You were the one asking for OSX on your ol' T23 rather than on a current MBP.
That doesn't mean I consider the performance comparable! If anything, you could treat that as a measure of just how important this is to me.
I respect your opinons, I just don't see them rising to the stature of universal truths.
But the idea that people are primarily buying Apple's laptops specifically because of the "museum quality" hardware rather than the software that runs on them is unchallangable?
I don't think people are really that interested in relinking with a new version of the library. In fact I would say 100% of the interesting modifications to libraries like mine are enhancements which require the calling program to be changed to take advantage of, thus the end user gets nothing by being able to relink.
I don't agree. And I don't think Stallman does, either: there's two reasons cited for source code access in his documents of the time, and enhancements are only one of them. The other is bug fixes, and bug fixes absolutely require relinking. Fixing a bug in a library when you can't actually use the fixed library is a complete waste of time.
I've several times been saved by being able to relink software that was distributed with static libraries in a relinkable format. Not having that option makes the availability of the library source code much much less... in fact, I'm better off having a relinkable application to having the source code of the library but no way to use it in the application, because at least with the relinkable application I can install a shim to fix a bug without doing things like binary patching that massively increase the risk of unforseen interactions.
It only takes one experience of being able to fix a bug in a program by relinking it right before a major demo to make one a complete convert to relinkable applications.
And, again, the relinking requirement is not as onerous as people claim. It doesn't require anyone expose their intellectual property at all.
There's no accounting for taste. Much of what you cite is personal preference. It is not borne out by neither the majority nor the trend of users.
The effect that Apple's keyboards have on RSI is not "taste", it's "physical pain".
Yes, of course it's personal preference, but...
It's a widely held one. The style over substance design of Apple's products is a widely cited reason for people buying generic Wintel boxes and running Linux desktops on them. I'm talking about people who would rather run Mac OS X but can't stomach the hardware.
Why then would more and more people be buying them?
Because of the operating system, and because Boot Camp makes the risk of *trying* the operating system out less of a risk... though a lot of the Boot Camp sales aren't going to be repeats.
====
And yes, of course the T23 isn't up to spec with the current Mac laptops. It's over three years older! Comparing the performance on that particular model with the Macbook Pro is like me dissing the Macbook Pro because your iBook G4 was slow... but if you want to compare it to a contemporary Apple laptop... be my guest. You'll find that the T23 was lighter than and had better specs than its Power PC counterparts.
The thing you're ignoring is that a cross-platform application should run more or less the same on all target platforms, unless you just really hate users running a different OS than you.
"Run more or less the same" doesn't mean "not following the OS user interface conventions".
The details of the user interface, like how edit boxes work or what input panes or panels are used, are not going to stymie users. The things that piss users off are changes in the layout or missing elements on one OS or another. You can maintain layout and functionality without making your application stand out like a zebra in a horse race.
I wouldn't give props to the Wiimote until people actually get some experience with it over time.
Don't forget how successful the Nintendo Power Glove wasn't. Personally, I think that's a shame because I like the idea of moving towards a more VR environment where you use real gestures to do stuff, but it flopped in the market. This baby step in the same direction may fare better... or not.
The LGPL has some weird extra conditions about allowing the end user to "relink the software with a new version of the library".
Those weird extra conditions are pretty important to making the GPL and LGPL meaningful. You have to be able to *use* the source you get, and it's highly desirable that you be able to use it with the system you got it with... that's one of the motivations for the GPL in the first place: read the preamble and see.
It seems to me that statically linking AND otherwise failing to make it possible for the user to relink the software with a new version of the library subverts the intent of the GPL. And it's not clear to me that it satisfies the GPL even with the classpath exception included... if it does, it does only by accident.
And it's not necessary to do that. Statically linking _and_ providing an relinkable version of the executable (which DOESN'T need to include the unlinked libraries, any modern linker can create a relinkable but otherwise statically-linked object out of a set of objects and libraries) satisfies the terms of the LGPL, but doesn't require you to maintain binary compatibility between revisions, expose your users to DLL hell, or require that you ship any intellectual property that you're not already providing.
No, I'm not kidding. From the consumer's point of view the styling is a mixed blessing at best. They pay a premium price for it and it frequently makes the computer harder to use. I don't have an iBook G4, but I have a Macbook Pro, and I had a Thinkpad T23. let's compare them, shall we?
I don't care about the designer or the sales department. I'm a consumer.
From the consumer's POV:
The single-button trackpad and the stylish "flat" keyboard on my macbook pro are just plain inferior to the beveled and angled keys with a longer throw on my old Thinkpad. Not just a little inferior, but "this keyboard is unusable because it makes my wrists and arms hurt" inferior. There's no generic dock for Apple laptops, just third party jam-in-the-ports port replicators that are specific to each box. My Thinkpad T23 dock worked with any T series Thinkpad and most of the other lines back to before the T-20.
I'm also a recovering system administrator. From the service/support POV:
It's also MUCH harder to service and support the Macbook Pro than the Thinkpad. It's harder to get into, and it's easier to break stuff when you get in. The computer isn't physically as rugged, the lid latch is harder to use and because the case and lid are both flat there's an unavoidable gap around the edge for foreign objects to get in between the screen and the keyboard unless you're so much more careful when you pack it. But all that makes it more stylish.
And from the user's POV:
My Macbook Pro weighs more than my Thinkpad did!
I know that Apple's stuff's gone into museums. very few consumer goods are in museums for being good consumer goods. A lot of "museum quality" products and designs are impractical: many of Frank Lloyd Wright's houses were a good bad example of that. "Museum Quality" is not a reason to buy a computer... it's a warning sign, it means you're more likely to have to put up with bad design in the name of style.
As far as I can tell, the LGPL does pretty much what people want and expect. At least it seems to do pretty much the same thing as the classpath expection. But the FSF goes and tells people "we don't want you using the LGPL", so people go "OK" and use the GPL... then turn it into the LGPL.
So that seems to make them responsible for the confusion.
Apple's purpose in releasing any of their source is to make sure people can write apps and add-ons that make OSX an uber-amazing OS that runs on museum-worthy hardware and makes people want to buy more of their systems.
Museum worthy?
Well, I'd say Apple's hardware belongs in a museum. Alongside the Cray 1 and CDC 6600 and IBM 360. And the Ford Model A.
I love OS X. I hate my Macbook Pro. I would happily pay Apple their margins on the Macbook Pro if they'd let me run OS X on a Thinkpad.
Two factual errors: gcc runtime can be linked with non-GPL applications
I didn't say it couldn't. I said that this isn't like gcc.
Java SE libraries are licensed with GPL2 + Classpath Exception (which is like LGPL but less restrictive).
Thanks for the clarification about the license, though it does raise the question of why the FSF is muddying the waters by having two licences that do the same thing.
Speaking of questions:
Do you know the purpose of this symbol: "?"? It indicates that the preceding sentence is not a statement of fact, but rather a question. Surely even anonymous cowards should be able to parse that.
You're right. The terms of the classpath exception are poorly worded and confusing, and lead me to believe otherwise.
Why didn't the FSF continue to use the LGPL, which is equivalent, instead of creating new variants of the GPL. Isn't this producing the same kind of license confusion they complain about with the Creative Commons license family?
Sure, just get Apple to come up with a Windows startup sound. Apple's startup chimes are modest, moderate, and above all brief. The Windows sound... my god, I hated it from the first time I heard it. It's boastful, vainglorious, arrogant, and every time I hear it I'm reminded of everything wrong with Windows.
I tried W2K on several different machines and found it to be very slow, and that applications crashed even more often than on Windows 9 -- for no good reason.
That statement is so far at odd with my experience of Windows NT, and that of everyone else that I have ever spoken to who has been in the position of administering both Windows 9x and NT networks, that I must assume you were consciously or not sabotaging yourself. Especially when your very next sentence more or less declares it:
I was also deeply suspicious of the large number of processes that were running and the fact that I couldn't easily find out at that time what some of them were.
There are no fewer components running in a Windows 9x system performing the same tasks. The fact that in NT so many of them were separated out into separate processes merely exposed the underlying complexity that was there all along. Exposed it, and let you *if you chose to* learn how to control it. If you choose not to learn, that's your choice, but you ought to at least recognise it.
I'm sorry, but what does a handful of domain controllers to handle the AAA for the workstations have to do with what you're running your server apps on? That's like saying you can't use Windows because you need to run TACACS+ for your firewall and modem pool.
I yield to no man in my bewilderment as to why the hell the 90% of computer savvy consumers who would be better off with Windows 98,
That's a funny way to spell Windows 2000. Or were you thinking of Windows NT 3.51?
Back in the '90s I insisted that all the computers on the engineering network run UNIX or Windows NT, *not* the traditional DOS based Windows. Despite having a couple of hundred users we only needed two admins to support the netork. The admin network had less than 20 people, two admins, and one pretty much spent all the time just going around un-fucking-up Windows boxes.
And their users were a LOT less "inventive" than ours.
When we finally got them upgraded to NT4 and 2000 they were *so* happy.
NT4 was a drop in reliability compared to NT 3.51, but it was still infinitely better than Windows 9x/ME. Windows 2000 was an improvement and since it was the first with native USB support I don't think there's any point sticking with 3.51 or 4, but I haven't seen anything in XP worth upgrading for unless you absolutely need Bluetooth and don't have a Widcomm stack for your hardware.
Even if Internet Explorer is reunning in a restricted account separated from the user, a virus can (a) use the infected application as a beachhead (as you note), and (b) attack resources the user owns on the web, including web passwords.
I'de be happy letting a 7 year old use Windows 2000 without antivirus. IF it was firewalled AND Internet Explorer and any other application that uses the HTML control was disabled.
No, not just hidden inside a "restricted" account. Completely disabled.
Slashdot Mirror
You were the one asking for OSX on your ol' T23 rather than on a current MBP.
That doesn't mean I consider the performance comparable! If anything, you could treat that as a measure of just how important this is to me.
I respect your opinons, I just don't see them rising to the stature of universal truths.
But the idea that people are primarily buying Apple's laptops specifically because of the "museum quality" hardware rather than the software that runs on them is unchallangable?
64-bit kernel mode drivers have to be signed or the OS won't load them. Period.
Another reason to avoid Vista.
I don't think people are really that interested in relinking with a new version of the library. In fact I would say 100% of the interesting modifications to libraries like mine are enhancements which require the calling program to be changed to take advantage of, thus the end user gets nothing by being able to relink.
I don't agree. And I don't think Stallman does, either: there's two reasons cited for source code access in his documents of the time, and enhancements are only one of them. The other is bug fixes, and bug fixes absolutely require relinking. Fixing a bug in a library when you can't actually use the fixed library is a complete waste of time.
I've several times been saved by being able to relink software that was distributed with static libraries in a relinkable format. Not having that option makes the availability of the library source code much much less... in fact, I'm better off having a relinkable application to having the source code of the library but no way to use it in the application, because at least with the relinkable application I can install a shim to fix a bug without doing things like binary patching that massively increase the risk of unforseen interactions.
It only takes one experience of being able to fix a bug in a program by relinking it right before a major demo to make one a complete convert to relinkable applications.
And, again, the relinking requirement is not as onerous as people claim. It doesn't require anyone expose their intellectual property at all.
There's no accounting for taste.
Much of what you cite is personal preference.
It is not borne out by neither the majority nor the trend of users.
The effect that Apple's keyboards have on RSI is not "taste", it's "physical pain".
Yes, of course it's personal preference, but...
It's a widely held one. The style over substance design of Apple's products is a widely cited reason for people buying generic Wintel boxes and running Linux desktops on them. I'm talking about people who would rather run Mac OS X but can't stomach the hardware.
Why then would more and more people be buying them?
Because of the operating system, and because Boot Camp makes the risk of *trying* the operating system out less of a risk... though a lot of the Boot Camp sales aren't going to be repeats.
====
And yes, of course the T23 isn't up to spec with the current Mac laptops. It's over three years older! Comparing the performance on that particular model with the Macbook Pro is like me dissing the Macbook Pro because your iBook G4 was slow... but if you want to compare it to a contemporary Apple laptop... be my guest. You'll find that the T23 was lighter than and had better specs than its Power PC counterparts.
Wikipedia link.
And it was still a great concept.
Wii'l see how well this one turns out. Too bad Okami's for the Playstation, there's a game that screams (howls?) for a Wiimote.
The thing you're ignoring is that a cross-platform application should run more or less the same on all target platforms, unless you just really hate users running a different OS than you.
"Run more or less the same" doesn't mean "not following the OS user interface conventions".
The details of the user interface, like how edit boxes work or what input panes or panels are used, are not going to stymie users. The things that piss users off are changes in the layout or missing elements on one OS or another. You can maintain layout and functionality without making your application stand out like a zebra in a horse race.
I wouldn't give props to the Wiimote until people actually get some experience with it over time.
Don't forget how successful the Nintendo Power Glove wasn't. Personally, I think that's a shame because I like the idea of moving towards a more VR environment where you use real gestures to do stuff, but it flopped in the market. This baby step in the same direction may fare better... or not.
The LGPL has some weird extra conditions about allowing the end user to "relink the software with a new version of the library".
Those weird extra conditions are pretty important to making the GPL and LGPL meaningful. You have to be able to *use* the source you get, and it's highly desirable that you be able to use it with the system you got it with... that's one of the motivations for the GPL in the first place: read the preamble and see.
It seems to me that statically linking AND otherwise failing to make it possible for the user to relink the software with a new version of the library subverts the intent of the GPL. And it's not clear to me that it satisfies the GPL even with the classpath exception included... if it does, it does only by accident.
And it's not necessary to do that. Statically linking _and_ providing an relinkable version of the executable (which DOESN'T need to include the unlinked libraries, any modern linker can create a relinkable but otherwise statically-linked object out of a set of objects and libraries) satisfies the terms of the LGPL, but doesn't require you to maintain binary compatibility between revisions, expose your users to DLL hell, or require that you ship any intellectual property that you're not already providing.
No, I'm not kidding. From the consumer's point of view the styling is a mixed blessing at best. They pay a premium price for it and it frequently makes the computer harder to use. I don't have an iBook G4, but I have a Macbook Pro, and I had a Thinkpad T23. let's compare them, shall we?
I don't care about the designer or the sales department. I'm a consumer.
From the consumer's POV:
The single-button trackpad and the stylish "flat" keyboard on my macbook pro are just plain inferior to the beveled and angled keys with a longer throw on my old Thinkpad. Not just a little inferior, but "this keyboard is unusable because it makes my wrists and arms hurt" inferior. There's no generic dock for Apple laptops, just third party jam-in-the-ports port replicators that are specific to each box. My Thinkpad T23 dock worked with any T series Thinkpad and most of the other lines back to before the T-20.
I'm also a recovering system administrator. From the service/support POV:
It's also MUCH harder to service and support the Macbook Pro than the Thinkpad. It's harder to get into, and it's easier to break stuff when you get in. The computer isn't physically as rugged, the lid latch is harder to use and because the case and lid are both flat there's an unavoidable gap around the edge for foreign objects to get in between the screen and the keyboard unless you're so much more careful when you pack it. But all that makes it more stylish.
And from the user's POV:
My Macbook Pro weighs more than my Thinkpad did!
I know that Apple's stuff's gone into museums. very few consumer goods are in museums for being good consumer goods. A lot of "museum quality" products and designs are impractical: many of Frank Lloyd Wright's houses were a good bad example of that. "Museum Quality" is not a reason to buy a computer... it's a warning sign, it means you're more likely to have to put up with bad design in the name of style.
THAT is the point.
There are a half a dozen pieces of Apple hardware in the permanent collection of the Museum of Modern Art in NYC.
What's your point? How does that make them better personal computers?
Conservative used to be a positive term too, you know.
As far as I can tell, the LGPL does pretty much what people want and expect. At least it seems to do pretty much the same thing as the classpath expection. But the FSF goes and tells people "we don't want you using the LGPL", so people go "OK" and use the GPL... then turn it into the LGPL.
So that seems to make them responsible for the confusion.
Apple's purpose in releasing any of their source is to make sure people can write apps and add-ons that make OSX an uber-amazing OS that runs on museum-worthy hardware and makes people want to buy more of their systems.
Museum worthy?
Well, I'd say Apple's hardware belongs in a museum. Alongside the Cray 1 and CDC 6600 and IBM 360. And the Ford Model A.
I love OS X. I hate my Macbook Pro. I would happily pay Apple their margins on the Macbook Pro if they'd let me run OS X on a Thinkpad.
Two factual errors: gcc runtime can be linked with non-GPL applications
I didn't say it couldn't. I said that this isn't like gcc.
Java SE libraries are licensed with GPL2 + Classpath Exception (which is like LGPL but less restrictive).
Thanks for the clarification about the license, though it does raise the question of why the FSF is muddying the waters by having two licences that do the same thing.
Speaking of questions:
Do you know the purpose of this symbol: "?"? It indicates that the preceding sentence is not a statement of fact, but rather a question. Surely even anonymous cowards should be able to parse that.
You're right. The terms of the classpath exception are poorly worded and confusing, and lead me to believe otherwise.
Why didn't the FSF continue to use the LGPL, which is equivalent, instead of creating new variants of the GPL. Isn't this producing the same kind of license confusion they complain about with the Creative Commons license family?
The classpath exception still requires you to release your main code under the GPL.
This isn't like GCC, where the runtime is negligable and easily replaced.
Releasing Java under the GPL rather than the LGPL may hurt more than it helps.
The issue isn't the error rate, it's the accountability.
How do you verify the votes? How do you determine the error rate if you can't verify the votes?
In this case there's a paper ballot as backup.
What about the systems that don't have that?
Sure, just get Apple to come up with a Windows startup sound. Apple's startup chimes are modest, moderate, and above all brief. The Windows sound... my god, I hated it from the first time I heard it. It's boastful, vainglorious, arrogant, and every time I hear it I'm reminded of everything wrong with Windows.
I tried W2K on several different machines and found it to be very slow, and that applications crashed even more often than on Windows 9 -- for no good reason.
That statement is so far at odd with my experience of Windows NT, and that of everyone else that I have ever spoken to who has been in the position of administering both Windows 9x and NT networks, that I must assume you were consciously or not sabotaging yourself. Especially when your very next sentence more or less declares it:
I was also deeply suspicious of the large number of processes that were running and the fact that I couldn't easily find out at that time what some of them were.
There are no fewer components running in a Windows 9x system performing the same tasks. The fact that in NT so many of them were separated out into separate processes merely exposed the underlying complexity that was there all along. Exposed it, and let you *if you chose to* learn how to control it. If you choose not to learn, that's your choice, but you ought to at least recognise it.
I'm sorry, but what does a handful of domain controllers to handle the AAA for the workstations have to do with what you're running your server apps on? That's like saying you can't use Windows because you need to run TACACS+ for your firewall and modem pool.
I yield to no man in my bewilderment as to why the hell the 90% of computer savvy consumers who would be better off with Windows 98,
That's a funny way to spell Windows 2000. Or were you thinking of Windows NT 3.51?
Back in the '90s I insisted that all the computers on the engineering network run UNIX or Windows NT, *not* the traditional DOS based Windows. Despite having a couple of hundred users we only needed two admins to support the netork. The admin network had less than 20 people, two admins, and one pretty much spent all the time just going around un-fucking-up Windows boxes.
And their users were a LOT less "inventive" than ours.
When we finally got them upgraded to NT4 and 2000 they were *so* happy.
NT4 was a drop in reliability compared to NT 3.51, but it was still infinitely better than Windows 9x/ME. Windows 2000 was an improvement and since it was the first with native USB support I don't think there's any point sticking with 3.51 or 4, but I haven't seen anything in XP worth upgrading for unless you absolutely need Bluetooth and don't have a Widcomm stack for your hardware.
Even if Internet Explorer is reunning in a restricted account separated from the user, a virus can (a) use the infected application as a beachhead (as you note), and (b) attack resources the user owns on the web, including web passwords.
I'de be happy letting a 7 year old use Windows 2000 without antivirus. IF it was firewalled AND Internet Explorer and any other application that uses the HTML control was disabled.
No, not just hidden inside a "restricted" account. Completely disabled.