Within every DRM system there will always be a way for the author to set the copy rights to allow freely made copies.
Don't count on it. Any such system means that somebody can "pirate" a copyrighted work by breaking the copy protection and recording the result, or filming off the display screen, or recording their own neurons, or whatever is needed to get around the DRM. Thus as long as somebody can make "non DRM" copies piracy is possible.
The biggest worry is that there will be no recording devices that can record without DRM, because such devices can be used for piracy. Yes, "old" devices will work, but this may be useless if the average person only owns a "new" playback machine because that is the only thing they can play popular movies on. The average person will be able to make their home movies but nobody can view them unless the original camera is attached to the Internet to authorize it, and the average person will accept this as normal and be perfectly satisfied.
The fact that unrestricted recording devices may very well be used for piracy 99% of the time will make it easy for them to convince everybody it is a good idea to get rid of them. The fact that the other 1% is *EVERY SINGLE BIT OF FREE SPEECH IN THE WORLD* will be ignored.
This was the purpose of patents, to encourage inventors to publicize their inventions without fear of them being stolen. In theory patents apply to things the author could still profit from by keeping it secret. Copyrights supposedly apply to things that you can only make money from by making it known.
On BeOS, you can use the *SAME* filename to open the same file whether you use the POSIX or the BeOS interface. This was not true of the NT "POSIX".
Love how you weasel out of the truth. Microsoft figured out a way to claim "POSIX compatability" without actually making it work, because the interoperability was ZERO with all the Windows applications. Rather than bragging about working on it, you should be ashamed of yourself.
The IBM PC had already been out for 2 years before 1985, you know. And there were CP/M machines in homes before that since 1977. And computers that could do word processing of some form existed since 1960.
I suppose though you think Bill Gates invented it all. But even then, he had already sold quite a bit of software, including MSDOS and Basic, for years by then.
This is wrong - "child" windows do not always float on the parent (sadly - the amazing vanishing dialog box remains a hassle in Gtk apps to this day)
This sounds like a GTK bug. I have never seen any X window manager developed in the last 10 years ignore the TRANSIENT_FOR hint. And on Windows it certainly obeys this rule.
Focus policy in X is the responsibility of the WM, and all the major ones allow you to switch off raise on click
The problem is that the program cannot *assumme* the raise-on-click is turned off. It is *TRIVIAL* to do raise-on-click yourself if you know the WM is not doing it. The opposity is physically impossible. For this reason the WM should NEVER do raise-on-click.
He's talking about the fact that you cannot delete a file while it is in use by another program.
Nothing to do with record locking, which I agree is a mess on both Unix and Windows, though increasingly irrelevant as most programs just treat files as atomic objects.
Although a lot better than Windows in that deletion and mv work atomically, Unix also has problems with the file locking he was talking about. Writing a file atomically is difficult, involving the need to figure out a temporary name, and to mv the result after closing, but only if the closing and writing is successful, and not mess up if the original file has unexpected permission or owner (or other metadata...). It would be nice if Linux started looking at changing things, and ignoring back-compatability with old record-oriented systems. As I see it there is only a need to write files with a "create" call, and the file is not visible to any other program until it is successfully closed. This would also get rid of all the race and security problems with temp files. I kind of doubt this will ever happen though, Linux seems stuck with complete panic about being back-compatable, while Microsoft is just sure to screw it up and make a complex interface...
No, "creation date", "modification date", and "name" are *also* metadata! Even the "size" is metadata (actually the array of data and length that most people consider the file's "contents" is another piece of metadata).
One of the big worries about WinFS is that the idiots at Microsoft will force something on us with a whole new API because they fail to realize that all data can be named/accessed through a single API. At the API level all the metadata is the same and it should be possible to remove the creation date (you may not have permission to do this, however).
The only solution on Windows and X is to use what Windows calls "child" windows (X calls them "transient" windows). Each window has a "parent" window and the child is always atop the parent window. If the user clicks in the docuement/parent window to raise it, all the toolbars (which are children of it) are raised as well.
Now this is far from perfect, but it is the only thing you get out of X or Windows that works reliably. The most obvious problem is that there is only one parent. This makes using the same toolbars for more than one document impossible (on X you can hack in the ability to change the parent once the window exists, but you cannot do this on Windows).
The real solution is to *STOP RAISING THE WINDOWS ON CLICK!!!!!* Unfortunatly the designers of Windows and X are way too stupid and set in their ways to ever realize this. But the truth is that it is trivial for a program to raise it's own window in response to a click, and it can thus enforce any window ordering it wants.
I'm actually suprised the fake zip code works. This is done to try to prevent stolen credit cards from working. Gas pumps are doing this too, and I have started seeing this in stores. You need to enter the zip code that is on the billing address of the credit card. This is because it is a piece of information the card holder probably knows that is not printed on the credit card.
Trust me if they want to identify you they already have your entire credit card number, they don't need your zip code!
My guess: most of the time only one thing is being read from the disk at a time (yes several files can be open and reading, but I would suspect that only one being read right at a moment is very common). Therefore it seems the best use of two heads is to have them read 1/2 of the same track. This means that the duplicate positioning machinery is being used to place the heads at exactly the same track. If instead you added another platter or platter side, you could reuse the positioning machinery and get the exact same speed increase for transfers. Adding more platters also doubles the capacity, so this is a big win over multiple heads.
Of course 2 heads halves the rotational delay to aquire a piece of data. So it does have some wins. But maybe not enough.
I believe they are talking about people who *can* login as root. Non "administrators" cannot login as root. Administrators must still do something and know a password to do anything dangerous to the system.
Because installation still asks for a root password, I am confused whether this has the sudo method where the user uses their own password or the root password to get root privledges.
Like a previous poster said, it would be better if the installer just shut up rather than printing confusing information messages. There is no reason to mention "administrator" or not. Or put a "adminsitrator" checkmark on the panel with a tooltip that says "this user may sudo...".
In case anybody is confused, I believe he is talking about termanal programs. The Delete key works fine in Qt/KDE.
The basic tty driver of Linux is unchanged from ancient Unix ones and probably cannot do much. It actually cannot do any editing other than backspace, thus it is impossible to get the cursor between characters so that delete-forward can do anything. However I would like to see them scrap the entire stty stuff and just hardcode it so *both* delete and backspace and also any common escape sequences from these keys deletes backwards. An awful lot of user annoyance can be fixed by changing from "what action does this" to "what does this action do" in all the code and configuration, so that there is more than one way to do things.
Many terminal programs such as all modern shells and programs using readline and similar put the tty in raw mode and thus can do any editing that can be achieved without any vertical movement. However I checked and lots of them don't recognize the Delete key (ie they don't recognize the escape sequence they send). This should all be fixed the same way as the tty drivers. Ie scrap any "configuration" and hard-code in all the possible sequences the keys may produce and make them all work.
I would say the warnings are quite equivalent. Both the WIndows and BSD warnings are bad, and like you said they are from engineers who panic that the user may not know something.
Your Windows warnings are just as bad and I am suprised you think they are clearer to the user.
Warning -another version of Windows was detected!
How about "It looks like Windows is already installed. Should I replace it?"
Warning - are you sure you want to delete this partiton?
What's a partition? If I don't delete it is the installation going to work? Probably better is "to install Windows a part of your disk will be erased and replaced with Windows. This is what is there now (some info about whatever is on the partion), is it ok to delete and replace this? If not there are (number) other places that Windows can be put..."
Warning - your hard drive is blank (yes it really does "WARN" you of this)
A huge number of users are going to say "oh no my hard disk is faulty I need to replace it?"
I would agree that all these stupid messages (and the BSD ones) should be removed. Especially if there is no alternative.
They are talking about the same thing that OS/X has, where only "administrators" can login/sudo to do things as root. They still have to type their *own* password in order to aquire root privledges. Non-administrators cannot do anything as root.
I believe the same thing can be achieved on Linux with sudo and removing the root password and useing sudoers to select who is an "administrator". But I'm not sure about that. If so it really should be the default behavior.
Just want to repeat what another poster said. The above post is false. Assumming there really is GPL code in SCO's kernel:
1. They *may* have to pay damages, if somebody sues them and wins. This is legally possible.
2. Under *NO* possible scenario do they "have" to open-source their kernel. This is FUD straight from Bill Gates.
3. Even if they *do* open-source the kernel, they are *still* liable for damages, because they were violating the copyright before they open-sourced it. Otherwise there would be a huge loophole in the GPL (just wait until the code is uselessly old and release it and you are absolved of all copyright violations?)
It is true that companies often decide to open source some piece of code that they put GPL code into, in exchange for a promise to drop any pending lawsuits, and/or just for good public relations. However there is absolutely no legal requirement that they do this, and doing it does not put them in a legally better situation than before.
Last time I checked 3D Studio Max and Photoshop have Mac OS/X versions.
You are one of the most dangerous clueless individuals I have ever seen. Very very scary. You seem to be taking the very fact that Microsoft is a monopoly and able to dictate it's own rules and destroy interoperability, and word it as though that is an *advantage*. Wow. The mind boggles.
The GPL allows me to retain control of my code. It is extremely capitalist. Programmers use it exclusively for reasons of Ayn-Randian greed. Look up "dual licensing" if you still don't have a clue. In today's world writing GPL software is the ONLY way to distribute software that you really own in a way where you may potentially make a profit or get credits or a job, unless you are the head of one of the dozen or so very large software firm. This is the real reason Microsoft attacks the GPL so much, it is really it's only competition.
"Communist" would apply a bit better to BSD code, as supposedly all your work is contributed to everybody to do what they want with it, though none of it really applies that well.
Other comparisons of Microsoft to controlled economies to reverse the argument are fun but irrelevant, as the original poster pointed out. Supposedly "ideal communisim" would not require a controlled economy.
Too many people complain that Linux users say "go fix it yourself, you have the source". They completely miss the fact that along with that solution there is "pay somebody to fix it". This idea that code will not run on Linux unless it is free is the biggest piece of FUD out there.
Yes, one response is "go write it yourself, you have the source". Another response is "pay for it to get written!". I think the complaint is that for some reason people expect everything to be given to them for free. Nobody goes to Microsoft and demands a feature for free.
There appears to be a belief that software will not work on Linux unless it is free, as though that is some technical limitation or requirement. This is very annoying to companines and people like me who hope to sell stuff for Linux. Every time somebody complains with "there will never be kitchen-design software for Linux because nobody will write it for free" they are buying into this FUD. You are doing the same thing.
IT IS POSSIBLE TO WRITE COMMERCIAL SOFTWARE FOR LINUX!!!
Okay. The big point is that *direct X drivers are broken too!*. You should try to point that out.
I would like confirmation that in fact it is possible to write an OpenGL LDDM driver and that it is not significantly more difficult than writing a DirectX one.
Otherwise I certainly can't blame Microsoft for making their software OpenGL simulation call their software DirectX simulation. It would in fact give some DirectX acceleration to OpenGL (though contrary to claims this already existed, the OpenGL software library calls the same driver API as DirectX does in XP).
One area they could go a long way with is to release this OpenGL library as open source, so that API's that they don't care about (ie OpenGL > 1.4) can be added.
The claim is that the hardware manufacturer no longer can make an OpenGL driver. It is impossible to pass OpenGL calls though to the driver, you must translate to Direct3D.
I don't know if this really is true however. Some Microsoft fanboy keeps claiming that the driver must be rewitten to the "LDDL" spec or "to support virtualization" but there is no backup information on this, especially an explanation as to why this is not required for existing Direct3D drivers.
Maya and every other advanced graphics appliation out there uses OpenGL and also uses the windowing system. This will seriously impact them. If this story is really true (and I have my doubts) then this is a direct attempt to force all high level appliations to switch to DirectX, thus making portable vesions very difficult or impossible.
Unlike office-type applications, these high-level programs are very popular on Linux and Mac. I think 50% of our sales are Linux. Yet we are very small, if OpenGL is not supported on Windows, we will have to drop half our market (hard to say which, though it is obvious Microsoft hopes we will drop Linux).
Slashdot Mirror
Within every DRM system there will always be a way for the author to set the copy rights to allow freely made copies.
Don't count on it. Any such system means that somebody can "pirate" a copyrighted work by breaking the copy protection and recording the result, or filming off the display screen, or recording their own neurons, or whatever is needed to get around the DRM. Thus as long as somebody can make "non DRM" copies piracy is possible.
The biggest worry is that there will be no recording devices that can record without DRM, because such devices can be used for piracy. Yes, "old" devices will work, but this may be useless if the average person only owns a "new" playback machine because that is the only thing they can play popular movies on. The average person will be able to make their home movies but nobody can view them unless the original camera is attached to the Internet to authorize it, and the average person will accept this as normal and be perfectly satisfied.
The fact that unrestricted recording devices may very well be used for piracy 99% of the time will make it easy for them to convince everybody it is a good idea to get rid of them. The fact that the other 1% is *EVERY SINGLE BIT OF FREE SPEECH IN THE WORLD* will be ignored.
This was the purpose of patents, to encourage inventors to publicize their inventions without fear of them being stolen. In theory patents apply to things the author could still profit from by keeping it secret. Copyrights supposedly apply to things that you can only make money from by making it known.
On BeOS, you can use the *SAME* filename to open the same file whether you use the POSIX or the BeOS interface. This was not true of the NT "POSIX".
Love how you weasel out of the truth. Microsoft figured out a way to claim "POSIX compatability" without actually making it work, because the interoperability was ZERO with all the Windows applications. Rather than bragging about working on it, you should be ashamed of yourself.
The IBM PC had already been out for 2 years before 1985, you know. And there were CP/M machines in homes before that since 1977. And computers that could do word processing of some form existed since 1960.
I suppose though you think Bill Gates invented it all. But even then, he had already sold quite a bit of software, including MSDOS and Basic, for years by then.
This is wrong - "child" windows do not always float on the parent (sadly - the amazing vanishing dialog box remains a hassle in Gtk apps to this day)
This sounds like a GTK bug. I have never seen any X window manager developed in the last 10 years ignore the TRANSIENT_FOR hint. And on Windows it certainly obeys this rule.
Focus policy in X is the responsibility of the WM, and all the major ones allow you to switch off raise on click
The problem is that the program cannot *assumme* the raise-on-click is turned off. It is *TRIVIAL* to do raise-on-click yourself if you know the WM is not doing it. The opposity is physically impossible. For this reason the WM should NEVER do raise-on-click.
He's talking about the fact that you cannot delete a file while it is in use by another program.
Nothing to do with record locking, which I agree is a mess on both Unix and Windows, though increasingly irrelevant as most programs just treat files as atomic objects.
Although a lot better than Windows in that deletion and mv work atomically, Unix also has problems with the file locking he was talking about. Writing a file atomically is difficult, involving the need to figure out a temporary name, and to mv the result after closing, but only if the closing and writing is successful, and not mess up if the original file has unexpected permission or owner (or other metadata...). It would be nice if Linux started looking at changing things, and ignoring back-compatability with old record-oriented systems. As I see it there is only a need to write files with a "create" call, and the file is not visible to any other program until it is successfully closed. This would also get rid of all the race and security problems with temp files. I kind of doubt this will ever happen though, Linux seems stuck with complete panic about being back-compatable, while Microsoft is just sure to screw it up and make a complex interface...
No, "creation date", "modification date", and "name" are *also* metadata! Even the
"size" is metadata (actually the array of data and length that most people consider the file's "contents" is another piece of metadata).
One of the big worries about WinFS is that the idiots at Microsoft will force something on us with a whole new API because they fail to realize that all data can be named/accessed through a single API. At the API level all the metadata is the same and it should be possible to remove the creation date (you may not have permission to do this, however).
The only solution on Windows and X is to use what Windows calls "child" windows (X calls them "transient" windows). Each window has a "parent" window and the child is always atop the parent window. If the user clicks in the docuement/parent window to raise it, all the toolbars (which are children of it) are raised as well.
Now this is far from perfect, but it is the only thing you get out of X or Windows that works reliably. The most obvious problem is that there is only one parent. This makes using the same toolbars for more than one document impossible (on X you can hack in the ability to change the parent once the window exists, but you cannot do this on Windows).
The real solution is to *STOP RAISING THE WINDOWS ON CLICK!!!!!* Unfortunatly the designers of Windows and X are way too stupid and set in their ways to ever realize this. But the truth is that it is trivial for a program to raise it's own window in response to a click, and it can thus enforce any window ordering it wants.
I'm actually suprised the fake zip code works. This is done to try to prevent stolen credit cards from working. Gas pumps are doing this too, and I have started seeing this in stores. You need to enter the zip code that is on the billing address of the credit card. This is because it is a piece of information the card holder probably knows that is not printed on the credit card.
Trust me if they want to identify you they already have your entire credit card number, they don't need your zip code!
My guess: most of the time only one thing is being read from the disk at a time (yes several files can be open and reading, but I would suspect that only one being read right at a moment is very common). Therefore it seems the best use of two heads is to have them read 1/2 of the same track. This means that the duplicate positioning machinery is being used to place the heads at exactly the same track. If instead you added another platter or platter side, you could reuse the positioning machinery and get the exact same speed increase for transfers. Adding more platters also doubles the capacity, so this is a big win over multiple heads.
Of course 2 heads halves the rotational delay to aquire a piece of data. So it does have some wins. But maybe not enough.
I believe they are talking about people who *can* login as root. Non "administrators" cannot login as root. Administrators must still do something and know a password to do anything dangerous to the system.
Because installation still asks for a root password, I am confused whether this has the sudo method where the user uses their own password or the root password to get root privledges.
Like a previous poster said, it would be better if the installer just shut up rather than printing confusing information messages. There is no reason to mention "administrator" or not. Or put a "adminsitrator" checkmark on the panel with a tooltip that says "this user may sudo...".
In case anybody is confused, I believe he is talking about termanal programs. The Delete key works fine in Qt/KDE.
The basic tty driver of Linux is unchanged from ancient Unix ones and probably cannot do much. It actually cannot do any editing other than backspace, thus it is impossible to get the cursor between characters so that delete-forward can do anything. However I would like to see them scrap the entire stty stuff and just hardcode it so *both* delete and backspace and also any common escape sequences from these keys deletes backwards. An awful lot of user annoyance can be fixed by changing from "what action does this" to "what does this action do" in all the code and configuration, so that there is more than one way to do things.
Many terminal programs such as all modern shells and programs using readline and similar put the tty in raw mode and thus can do any editing that can be achieved without any vertical movement. However I checked and lots of them don't recognize the Delete key (ie they don't recognize the escape sequence they send). This should all be fixed the same way as the tty drivers. Ie scrap any "configuration" and hard-code in all the possible sequences the keys may produce and make them all work.
All these apply to Linux as well.
I would say the warnings are quite equivalent. Both the WIndows and BSD warnings are bad, and like you said they are from engineers who panic that the user may not know something.
Your Windows warnings are just as bad and I am suprised you think they are clearer to the user.
Warning -another version of Windows was detected!
How about "It looks like Windows is already installed. Should I replace it?"
Warning - are you sure you want to delete this partiton?
What's a partition? If I don't delete it is the installation going to work? Probably better is "to install Windows a part of your disk will be erased and replaced with Windows. This is what is there now (some info about whatever is on the partion), is it ok to delete and replace this? If not there are (number) other places that Windows can be put..."
Warning - your hard drive is blank (yes it really does "WARN" you of this)
A huge number of users are going to say "oh no my hard disk is faulty I need to replace it?"
I would agree that all these stupid messages (and the BSD ones) should be removed. Especially if there is no alternative.
They are talking about the same thing that OS/X has, where only "administrators" can login/sudo to do things as root. They still have to type their *own* password in order to aquire root privledges. Non-administrators cannot do anything as root.
I believe the same thing can be achieved on Linux with sudo and removing the root password and useing sudoers to select who is an "administrator". But I'm not sure about that. If so it really should be the default behavior.
Just want to repeat what another poster said. The above post is false. Assumming there really is GPL code in SCO's kernel:
1. They *may* have to pay damages, if somebody sues them and wins. This is legally possible.
2. Under *NO* possible scenario do they "have" to open-source their kernel. This is FUD straight from Bill Gates.
3. Even if they *do* open-source the kernel, they are *still* liable for damages, because they were violating the copyright before they open-sourced it. Otherwise there would be a huge loophole in the GPL (just wait until the code is uselessly old and release it and you are absolved of all copyright violations?)
It is true that companies often decide to open source some piece of code that they put GPL code into, in exchange for a promise to drop any pending lawsuits, and/or just for good public relations. However there is absolutely no legal requirement that they do this, and doing it does not put them in a legally better situation than before.
I'm not RMS's greatest fan. The GPL is really clever, but not for exactly what he thinks.
Last time I checked 3D Studio Max and Photoshop have Mac OS/X versions.
You are one of the most dangerous clueless individuals I have ever seen. Very very scary. You seem to be taking the very fact that Microsoft is a monopoly and able to dictate it's own rules and destroy interoperability, and word it as though that is an *advantage*. Wow. The mind boggles.
Um, accelerated OpenGL on multiple monitors works for us. This is however on a single card, perhaps it does not work on multiple ones...
The GPL allows me to retain control of my code. It is extremely capitalist. Programmers use it exclusively for reasons of Ayn-Randian greed. Look up "dual licensing" if you still don't have a clue. In today's world writing GPL software is the ONLY way to distribute software that you really own in a way where you may potentially make a profit or get credits or a job, unless you are the head of one of the dozen or so very large software firm. This is the real reason Microsoft attacks the GPL so much, it is really it's only competition.
"Communist" would apply a bit better to BSD code, as supposedly all your work is contributed to everybody to do what they want with it, though none of it really applies that well.
Other comparisons of Microsoft to controlled economies to reverse the argument are fun but irrelevant, as the original poster pointed out. Supposedly "ideal communisim" would not require a controlled economy.
Absolutely agree.
Too many people complain that Linux users say "go fix it yourself, you have the source". They completely miss the fact that along with that solution there is "pay somebody to fix it". This idea that code will not run on Linux unless it is free is the biggest piece of FUD out there.
Yes, one response is "go write it yourself, you have the source". Another response is "pay for it to get written!". I think the complaint is that for some reason people expect everything to be given to them for free. Nobody goes to Microsoft and demands a feature for free.
There appears to be a belief that software will not work on Linux unless it is free, as though that is some technical limitation or requirement. This is very annoying to companines and people like me who hope to sell stuff for Linux. Every time somebody complains with "there will never be kitchen-design software for Linux because nobody will write it for free" they are buying into this FUD. You are doing the same thing.
IT IS POSSIBLE TO WRITE COMMERCIAL SOFTWARE FOR LINUX!!!
Parent post was a parady of a typical attack against Michael Moore (or I guess any political documentary somebody disagrees with).
Okay. The big point is that *direct X drivers are broken too!*. You should try to point that out.
I would like confirmation that in fact it is possible to write an OpenGL LDDM driver and that it is not significantly more difficult than writing a DirectX one.
Otherwise I certainly can't blame Microsoft for making their software OpenGL simulation call their software DirectX simulation. It would in fact give some DirectX acceleration to OpenGL (though contrary to claims this already existed, the OpenGL software library calls the same driver API as DirectX does in XP).
One area they could go a long way with is to release this OpenGL library as open source, so that API's that they don't care about (ie OpenGL > 1.4) can be added.
The claim is that the hardware manufacturer no longer can make an OpenGL driver. It is impossible to pass OpenGL calls though to the driver, you must translate to Direct3D.
I don't know if this really is true however. Some Microsoft fanboy keeps claiming that the driver must be rewitten to the "LDDL" spec or "to support virtualization" but there is no backup information on this, especially an explanation as to why this is not required for existing Direct3D drivers.
Maya and every other advanced graphics appliation out there uses OpenGL and also uses the windowing system. This will seriously impact them. If this story is really true (and I have my doubts) then this is a direct attempt to force all high level appliations to switch to DirectX, thus making portable vesions very difficult or impossible.
Unlike office-type applications, these high-level programs are very popular on Linux and Mac. I think 50% of our sales are Linux. Yet we are very small, if OpenGL is not supported on Windows, we will have to drop half our market (hard to say which, though it is obvious Microsoft hopes we will drop Linux).