"I don't get all these new gadgets in cars these days. They've got this thing called 'passive restraint'. How it works is you sit down behind the wheel and the seat belts grab you and strap you in firmly. Then a deep voice comes from the radio saying, 'We have you now Mr. Bond...'"
Re:Whose task is copy&paste
on
The Power of X
·
· Score: 1
Well of course! If you are not a professional software developer this might not make sense, but it's a very common thing to do. You do the hard stuff first because it's hard and will take the longest. The easy stuff you can fill in along the way. That's probably why there are so many unfinished projects on Sourceforge, because a lot of people decide to work on the easy stuff first...
First of all, my survey was on configuration dialogs. I didn't just pick this area out of my hat, but it just happened to be a major inconstancy I tripped over the other day.
You are right however that keyboard shortcuts are more consistant under Windows than elsewhere. This is despite the fact that the Windows programmer STILL has to connect most of them up by hand. A few like Alt-F4 might be hardwired by the system to specific events, but most require cooperation by the developer. Ditto for standard accelerators.
But direct your anger at the correct party. Microsoft's standards are for Microsoft. Solaris doesn't use it, despite your assertion that it does. What Solaris does instead is to follow the Open Group's keyboard standard. For those keystrokes that the OG doesn't ignore, they pretty much follow the Microsoft way (Alt-F4 is used for "close" and not "close window"). The problem is that the Open Group standard is very hard to find and very expensive when you do. So Free Software developers simply don't bother with it. That's where the inconsistancy is coming from, some developers will use Microsoft's standards while others will studiously avoid them.
but because it follows the definitive sci-fi paradigm of changing the setting in a pseudo-science-ish speculative way
And just why doesn't Star Wars fit this bill? We have alien cultures, FTL travel, robotics, and even midichlorians. It sounds pseudo-science-ish to me!
The point is, if Star Wars isn't SciFi despite its pseudo-science trappings, then neither is the X-Men. The only thing speculative about X-Men is how society would react to the sudden emergence of magic.
which is why SF actually stands for *speculative* fiction
That's a backronym. It was originally "science fiction" and the term "speculative fiction" came about later.
That's the problem with modern scifi, it's unfathomable unless you first void your mind of all content.
Seriously, calling a X-Men scifi is as silly as calling Superman scifi. Oh wait, you probably do. Maybe if I take LOTR and throw in some technobabble I can get you to think it's scifi too...
X-Men!!?! How can you possibly call X-Men "science fiction"? Star Wars has a thousand times the imaginative speculative science that X-Men does, yet you call it fantasy.
Mutations have a basis in scientific fact, but mutations that cause "magic" do not. Magneto's magnetism is magical. Cyclop's eyesight is magical. To suggest that these and other powers can arise through the mutation of DNA is ludicrous.
I've never checked what final size I get with an initial install, as I usually go install lot of stuff on top of it immediately afterwards. But I can give you some hints. I don't know if these will work for you are not, but give it a shot.
Don't install the source code if you don't need it, or remove it afterwards if you do. Don't include Linux compatibility. Don't install games, profiled libraries, pre-catted man pages. The 3.x and 4.x compat libs are pretty small, but leave them out anyway if you don't need them.
Don't install the X.org/XFree86 metapackage but use the individual component packages instead, so you won't be sucking down a lot of stuff you won't need, like docs and cyrillic fonts.
Re:The Misinformation Campaign Rolls Along...
on
The Power of X
·
· Score: 2, Informative
Cairo. DBUS. HAL. the.desktop standards.
Excepting the.desktop format, none of them are fd.o standards, and even the.desktop format is still a "draft". Havoc's unilateral pronouncements that something is a "standard" means nothing more than he wishes it was. Most of the real de facto standards (like.desktop) were created by the "little guys" of KDE and GNOME working together without the benefit of Havoc's blessings.
In all Open Source projects, the people who get things done and the people who strut about crowing are two separate groups with very little intersection. Freedesktop.org is no different.
Re:unified desktop
on
The Power of X
·
· Score: 3, Insightful
The Inconstancy of the Windows Desktop
One of the more insistant and vocal themes heard in the desktop debate is that that Unix desktop needs to be like Windows. It is said that multiple widget toolkits, inconsistant dialogs, and other evidences of a decentralized development model must be removed before the masses will accept a Unix destkop. This cry for uniformity can be especially shrill, almost as if the very survival of a certain free operating system depended upon it. But is the underlying premise true? Is Windows really a consistant and uniform desktop?
The answer is resoundingly negative.
While conducting a quick survey of configuration dialogs under Windows, in an attempt to understand what a newbie user of my software would be familiar with, I discovered that there was no standard procedure for these dialogs. Even configuration dialogs from the same manufacturer varied wildly. By all Slashdot accounts, Windows users must certainly be mentally damaged from their constant exposure to such inconsistant interfaces.
Where is the configuration dialog located for a Windows application? Using the Windows system I use every day at work, I discovered that even this simple item was highly variable. Microsoft Word had two configuration dialogs, "Tools->Customize" and "Tools->Options", while Microsoft Outlook added an additional "Tools->Services". Microsoft WordPad had only one under a completely different menu "View->Options". Moving on to non-Microsoft products, I see that Adobe Reader and Quicktime Player have "Edit->Preferences". But lest you think those are consistant, Adobe Reader has a single dialog, while Quicktime Player has a submenu of three dialogs. Firefox and Roxio Creator Classic follow the WordPad model of placement.
What about the dialog contents themselves? Microsoft Word has modal tabbed dialogs, while Microsoft Outlook has a modeless tabbed dialog without a help button. Adobe Reader and Firefox have modal dialogs using a listbox instead of tabs to separate the pages. Quicktime Player is similar, but uses a combobox instead of a listbox. Some of these dialogs had help buttons while the rest lacked them.
Okay, what about the look and feel? Certainly the Windows platform has a consistant widget set? Sadly, no. Adobe Reader has an almost-but-not-quite Win2K look, that matches neither the Windows Classic nor Luna themes that comes with Windows XP. Roxio Creator Classic has a "brushed plastic" look with odd splitter controls. Quicktime player has, of course, a look and feel straight out of another operating system! Comparing native Microsoft applications only improves matters slightly. Microsoft Word has a completely different toolbar style than Microsoft WordPad! I could continue on to some truly egregious examples of inconsistancy, but I'll leave that as an exercise to the reader.
I think by now that I have thoroughly debunked the notion that the Windows desktop is uniform and consistant. The question remains though, is the Unix desktop better? The answer is similarly, "no". But since Windows isn't consistant, the urgency of the question is clearly lessoned. Newbies aren't going to be rendered insane by seeing Evolution running alongside Konqueror. They aren't going to go running back to Windows when their distro forgot to include Plastik icons with Mozilla.
Corporations are moving towards all web-apps. In my company I have seen one rolled out to replace a desktop app approximately once a month for the past year.
Unfortunately these webapps all require Windows Internet Explorer running on a Windows NT/XP desktop. It's not about thin clients, it's about being able to manage a multinational corporation from a single IT building in New Jersey.
Re:KDE and Knome infect X ?
on
The Power of X
·
· Score: 1
Don't listen to the propaganda coming from Freedesktop.org then. he politics promulgated by the fd.o leadership are regrettable, but fortunately no one much listens to Daniel and Havoc except Daniel and Havoc. The software that fd.o hosts is completely independent of fd.o. In this regard it's not much different from freshmeat. The rest of fd.o still serves as a useful clearinghouse for desktop interoperability standards. While Havoc has a lot of clout within Redhat and GNOME, he has no authority within any other destkop's developer community.
Re:I'm using the new X.org
on
The Power of X
·
· Score: 1
More or less nothing has changed in XFree86 since the license change.
Except that X.org starts up about ten times slower than XFree86 in exactly the same environment. Everything else seems just the same, just the startup is dog slow. I have no idea why.
Re:Upgrading from 5.2.1 to 5.3-BETA1 a little bump
on
FreeBSD 5.3 Beta1
·
· Score: 2, Informative
A few weeks ago I upgraded my 5.2.1 laptop to 5-CURRENT, and the build stopped with a message about mergemaster right at the very beginning. No need to wait around an hour to discover your mistake...
Re:Whose task is copy&paste
on
The Power of X
·
· Score: 1
Even your 1984 Mac paint programs had to actually code this functionality into the software. That didn't happen instantly and by magic. KolourPaint is brand new, as in "it's still in the middle of being potty trained" new. Give it a tiny bit of time and that crucial bit of functionality will be added.
In the meantime I expect that KolourPaint can do thinks your 1984 Mac programs could only dream about. Like color, transparency, etc...
Try using something like WindowMaker and some non-GTK non-QT apps.
I'll have to disagree with you here, because you're comparing apples to oranges. Those KDE and GNOME apps seem slightly slower and less responsive because they've been jammed packed full of functionality. The KDE desktop does about a hundred times as much as the bare WindowMaker window manager. A KDE application is going to pull in ten times the functionality from the KDE librarieas than a bare xlib application will get from X.
The very same thing holds true under Windows, but few people are honest enough to admit it. Windows 95 was snappy and responsive under a 90MHz Pentium and 8MB RAM. But try putting Windows XP on the same system! The difference is that Windows 95 didn't have a tenth the functionality of XP, and that's not even counting the eye candy. When I compare the workstation in front of me between Windows XP Pro and FreeBSD with KDE, I am finding that while XP apps may sometimes start up faster, they most certainly are NOT more responsive. Your situation may be different, but in my work environment that is what I am seeing.
The Misinformation Campaign Rolls Along...
on
The Power of X
·
· Score: 1
In other words, this is merely a hosting site. Except for its smaller size and narrower focus, it's not any different from freshmeat or savannah. The software at fd.o is not official and not required for desktops to use. They are not standards which must be adopted for "freedesktop compliance". While there have been a few people clamouring to make X.org the "official" X11 of GNOME and KDE, frankly it's not going to happen.
To take a specific assertion, "While freedesktop.org is, in many ways, a fairly loosely organized community project, we're all really minions of Havoc": this is completely wrong. Havoc Pennington is an employee of Redhat and a GNOME developer. He has no authority over X.org, KDE, XFCE or any other project outside of Redhat and GNOME. He has extremely little input to any of the projects hosted at fd.o.
Freedesktop.org started out as a great idea, and for a while it was very useful. But then it got infected with politics, and quickly lost most of their relevancy. If not for their recent hosting of independent software projects, they would have faded into obsurity. Frankly, would anyone care about fd.o if X.org wasn't hosted there?
This doesn't explain Syllable, NewOS, OpenBeOS, or ReactOS, which are GPL and mentioned in the granparent post. Neither does it address the issue of why the Free Software community behaves as if there was one and only one Free Software operating system.
OK, I busted my ass on this great piece of software, you can steal it, change the name and sell it.
Actually that's not true, because YOU didn't write the software. According to your anecdote, it was someone named "Max Baker".
So let me rephrase your objection: "OK, you busted your ass on this gret piece of software, and now I am bitching about because you don't care if other people change its name and sell it." Sounds like a classic case of busybodyism if you ask me.
Well, if the UN isn't sponsoring it, I guess I'll go ahead and celebrate Free Software. The UN isn't about freedom, so if it ever officially supported Free Software, they would only be doing so with a strangely bizarre definition of freedom.
I've seen the UN talk about Free Software, and it seems to me that their definition of "freedom" means "collectivist enforcement of sharing". They want involuntary copyleft reciprocation instead of voluntary giving. To them it's not about removing artificial government controls (eg, copyright and patents) on information, as is shown by their near complete silence on abolishing state granted intellectual property privileges.
See what types of calls to Win98 the driver is trying to do and then "translate" those to the equivalent calls in XP and vice versa on the return side--no knowledge of the device needed.
XP's problem was that users were migrating to a completely different operating system. DOS and NT/XP are as different as CPM and VMS. The only reason so many W98 drivers work in WXP is because Microsoft threw several abstraction layers between the kernel and the driver. Frankly, I think it's amazing that so many still work.
But despite any abstraction layers that may exist, the kernel and drivers can still have bugs in them that won't show up until the next release. I develop an embedded system and routinely run across bugs that have lain hidden for years. The easiest way to uncover them is to move to a more recent version of the RTOS.
In the case of the FreeBSD NVidia driver, I was shit out of luck for six months because NVidia was not updating their driver. It was a major pain in the ass because the old binary was crashing the kernel about once a day.
NVidia argued that this was FreeBSD's fault for providing a moving 5.x target, and said they weren't going to update the driver until 5.x was marked stable. But the funny thing is that NONE of the open source XFree86/X.org DRI drivers had this problem.
I'm not taking this anti-binary stance because I'm some sort of "freedom" ideologue, I'm taking this stance because I ran headlong into the "pragmatic" brick wall of closed source drivers.
...why would it not work with later versions of the OS?
That is where your argument completely falls apart. That binary is *NOT* going to work with different kernels. My friend has a Canon printer that will not work under WinXP desptite the fact that he has binary drivers for Win98. A coworker has a card that will not work with Linux 2.6 even though he has a binary driver for 2.4. I myself have experienced extreme difficulties using an binary video driver with FreeBSD for the same reasons.
The source code will at least let you keep up with minor kernel interface variances, and easily port across major interface variances.
Another drawback to binary drivers is that you have no fix if they start interfering with hardware you by later on. While there won't be tens of thousands of developers eager to fix the source based driver for you, there will still be scads more than that bankrupt hardware company will provide.
Slashdot Mirror
"I don't get all these new gadgets in cars these days. They've got this thing called 'passive restraint'. How it works is you sit down behind the wheel and the seat belts grab you and strap you in firmly. Then a deep voice comes from the radio saying, 'We have you now Mr. Bond...'"
Well of course! If you are not a professional software developer this might not make sense, but it's a very common thing to do. You do the hard stuff first because it's hard and will take the longest. The easy stuff you can fill in along the way. That's probably why there are so many unfinished projects on Sourceforge, because a lot of people decide to work on the easy stuff first...
I get it! If the story has depth and meaning then it's science fiction, but if it's just cheap entertainment then it's fantasy.
First of all, my survey was on configuration dialogs. I didn't just pick this area out of my hat, but it just happened to be a major inconstancy I tripped over the other day.
You are right however that keyboard shortcuts are more consistant under Windows than elsewhere. This is despite the fact that the Windows programmer STILL has to connect most of them up by hand. A few like Alt-F4 might be hardwired by the system to specific events, but most require cooperation by the developer. Ditto for standard accelerators.
But direct your anger at the correct party. Microsoft's standards are for Microsoft. Solaris doesn't use it, despite your assertion that it does. What Solaris does instead is to follow the Open Group's keyboard standard. For those keystrokes that the OG doesn't ignore, they pretty much follow the Microsoft way (Alt-F4 is used for "close" and not "close window"). The problem is that the Open Group standard is very hard to find and very expensive when you do. So Free Software developers simply don't bother with it. That's where the inconsistancy is coming from, some developers will use Microsoft's standards while others will studiously avoid them.
but because it follows the definitive sci-fi paradigm of changing the setting in a pseudo-science-ish speculative way
And just why doesn't Star Wars fit this bill? We have alien cultures, FTL travel, robotics, and even midichlorians. It sounds pseudo-science-ish to me!
The point is, if Star Wars isn't SciFi despite its pseudo-science trappings, then neither is the X-Men. The only thing speculative about X-Men is how society would react to the sudden emergence of magic.
which is why SF actually stands for *speculative* fiction
That's a backronym. It was originally "science fiction" and the term "speculative fiction" came about later.
Actually IBM is doing this because they got sued by SCO...
...what makes this event so cool is that it is organized entirely by volunteers...
I'm so glad they've stopped the practice of involuntary servitude...
That's the problem with modern scifi, it's unfathomable unless you first void your mind of all content.
Seriously, calling a X-Men scifi is as silly as calling Superman scifi. Oh wait, you probably do. Maybe if I take LOTR and throw in some technobabble I can get you to think it's scifi too...
X-Men
X-Men!!?! How can you possibly call X-Men "science fiction"? Star Wars has a thousand times the imaginative speculative science that X-Men does, yet you call it fantasy.
Mutations have a basis in scientific fact, but mutations that cause "magic" do not. Magneto's magnetism is magical. Cyclop's eyesight is magical. To suggest that these and other powers can arise through the mutation of DNA is ludicrous.
I've never checked what final size I get with an initial install, as I usually go install lot of stuff on top of it immediately afterwards. But I can give you some hints. I don't know if these will work for you are not, but give it a shot.
Don't install the source code if you don't need it, or remove it afterwards if you do. Don't include Linux compatibility. Don't install games, profiled libraries, pre-catted man pages. The 3.x and 4.x compat libs are pretty small, but leave them out anyway if you don't need them.
Don't install the X.org/XFree86 metapackage but use the individual component packages instead, so you won't be sucking down a lot of stuff you won't need, like docs and cyrillic fonts.
Cairo. DBUS. HAL. the .desktop standards.
.desktop format, none of them are fd.o standards, and even the .desktop format is still a "draft". Havoc's unilateral pronouncements that something is a "standard" means nothing more than he wishes it was. Most of the real de facto standards (like .desktop) were created by the "little guys" of KDE and GNOME working together without the benefit of Havoc's blessings.
Excepting the
In all Open Source projects, the people who get things done and the people who strut about crowing are two separate groups with very little intersection. Freedesktop.org is no different.
The Inconstancy of the Windows Desktop
One of the more insistant and vocal themes heard in the desktop debate is that that Unix desktop needs to be like Windows. It is said that multiple widget toolkits, inconsistant dialogs, and other evidences of a decentralized development model must be removed before the masses will accept a Unix destkop. This cry for uniformity can be especially shrill, almost as if the very survival of a certain free operating system depended upon it. But is the underlying premise true? Is Windows really a consistant and uniform desktop?
The answer is resoundingly negative.
While conducting a quick survey of configuration dialogs under Windows, in an attempt to understand what a newbie user of my software would be familiar with, I discovered that there was no standard procedure for these dialogs. Even configuration dialogs from the same manufacturer varied wildly. By all Slashdot accounts, Windows users must certainly be mentally damaged from their constant exposure to such inconsistant interfaces.
Where is the configuration dialog located for a Windows application? Using the Windows system I use every day at work, I discovered that even this simple item was highly variable. Microsoft Word had two configuration dialogs, "Tools->Customize" and "Tools->Options", while Microsoft Outlook added an additional "Tools->Services". Microsoft WordPad had only one under a completely different menu "View->Options". Moving on to non-Microsoft products, I see that Adobe Reader and Quicktime Player have "Edit->Preferences". But lest you think those are consistant, Adobe Reader has a single dialog, while Quicktime Player has a submenu of three dialogs. Firefox and Roxio Creator Classic follow the WordPad model of placement.
What about the dialog contents themselves? Microsoft Word has modal tabbed dialogs, while Microsoft Outlook has a modeless tabbed dialog without a help button. Adobe Reader and Firefox have modal dialogs using a listbox instead of tabs to separate the pages. Quicktime Player is similar, but uses a combobox instead of a listbox. Some of these dialogs had help buttons while the rest lacked them.
Okay, what about the look and feel? Certainly the Windows platform has a consistant widget set? Sadly, no. Adobe Reader has an almost-but-not-quite Win2K look, that matches neither the Windows Classic nor Luna themes that comes with Windows XP. Roxio Creator Classic has a "brushed plastic" look with odd splitter controls. Quicktime player has, of course, a look and feel straight out of another operating system! Comparing native Microsoft applications only improves matters slightly. Microsoft Word has a completely different toolbar style than Microsoft WordPad! I could continue on to some truly egregious examples of inconsistancy, but I'll leave that as an exercise to the reader.
I think by now that I have thoroughly debunked the notion that the Windows desktop is uniform and consistant. The question remains though, is the Unix desktop better? The answer is similarly, "no". But since Windows isn't consistant, the urgency of the question is clearly lessoned. Newbies aren't going to be rendered insane by seeing Evolution running alongside Konqueror. They aren't going to go running back to Windows when their distro forgot to include Plastik icons with Mozilla.
Corporations are moving towards all web-apps. In my company I have seen one rolled out to replace a desktop app approximately once a month for the past year.
Unfortunately these webapps all require Windows Internet Explorer running on a Windows NT/XP desktop. It's not about thin clients, it's about being able to manage a multinational corporation from a single IT building in New Jersey.
Don't listen to the propaganda coming from Freedesktop.org then. he politics promulgated by the fd.o leadership are regrettable, but fortunately no one much listens to Daniel and Havoc except Daniel and Havoc. The software that fd.o hosts is completely independent of fd.o. In this regard it's not much different from freshmeat. The rest of fd.o still serves as a useful clearinghouse for desktop interoperability standards. While Havoc has a lot of clout within Redhat and GNOME, he has no authority within any other destkop's developer community.
More or less nothing has changed in XFree86 since the license change.
Except that X.org starts up about ten times slower than XFree86 in exactly the same environment. Everything else seems just the same, just the startup is dog slow. I have no idea why.
A few weeks ago I upgraded my 5.2.1 laptop to 5-CURRENT, and the build stopped with a message about mergemaster right at the very beginning. No need to wait around an hour to discover your mistake...
Even your 1984 Mac paint programs had to actually code this functionality into the software. That didn't happen instantly and by magic. KolourPaint is brand new, as in "it's still in the middle of being potty trained" new. Give it a tiny bit of time and that crucial bit of functionality will be added.
In the meantime I expect that KolourPaint can do thinks your 1984 Mac programs could only dream about. Like color, transparency, etc...
Try using something like WindowMaker and some non-GTK non-QT apps.
I'll have to disagree with you here, because you're comparing apples to oranges. Those KDE and GNOME apps seem slightly slower and less responsive because they've been jammed packed full of functionality. The KDE desktop does about a hundred times as much as the bare WindowMaker window manager. A KDE application is going to pull in ten times the functionality from the KDE librarieas than a bare xlib application will get from X.
The very same thing holds true under Windows, but few people are honest enough to admit it. Windows 95 was snappy and responsive under a 90MHz Pentium and 8MB RAM. But try putting Windows XP on the same system! The difference is that Windows 95 didn't have a tenth the functionality of XP, and that's not even counting the eye candy. When I compare the workstation in front of me between Windows XP Pro and FreeBSD with KDE, I am finding that while XP apps may sometimes start up faster, they most certainly are NOT more responsive. Your situation may be different, but in my work environment that is what I am seeing.
I see the misinformation campaign continues. Let me quote from the site itself: "None of this is "endorsed" by anyone or implied to be standard software, remember that freedesktop.org is a collaboration forum, so anyone is encouraged to host stuff here if it's on-topic".
In other words, this is merely a hosting site. Except for its smaller size and narrower focus, it's not any different from freshmeat or savannah. The software at fd.o is not official and not required for desktops to use. They are not standards which must be adopted for "freedesktop compliance". While there have been a few people clamouring to make X.org the "official" X11 of GNOME and KDE, frankly it's not going to happen.
To take a specific assertion, "While freedesktop.org is, in many ways, a fairly loosely organized community project, we're all really minions of Havoc": this is completely wrong. Havoc Pennington is an employee of Redhat and a GNOME developer. He has no authority over X.org, KDE, XFCE or any other project outside of Redhat and GNOME. He has extremely little input to any of the projects hosted at fd.o.
Freedesktop.org started out as a great idea, and for a while it was very useful. But then it got infected with politics, and quickly lost most of their relevancy. If not for their recent hosting of independent software projects, they would have faded into obsurity. Frankly, would anyone care about fd.o if X.org wasn't hosted there?
Otherwise they would be more popular.
This doesn't explain Syllable, NewOS, OpenBeOS, or ReactOS, which are GPL and mentioned in the granparent post. Neither does it address the issue of why the Free Software community behaves as if there was one and only one Free Software operating system.
OK, I busted my ass on this great piece of software, you can steal it, change the name and sell it.
Actually that's not true, because YOU didn't write the software. According to your anecdote, it was someone named "Max Baker".
So let me rephrase your objection: "OK, you busted your ass on this gret piece of software, and now I am bitching about because you don't care if other people change its name and sell it." Sounds like a classic case of busybodyism if you ask me.
Well, if the UN isn't sponsoring it, I guess I'll go ahead and celebrate Free Software. The UN isn't about freedom, so if it ever officially supported Free Software, they would only be doing so with a strangely bizarre definition of freedom.
I've seen the UN talk about Free Software, and it seems to me that their definition of "freedom" means "collectivist enforcement of sharing". They want involuntary copyleft reciprocation instead of voluntary giving. To them it's not about removing artificial government controls (eg, copyright and patents) on information, as is shown by their near complete silence on abolishing state granted intellectual property privileges.
See what types of calls to Win98 the driver is trying to do and then "translate" those to the equivalent calls in XP and vice versa on the return side--no knowledge of the device needed.
XP's problem was that users were migrating to a completely different operating system. DOS and NT/XP are as different as CPM and VMS. The only reason so many W98 drivers work in WXP is because Microsoft threw several abstraction layers between the kernel and the driver. Frankly, I think it's amazing that so many still work.
But despite any abstraction layers that may exist, the kernel and drivers can still have bugs in them that won't show up until the next release. I develop an embedded system and routinely run across bugs that have lain hidden for years. The easiest way to uncover them is to move to a more recent version of the RTOS.
In the case of the FreeBSD NVidia driver, I was shit out of luck for six months because NVidia was not updating their driver. It was a major pain in the ass because the old binary was crashing the kernel about once a day.
NVidia argued that this was FreeBSD's fault for providing a moving 5.x target, and said they weren't going to update the driver until 5.x was marked stable. But the funny thing is that NONE of the open source XFree86/X.org DRI drivers had this problem.
I'm not taking this anti-binary stance because I'm some sort of "freedom" ideologue, I'm taking this stance because I ran headlong into the "pragmatic" brick wall of closed source drivers.
...why would it not work with later versions of the OS?
That is where your argument completely falls apart. That binary is *NOT* going to work with different kernels. My friend has a Canon printer that will not work under WinXP desptite the fact that he has binary drivers for Win98. A coworker has a card that will not work with Linux 2.6 even though he has a binary driver for 2.4. I myself have experienced extreme difficulties using an binary video driver with FreeBSD for the same reasons.
The source code will at least let you keep up with minor kernel interface variances, and easily port across major interface variances.
Another drawback to binary drivers is that you have no fix if they start interfering with hardware you by later on. While there won't be tens of thousands of developers eager to fix the source based driver for you, there will still be scads more than that bankrupt hardware company will provide.