmly commited to Java on Linux for future phones. I know Orange is an MS pawn but it's still disappointing to see Motorola lending credence to it.
Not only are they going to ship an MS phone, but they're also pulling out of Symbian. Symbian is quickly turning into an all Nokia game, and as it does that it's going to be harder and harder for 3rd parties to justify using their competitors software. Microsoft of course doesn't compete against cell phone vendors, so the pressure to ship MS phones will go up.
That's not to imply that the pressure to ship Linux phones won't go up too, but I suspect there's a lot more work for a cell phone vendor to convert Linux into a cell phone OS than MS Smartphone software. I say that simply because MS's target is cell phone's and they can design for it, that's harder for the open source community to do simply bceause they don't have the hardware.
Re:Dead? Well, probably not. Mostly.
on
Java vs .NET
·
· Score: 1
The thing to remember about.NET is that it's really Microsoft's Java repackaged in a form that Sun can't sue them over. It has most of the JDK 1.1 libraries pretty much intact but with package and method names changed. But they did a very nice job in fixing a lot of issues that Java has.... In particular.NET's assembly management beats the heck out of collections of jar files.
There's many other differences between.NET & Java than just this one. If you look closely at.NET you can see where it evolves from. For example the assembly system you mention evolves from Fusion.
Really the right way to think of.NET is Microsoft's evolution of COM/COM+". Part of.NET is the COM+ Runtime. One of the great things about.NET mentioned here is metadata. Metadata is what makes.NET web services rock.
A good consice article discusses Visual Studio 7, COM+2.0, Fusion 2.0, etc...
What isn't ever mentioned in any of these articles of course is a JIT. And of course IL is a stack based language just like Java bytecodes, so it appears Microsoft took a page from Sun on this.
And of course anyone familar with Microsoft's UI (WFC) for Java will know that WinForms is very similar. But once again this is something that's pure Microsoft.
So really when you get down to it the similarities between the two just consist of the JIT and many of the base class libraries. But how much can console I/O, sockets, files, and other basic things differ? Everything else in.NET is pure Microsoft.
There probably aren't too many people out there that know whether there's any actual code shared, and of course no one at Microsoft is saying. But when you look at the publicly documented evolution, and you look at the pretty significant differences, it becomes obvious that.NET has it's own unique history despite superficial appearances of similarity.
Slashdot Mirror
Also note that only Client (Windows and CALs), Information Worker (Office, Project, Etc) and Server and Tools (Exchange, SQL, Windows Server, and Visual Studio) are the only three groups that make money. This is an extremely successful one hit pony, they have never made money at anything else.
It's funny that they're a one hit pony when you just listed 7 different products that they're making money on across 3 different divisions. So depending on how you're counting, MS is either a 7 trick or 3 trick pony.
Not to mention you've certainly left off a large number of products within those divisions (thinks like BizTalk server and the other stuff rolled up into Small Business Server that's available seperately; MS Money, Encarta and other consumer software. Games like Flight Simulator and other PC games MS publishes but doesn't create). Certainly these businesses make money otherwise MS would have dropped them a long time ago. They're profit is probably rolled up into one of the 3 divisions you've mentioned, but again, it shows that MS has a wide selection of software available. So depending on how you're counting, MS is either a more than 7 or 3 trick pony.
Now you might argue that the "one trick" is selling software, but then by that logic MOST companies are one trick ponies. WalMart's just doing it's one retailing trick. But anyway, MS recently posted a profit on MSN for the first time. So depending on how you're counting, MS is either a more than 7, 3 trick, or 2 trick pony.
The point of using Linux is that you can find out, and the OS will deliver you an honest list, unlike Windows, where you can have a daemon not listed in the Task window.
Try:
netstat -a -n -o | findstr LISTEN
The number on the right is the process ID. That process is listening on an open port. The port number comes after the first : on each line. Was that so hard?
First off, it's the CLR, not the CRL. Next, what's your definition of using the CLR to it's full capacity?
In C# you cannot do tailcalls, and you cannot do filter statements for exceptions (essentially where you conditionally execute a catch block). You also can't blend native & managed code in a single executable. In VB.NET you can do exception filters. But you don't get signed integers. In C++ you can mix native & managed code in a single executable. You also get explicit control over boxing which isn't available to C#. In Whidbey VB.NET will have Edit and Continue support, while C# won't. Not to mention that C# builds language construcuts that the CLR doesn't provde, and other languages do the same thing.
So there are differences between the languages. There are differences between the features of the CLR they expose. And if anything they are diverging, not converging.
Other articles, such as this one mention that this is about Avalon.
Avalon, for those of you who haven't been paying attention, is Microsoft's new UI programming paradigm for Longhorn. And it's all based, and written, on ".net". I use ".net" in quotes because what that really means is that it runs in the CLR and is written in C#, managed C++, VB, or anything else that compiles to IL such as XAML. Avalon has also been been described as being much like ASP.NET w/ it's declarative markup syntax. So it's got that going for it too.
So you can be sure that the ".net" technologies are living on, even if the brand may be deprecated (which doesn't necessarily appear to be happening [VS.NET is still alive and well] but who can tell what the MS marketing machine will come up with next).
Windows has finally reached a stability that Linux/Unix/Mac has been boasting for a very long time.
Minor nitpick: Mac has not been boasting about stability for very long. Infact, I'd say Microsoft has been boasting about stability since NT 3.5. Mac has been boasting about stability since OS X. Of course that also means MS has actually been boasting about stability for at least as long as Linux. That leaves Unix which these days is a small player (Windows = ~50%, Linux = ~25%, everyone else fights for the remaining scraps). If you're stable and no one wants to use your product, what's it really matter?
The problem for a cracker is that they need a user's login and password, not just the root password; and even then that user has to be allowed to use the su command.
It's called Escalation of Privledge. Hackers don't need to have an account that is designed to become root. Hackers don't need the root password.
Hackers compromise a setuid program or a service that's only available locally. It just takes 1 user account. Traditionally there's a lot less remote exploits than there are local escalation of privledge attacks.
Except the meta-data rich world you describe still requires some sort of absolute referencing system, like say, filepaths. I agree that the user shouldn't always have to know the filepath, but I don't see any way for the computer itself to avoid having an index relating item (i.e. file) X to physical location Y.
I also don't see any reasonable way to make this list non-hierarchial -- if it was you'd have to search through the index, entry by entry, every time youwanted to look up anything. Again, the user doesn't have to see this, but sorted list with hierarchial entry points seem a very reasonable way to organize an index..
The computer certainly needs to index an item to a physical location. What the computer doesn't necessarily need to do is build up one true hierarchy of files & folders. All it really needs is an inode & it knows the file. That implies that in a database it has an inode column & a whole bunch of other metadata columns. You could imagine one of those is "Directory", and the directory could have a parent directory as well. But you could have multiple such hierarchies. Anyway, I suspect WinFS will still have one true hierarchy though (as NTFS isn't going away) - but if you truly had a database filesystem it seems reasonable to NOT have one true hierachy.
So they get to break all the existing applications - yet again.
I still daily run code that I first wrote on Unix 18 years ago, it has evolved at bit but is largely unchanged in 10 years.
It's funny you mention that, because they ran VisiCalc (search the article for VisiCalc) at the PDC. That's 20 years old, and it's a completely unmodified binary. Compare that to your 18 year old code, that's evolved, and is only "largely unchanged".
I'm sure you could find some software that could do that for Unix, but the hard part would probably be finding the Unix that has been on the same hardware for 20 years. It's not Linux, it's not Solaris; SGI maybe? They've been MIPS all the way I believe.
Anyway, the point being that WinFX, like every other major Microsoft upgrade, is backwards compatible. You like OLE? DDE? Win16? DOS? You can still do it all & run it all on Windows today and apparently tommorow too.
But the thing is, and what I think the original poster was saying, that doesn't seem to be the case. Microsoft appears to have a form that you fill out saying that you want to license something, and then they "get back to you" about the fees. That doesn't sound like the behavior of an entity that intends to use the same royalty schedule for IBM and SCO. In fact, it sounds suspiciously like an entity that's trying to avoid licensing its technology at all, except to carefully-chosen partners. (As "proof" that its complying)
The Justice department (and the enforcement panel) can certainly inspect the fees that are being charged here.
Also convenient that you have to sign an NDA to look at the protocol list. This prevents you from telling the public, for example, that there are no protocols on the list. Or that Microsoft wanted to know what you'd use it for before giving you pricing information. Or any one of a dozen other violations.
First, given that there are licencees, obviously there are protocols to be licensed. And there's no doubt that the Justice department & friends have seen the list. Obviously this process is going to be looked at very closely by the Justice department. Now the Judge wants an even CLOSER look. If your allegations were true they would be violating the terms, and I think we'd see something significant. What, do you think no one's paying any attention to what Microsoft's doing these days?
I agree that platforms themselves should be free from control.
Being that your goal is to write closed source software on the open source platform you should watch out for KDE. It's GPL only, so if you link against it (e.g., write any KDE app) you have to go GPL as well. You can of course buy a Qt license to make your software non-free, but you still can't link against KDE! So you can only write Qt apps, not KDE apps.
Gnone & GTK don't have this GPL problem (it's LGPL) so you're free to control your own source code. So you can write Gnome apps that are closed source. Is it any wonder that most large companies support Gnome over KDE?
And why would no one want to license SMB or CIFS? Is it maybe because Samba offers a (nearly?) complete implementation for free? And that Microsoft has done nothing to prevent this? Nahh... that couldn't be the reason at all.
Imagine if Samba didn't exist: There'd be a lot more people clamoring at the door to license these protocols. They'd be able to sell a product for Unix/Windows interoperability and make a bunch of money. Obviously there's demand for it.
The Bush administration may have changed the DOJ's perspective on the case, but ultimately, it's up to the judge, not the DOJ, to determine what the actual penalty would be, regardless of the DOJ's opinion
Even this is debatable. It's easy to forget the finer points of the long legal battle, but Microsoft has been involved in negotiations with the Department of Justice many times before Bush was in office. A great book that covers this is World War 3.0 (unfortunately I read the original, not this newly updated copy).
Anyway, to make a long story short, there was a point when Microsoft had made an agreement with the Clinton administration. It was the STATES that blocked it; and of course it was the states that didn't sign on to the final agreement.
So while everyone constantly moans about Bush settling with Microsoft it was really the case that the states steadfastly refused to do so (and it remains that way to this day). So that change on perspective was at most "fuck the other states" rather than "fuck prosecuting Microsoft."
At least they're trying to shake things up on the software front, with their Java Enterprise System, which costs companies only $100 per user.
You left off the "per year" on that cost. 3 years is a fast upgrade cycle. 4 years is probably the most common, but many companies still take longer than that.
Microsoft is a monopoly.
Uhh, let's see. They're talking about how Mac OS X does what they want (bidi) and apparently enough people use it to complain that Office doesn't support this on OS X. They are also contemplating a switch to OpenOffice.
Remind me, what product is it that Microsoft holds monopoly power for again? From this it certainly doesn't sound like it's operating systems or office productivity software.
fuck head.
You are both correct and incorrect about this. Opera does mimic the beginning of MSIE's user agent (eg, "Mozilla/4.0 (compatible; MSIE 6.0; MSIE 5.5; Windows 98)") but it does ALSO include an Opera specific portion of the user agent string (the full string is "Mozilla/4.0 (compatible; MSIE 6.0; MSIE 5.5; Windows 98) opera 7.0 [en]"). So most websites that are MSIE aware but not Opera aware will identify the browser as being MSIE. But a web site can still be Opera aware. I don't know if Google is doing this or not (I suspect not as they don't call out Opera in their stats) but they could be lumping it in with Other. (I wonder if I'll get called a jack-monkey for posting raw facts again) :)
I know I shouldn't feed the trolls, but I can't resist, you're just too stupid. Having a capacity for intelligent conversation obviously excludes yourself. At least I can figure out how to create an account. What the hell is a jack-monkey anyway? Insults don't do much good if they're not universally understood fuck head.
Uhh, check out Google's Zeitgeist. It includes browser usage stats, and just about everyone is using IE. I'm sure you're aware of this and were just trying to be pedantic, but you're just being stupid.
But it doesn't work if you have the latest patches.
Not only do many of the respondents take the attitude of it beats everything else, but this is also a non-scientific self selecting survey:
From the article:
The Reader Question is by no means scientific; we solicit responses on InternetWeek.com and in the newsletter and make no effort to qualify respondents. QuickPolls are the Internet journalism equivalent of man-in-the-street interviews and call-in talk radio.
And who is reading InternetWeek.com? I'm betting they don't see many AOL users, secretary's, accountants, or other "non-computer people". Rather these are people who are probable intimiately familiar with computers. For your typical user (let alone from a scientific poll) the answers may be completely different.
The first comment in the article even mentions this: Only 5% of their users would be "good candidates" of using Linux on the desktop.
SGI has remedied the *MINOR* breaches it previously had, and is free. Its own analysis proved that.
They gave the code to the entire world. Can I still go to kernel.org and get an old kernel with the code? I suspect the answer is yes, and if it isn't, I suspect the code is floating around somewhere still. How do you remedy that?
The very fact that Unix-like OSs have a concept of a "root" account (which the Windows "equivalent", "administrator", does not even come CLOSE to matching in terms of actual separation of permissions), makes it all but invincible to virii.
;)
:)
Lots of other people have already corrected you on many points, but no one commented on "seperation of permission" so I'll go ahead and tell you why you're wrong about this one
"root" gets full control of your box. In Unix-like OSes there is no "half-root" that can only do some super powerful things, and you can't easily & efficiently break out security roles.
Certainly "everything is a file" (except for when it isn't) so you can have a group own a file, and have multiple users be in that group. For example you can't have one user have full control, another user have only read access, another user only have write access, and a fourth user have both read&write access, while everyone else has no access. It's impossible in traditional Unix (some Unix's do support ACLs, but it's certainly not a standard feature). I imagine many syscalls are probably (although I haven't investigated) root or not root.
NT, on the other hand, does actually contain more seperation of privledges. Certainly there is still an Administrators group, and anyone belonging to this group has full access to the machine. NT uses security groups to give permission sets to users. So far we sound like Unix. But under NT the groups are used for providing broad different types of access. To list a few: Debugging Users, DNS admin, DHCP admin, certificate publishers, group policy admin, and I'm sure there's many more. And with NT's system you never get into a jam where you can't assign the permission combinations you want. ACLs provide all the flexibility you want.
And you can also take steps to provide interesting degrees of flexibility. Let me give you an example: I could make a InstallsNewSoftware group that has write access to %PROGRAMFILES% & access to write to a subset of the registry. Hell, I could even NOT give them the ability to delete files in %PROGRAMFILES%.
Under those conditions most programs would install. And guess what? Windows XP ships with a power users group which is like this, and even better. It also allows power users to create new user accounts & modify the accounts they've created (interestingly NOT the accounts they didn't create). They also get the ability to update settings in the control panel such as printers, date/time, & power options.
But they can't mess up the operating system. That's a pretty heavy set of privledge seperation that ships out of the box in a 2 year old Microsoft product. It beats anything the major Linux distributors ship today. Hopefully you can now affort to go and buy a clue
As much as I detest Microsoft, Motorola has made the terrible decision of putting out a single MS powered phone (for Orange SA in Europe). Motorola then proceeded to restate that they're firmly commited to Java on Linux for future phones. I know Orange is an MS pawn but it's still disappointing to see Motorola lending credence to it.
Not only are they going to ship an MS phone, but they're also pulling out of Symbian. Symbian is quickly turning into an all Nokia game, and as it does that it's going to be harder and harder for 3rd parties to justify using their competitors software. Microsoft of course doesn't compete against cell phone vendors, so the pressure to ship MS phones will go up.
That's not to imply that the pressure to ship Linux phones won't go up too, but I suspect there's a lot more work for a cell phone vendor to convert Linux into a cell phone OS than MS Smartphone software. I say that simply because MS's target is cell phone's and they can design for it, that's harder for the open source community to do simply bceause they don't have the hardware.
The thing to remember about .NET is that it's really Microsoft's Java repackaged in a form that Sun can't sue them over. It has most of the JDK 1.1 libraries pretty much intact but with package and method names changed. But they did a very nice job in fixing a lot of issues that Java has. ... .NET's assembly management beats the heck out of collections of jar files.
.NET & Java than just this one. If you look closely at .NET you can see where it evolves from. For example the assembly system you mention evolves from Fusion.
.NET is Microsoft's evolution of COM/COM+". Part of .NET is the COM+ Runtime. One of the great things about .NET mentioned here is metadata. Metadata is what makes .NET web services rock.
.NET is pure Microsoft.
.NET has it's own unique history despite superficial appearances of similarity.
In particular
There's many other differences between
Really the right way to think of
A good consice article discusses Visual Studio 7, COM+2.0, Fusion 2.0, etc...
What isn't ever mentioned in any of these articles of course is a JIT. And of course IL is a stack based language just like Java bytecodes, so it appears Microsoft took a page from Sun on this.
And of course anyone familar with Microsoft's UI (WFC) for Java will know that WinForms is very similar. But once again this is something that's pure Microsoft.
So really when you get down to it the similarities between the two just consist of the JIT and many of the base class libraries. But how much can console I/O, sockets, files, and other basic things differ? Everything else in
There probably aren't too many people out there that know whether there's any actual code shared, and of course no one at Microsoft is saying. But when you look at the publicly documented evolution, and you look at the pretty significant differences, it becomes obvious that