There may be a way to "hijack" the selection buffer (what the middle mouse pastes) as well as the clipboard. Perhaps display the evil text in a text widget, make sure it has focus, and set the selection on it. Maybe you can even fake mouse events to the widget if that is the only way to change the selection.
It has nothing to do with text verses other types of data.
The problem is that the original "selection" was replaced when the user selected text. IE selecting text was pretty much identical to selecting text and doing an immediate Copy on Windows.
There were a lot of confused attempts to modify the selection to be more like the clipboard, until finally around 1997 everybody realized that they were two different things. In fact the selection is really the drag & drop buffer. You would not expect drag & drop to modify the clipboard, therefore these should be two buffers. I guess Windows has the same thing though they call the selection the drag & drop.
I keep hearing that this is addressed but I never see it work, even on the newest Gnome systems and GTK or Qt programs.
Windows actually has the same bug, but it sends a WM_RENDERALLFORMATS message when the program is exiting and this is used as a signal that the program should write it's selected text to the global store.
The "selection" is what came first, actually. The "selection" is much better thought of as a drag & drop. Clicking the middle mouse button is the same as dragging and dropping the current selection, with the huge advantage that you can rearrange windows and launch programs between the start and end of the drag & drop. So in effect X had drag & drop initially, but no clipboard.
Unfortunatly for a long time X users had no idea what they had and kept thinking it was a clipboard. The ideas did not merge very well, primarily because selecting text in effect immediatly did a copy. The largest problem with this is that it is impossible to select text and replace it with the previous copy. It also screwed up users who did not expect selection to mess with the previous copy.
Finally around 1997 they added the second clipboard and the toolkits were fixed to use it correctly. Now what is needed is for them to realize that drag & drop and the selection are the same thing, and ideally all applications and toolkits fixed so the results are identical whether you do a drag & drop or you click the middle mouse button. This would be a good deal better than drag & drop on other systems and a good use of the middle mouse button.
Have the same problem on my Ubuntu laptop. You can fix it with "sudo hdparm -B 192/dev/sda" but this has several problems: first it completely disables the head parking. Second is that, despite trying to follow the instructions, I cannot get it to automatically run this when woken from sleep, I have to manually type it (I made a button do it). This is not good, Ubuntu.
More interesting is that somebody *finally* did some research to find out what Windows was doing. Apparently it is not doing a secret handshake to tell the disk to act right, nor changing it from it's default setting. What actually happens is that Windows (both XP and Vista) continuously access the disk all the time as long as it has not entered sleep mode. What the disk manufacturers have done is adjust the head-park time as tiny as possible so that it is the most efficient "detect if Windows is sleeping" algorithim. I'm not sure if there is any fix other than to duplicate the Windows bug in Ubuntu.
Your original post did not mention capital gains tax at all, it talked about "income" and taxes. Capital gains tax is less than 50% even under Obama's plan.
Right from the second page, if you bother hitting "next page". Not exactly a ringing endorsement, but certainly not the dismissal you claim, and he expects to use federal money to encourage nuclear power:
Asked his views on nuclear power in Jacksonville, Florida on Friday, Obama said, "I think that nuclear power should be in the mix when it comes to energy." But he added, "I don't think it's our optimal energy source because we haven't figured out how to store the waste safely or recycle the waste."
Obama supports using federal research and development dollars to explore whether nuclear waste can be stored safely for reuse.
Why the hell do your "companies" go out of business exactly one year after they become profitable?
I have another plan for you: make 1 company. After it becomes profitable (2-3 years), KEEP it and keep making 500K a year and paying 50%. You would then make about 490K per year and pay 50%.
You don't seem to understand that Linux stores the configuration in the users home directory, not in a central location. Thus two users do not clobber each other's settings.
Most of your other points are valid, but some of them are not because of this.
I think everybody is over-estimating what OSX is doing.
It does not magically connect your application to the services.
It is their text-editor widget that has code added to it to talk to the services. It is not amazing in any way. You cannot spell-check any data other than is selected in the text-editor box, unless the program is actually rewritten to be able to talk to the services. Same as what would be written on Linux.
That does not excuse the fact that little/nothing has been done on Linux. The service api should exist. And it should be SIMPLE (ie it should involve calling popen() and no other calls). And then the toolkits (KDE and Qt and maybe others) can modify their text-editor widgets to call it, and that will "magically make all programs use the services". Of course real integration will require programs to be rewritten, exactly like they are on OS/X.
I think the worry is that Microsoft will release under a license that is purposely incompatible with the GPL, which means it will also be incompatible with the BSD license.
In case you don't believe me, it should be pretty obvious. BSD code can be used in GPL projects. Therefore any rule that prevents the code from being used in a GPL project will also prevent it from being released as BSD licensed, because part of being BSD licensed is the ability to use it in GPL code.
There is a much more obvious reason why avoidance of Vista actually means bad things for Linux as well. The reason people don't go to Vista is unfamiliarity and fear. These exact same things work against Linux. If even Microsoft cannot overcome that, despite the fact that they can literally force you to get a "free" copy of Vista if you need to get a new computer, then it seems pretty hopeless for Linux to compete with XP (and 98 and NT, which still exist in droves).
I think the problem is that it would be *FAR* more effective to say nothing. If the post really is childish, that will be quite clear to everybody reading it. Responding to it indicates that it upset you.
The M$ stuff does amuse me, though. In this case I think the problem is that it is *not* childish or whatever. Nobody responds if somebody says "microsucks" because the post stands on it's own for belittling the poster. But since M$ really is not that childish, the Microsoft defenders feel the *have* to post to try to convince people it is.
Too many people here say that Microsoft needs to release Windows or something under the GPL to get their trust. That won't happen, and is not necessary.
Conversely, what Microsoft wants is to add Windows-specific extensions to existing GPL projects. That is probably the extent of their evil plans.
What I want to see Microsoft do:
Rather than just say "here is some extra code and ifdefs so it runs on Windows" is make changes to Windows so "your Linux code compiles without changes". Now let's ignore X, which is a mess, and I can't blame them for not emulating that. But they need to provide a working, default, POSIX-like environment. This means that all files can be clearly named with as string starting with "/", and that if you truncate after a slash and do readdir, you get a list of the files there. It means that any api that returns a filename returns forward slashes. It means that there is no difference between text and binary files (interesting that they still have relics of 1950's mechanical teletype designs in their system, yet claim they are the "modern" one).
Somebody above mentioned how "nice" they were about trying to get people to develop for Windows CE. Yet the whole point about developing for CE is that the code does not port to other machines. What I would like to see is them saying "hey you can run the same programs on CE as on those Linux phones, without any changes, because we will help you".
OSS has done enough crap trying to be compatible with them it is time for them to show that they really mean it by doing some work on their own.
Other things: drop OOXML and actually list those "200 patents".
Let's fuck up all their de-facto standards with GPL-only extensions and see how they like it.
First they could use the GPL code to implement the standards if they wanted to, and release the relevant tools GPL. I'm sure they are clever enough to wall it off into an application so that they don't have to release stuff they don't want to.
Second it is highly unlikely for any standard to be accepted even on Linux unless the reference is at least LGPL, or GPL with the linking exception, or BSD. Even the FSF realizes this and releases anything they want to establish as a standard as BSD.
Most importantly the GPL covers an implementation and not the standard. They can just write their own code to implement it. Even if they can't figure it out from the docs, the GPL explicitly allows reverse engineering. This is on purpose and part of the GPL design.
Um, the "viral" thing is used by people arguing *against* the GPL, not for it!
"Viral" implies that it will infect you by you simply being in the same room. In reality you have to actually *use* the GPL code, and redistribute the result, for it to have any effect on you, which requires a conscious effort. For this reason GPL proponents certainly don't use "viral".
Therefore his quote is actually just the same as GPL opponents use, substituting "cancer" for "viral".
It certainly is funny. Especially the hilarious responses from the M$ fanboys, such as the above, their feelings so hurt that they can't type straight and they can't realize that it would be far more effective to shut up.
Also funny is how they go crazy if you say "M$" in a post. "Oh that's so childish! I'm going to tell my mom! You are a little little child! Childish! Childish! Childish! Because I say so you are a little child!" As though saying it enough times will make their dreams come true.
If you check my posting history you will see I used to write "Microsoft" but I always try to remember to type M$ nowadays, just because of the hilarious immature responses from them...
We don't need more files to read. They need to look at how Unix was originally designed.
There should be a little program called "start" or "open" that takes a URL as an argument and does whatever the hell the user expects a double-click on that to do it. I should be able to use it with 1 line of C (an execl() call). No library and no file I have to parse.
They should also look at using Unix ideas for solving GUI issues. The file chooser should be a program you exec. You can either wait for it to finish (it prints the chosen file on stdout), or use pipes for bidirectional communication. And all the toolkits should call that program and scrap their built-in filechooser (although those will probably be used to implement the first ones). If this is done I think within months Linux will switch from having the worst file chooser to by far the best one, because there will be lots of competing ones, and it will be trivial for users to try different ones.
All the other popus (alerts, little dialog boxes, etc) should be done with exec as well.
That's not "binary compatability", that is generally called "DLL hell" (I suppose there is a Unix name for it, but it is the same thing). I agree that missing libraries and/or programs requiring older versions of libraries is a big problem on Linux so that it is impossible to install most precompiled software. For the older versions I always end up symlinking the newer one to the old name, which seems to work mostly, but in that case I certainly am introducing binary incompatability.
Slashdot Mirror
There may be a way to "hijack" the selection buffer (what the middle mouse pastes) as well as the clipboard. Perhaps display the evil text in a text widget, make sure it has focus, and set the selection on it. Maybe you can even fake mouse events to the widget if that is the only way to change the selection.
It has nothing to do with text verses other types of data.
The problem is that the original "selection" was replaced when the user selected text. IE selecting text was pretty much identical to selecting text and doing an immediate Copy on Windows.
There were a lot of confused attempts to modify the selection to be more like the clipboard, until finally around 1997 everybody realized that they were two different things. In fact the selection is really the drag & drop buffer. You would not expect drag & drop to modify the clipboard, therefore these should be two buffers. I guess Windows has the same thing though they call the selection the drag & drop.
I keep hearing that this is addressed but I never see it work, even on the newest Gnome systems and GTK or Qt programs.
Windows actually has the same bug, but it sends a WM_RENDERALLFORMATS message when the program is exiting and this is used as a signal that the program should write it's selected text to the global store.
That's not the same. What is wanted is to copy the most recently selected object, whether or not Copy was done after the selection.
The "selection" is what came first, actually. The "selection" is much better thought of as a drag & drop. Clicking the middle mouse button is the same as dragging and dropping the current selection, with the huge advantage that you can rearrange windows and launch programs between the start and end of the drag & drop. So in effect X had drag & drop initially, but no clipboard.
Unfortunatly for a long time X users had no idea what they had and kept thinking it was a clipboard. The ideas did not merge very well, primarily because selecting text in effect immediatly did a copy. The largest problem with this is that it is impossible to select text and replace it with the previous copy. It also screwed up users who did not expect selection to mess with the previous copy.
Finally around 1997 they added the second clipboard and the toolkits were fixed to use it correctly. Now what is needed is for them to realize that drag & drop and the selection are the same thing, and ideally all applications and toolkits fixed so the results are identical whether you do a drag & drop or you click the middle mouse button. This would be a good deal better than drag & drop on other systems and a good use of the middle mouse button.
It really should just have text on it, preferably the word "meta".
(yea I know that newer Linux uses the X "super" keysym for it, but it really should be "meta" anyway, that is a much better word).
Have the same problem on my Ubuntu laptop. You can fix it with "sudo hdparm -B 192 /dev/sda" but this has several problems: first it completely disables the head parking. Second is that, despite trying to follow the instructions, I cannot get it to automatically run this when woken from sleep, I have to manually type it (I made a button do it). This is not good, Ubuntu.
More interesting is that somebody *finally* did some research to find out what Windows was doing. Apparently it is not doing a secret handshake to tell the disk to act right, nor changing it from it's default setting. What actually happens is that Windows (both XP and Vista) continuously access the disk all the time as long as it has not entered sleep mode. What the disk manufacturers have done is adjust the head-park time as tiny as possible so that it is the most efficient "detect if Windows is sleeping" algorithim. I'm not sure if there is any fix other than to duplicate the Windows bug in Ubuntu.
Your original post did not mention capital gains tax at all, it talked about "income" and taxes. Capital gains tax is less than 50% even under Obama's plan.
Right from the second page, if you bother hitting "next page". Not exactly a ringing endorsement, but certainly not the dismissal you claim, and he expects to use federal money to encourage nuclear power:
Asked his views on nuclear power in Jacksonville, Florida on Friday, Obama said, "I think that nuclear power should be in the mix when it comes to energy." But he added, "I don't think it's our optimal energy source because we haven't figured out how to store the waste safely or recycle the waste."
Obama supports using federal research and development dollars to explore whether nuclear waste can be stored safely for reuse.
Makes sense, but isn't this capital gains tax, not income tax?
Is the 500K the sale price, or the sale price minus the investment made in the original company?
Obama never said anything against nuclear power.
Why the hell do your "companies" go out of business exactly one year after they become profitable?
I have another plan for you: make 1 company. After it becomes profitable (2-3 years), KEEP it and keep making 500K a year and paying 50%. You would then make about 490K per year and pay 50%.
You don't seem to understand that Linux stores the configuration in the users home directory, not in a central location. Thus two users do not clobber each other's settings.
Most of your other points are valid, but some of them are not because of this.
I think everybody is over-estimating what OSX is doing.
It does not magically connect your application to the services.
It is their text-editor widget that has code added to it to talk to the services. It is not amazing in any way. You cannot spell-check any data other than is selected in the text-editor box, unless the program is actually rewritten to be able to talk to the services. Same as what would be written on Linux.
That does not excuse the fact that little/nothing has been done on Linux. The service api should exist. And it should be SIMPLE (ie it should involve calling popen() and no other calls). And then the toolkits (KDE and Qt and maybe others) can modify their text-editor widgets to call it, and that will "magically make all programs use the services". Of course real integration will require programs to be rewritten, exactly like they are on OS/X.
I think the worry is that Microsoft will release under a license that is purposely incompatible with the GPL, which means it will also be incompatible with the BSD license.
In case you don't believe me, it should be pretty obvious. BSD code can be used in GPL projects. Therefore any rule that prevents the code from being used in a GPL project will also prevent it from being released as BSD licensed, because part of being BSD licensed is the ability to use it in GPL code.
There is a much more obvious reason why avoidance of Vista actually means bad things for Linux as well. The reason people don't go to Vista is unfamiliarity and fear. These exact same things work against Linux. If even Microsoft cannot overcome that, despite the fact that they can literally force you to get a "free" copy of Vista if you need to get a new computer, then it seems pretty hopeless for Linux to compete with XP (and 98 and NT, which still exist in droves).
I think the problem is that it would be *FAR* more effective to say nothing. If the post really is childish, that will be quite clear to everybody reading it. Responding to it indicates that it upset you.
The M$ stuff does amuse me, though. In this case I think the problem is that it is *not* childish or whatever. Nobody responds if somebody says "microsucks" because the post stands on it's own for belittling the poster. But since M$ really is not that childish, the Microsoft defenders feel the *have* to post to try to convince people it is.
Too many people here say that Microsoft needs to release Windows or something under the GPL to get their trust. That won't happen, and is not necessary.
Conversely, what Microsoft wants is to add Windows-specific extensions to existing GPL projects. That is probably the extent of their evil plans.
What I want to see Microsoft do:
Rather than just say "here is some extra code and ifdefs so it runs on Windows" is make changes to Windows so "your Linux code compiles without changes". Now let's ignore X, which is a mess, and I can't blame them for not emulating that. But they need to provide a working, default, POSIX-like environment. This means that all files can be clearly named with as string starting with "/", and that if you truncate after a slash and do readdir, you get a list of the files there. It means that any api that returns a filename returns forward slashes. It means that there is no difference between text and binary files (interesting that they still have relics of 1950's mechanical teletype designs in their system, yet claim they are the "modern" one).
Somebody above mentioned how "nice" they were about trying to get people to develop for Windows CE. Yet the whole point about developing for CE is that the code does not port to other machines. What I would like to see is them saying "hey you can run the same programs on CE as on those Linux phones, without any changes, because we will help you".
OSS has done enough crap trying to be compatible with them it is time for them to show that they really mean it by doing some work on their own.
Other things: drop OOXML and actually list those "200 patents".
I fail to see "Winblows" or "microsucks" or any of your other words in the above quote.
Meanwhile it is trivial to find the word "cancer" in a Microsoft quote.
BZZT. Please try again.
Let's fuck up all their de-facto standards with GPL-only extensions and see how they like it.
First they could use the GPL code to implement the standards if they wanted to, and release the relevant tools GPL. I'm sure they are clever enough to wall it off into an application so that they don't have to release stuff they don't want to.
Second it is highly unlikely for any standard to be accepted even on Linux unless the reference is at least LGPL, or GPL with the linking exception, or BSD. Even the FSF realizes this and releases anything they want to establish as a standard as BSD.
Most importantly the GPL covers an implementation and not the standard. They can just write their own code to implement it. Even if they can't figure it out from the docs, the GPL explicitly allows reverse engineering. This is on purpose and part of the GPL design.
Um, the "viral" thing is used by people arguing *against* the GPL, not for it!
"Viral" implies that it will infect you by you simply being in the same room. In reality you have to actually *use* the GPL code, and redistribute the result, for it to have any effect on you, which requires a conscious effort. For this reason GPL proponents certainly don't use "viral".
Therefore his quote is actually just the same as GPL opponents use, substituting "cancer" for "viral".
It certainly is funny. Especially the hilarious responses from the M$ fanboys, such as the above, their feelings so hurt that they can't type straight and they can't realize that it would be far more effective to shut up.
Also funny is how they go crazy if you say "M$" in a post. "Oh that's so childish! I'm going to tell my mom! You are a little little child! Childish! Childish! Childish! Because I say so you are a little child!" As though saying it enough times will make their dreams come true.
If you check my posting history you will see I used to write "Microsoft" but I always try to remember to type M$ nowadays, just because of the hilarious immature responses from them...
Huh? I certainly didn't miss that, that was my entire point!
We don't need more files to read. They need to look at how Unix was originally designed.
There should be a little program called "start" or "open" that takes a URL as an argument and does whatever the hell the user expects a double-click on that to do it. I should be able to use it with 1 line of C (an execl() call). No library and no file I have to parse.
They should also look at using Unix ideas for solving GUI issues. The file chooser should be a program you exec. You can either wait for it to finish (it prints the chosen file on stdout), or use pipes for bidirectional communication. And all the toolkits should call that program and scrap their built-in filechooser (although those will probably be used to implement the first ones). If this is done I think within months Linux will switch from having the worst file chooser to by far the best one, because there will be lots of competing ones, and it will be trivial for users to try different ones.
All the other popus (alerts, little dialog boxes, etc) should be done with exec as well.
That's not "binary compatability", that is generally called "DLL hell" (I suppose there is a Unix name for it, but it is the same thing). I agree that missing libraries and/or programs requiring older versions of libraries is a big problem on Linux so that it is impossible to install most precompiled software. For the older versions I always end up symlinking the newer one to the old name, which seems to work mostly, but in that case I certainly am introducing binary incompatability.