One thing Apple learned is that they'd conditioned their customers to expect a good quality floppy drive (with mechanical eject). The Clone machines with standard floppies didn't sell well to mac fans.
Unlike $10 PC Floppies, the replacement part value of a Mac Floppy drive was $50-$75. I don't know what the wholesale costs are, but I think it's safe to say that Apple was losing at least $20 per computer shipped relative to PC companies.
And, while I'm blabbering on the subject, has anyone here ever used a floppy drive from an older IBM brand machine? Wonderful devices -- fast, quiet, and stable. IBM even standardized on 2.88MB for a while. Of course, who wanted to pay for a nice floppy drive? People settle every day for the ones that make mysterious and random 'crunk' sounds while Windows 95 is booting. The things are screaming "unreliable" at you.
But that's the general rule of consumer computing hardware - don't expect any real quality unless it translates directly into higher Quake benchmarks. --
The jist of your post is that PC BIOSes suck in multifaceted ways, the largest two being that they're completely retarded when dealing with multiple OSes on the same computer, and there's no configuration/updating interface from any OS more modern than DOS.
Unfortunately, the fish rots from the head down, and it's probably going to have to be Intel/(Microsoft)/Compaq/IBM/Dell that will have to solve that problem, not "PC Chips".
That having been said, it would be nice if a FreeDOS boot disk creator was included with Linux distributions to make handing BIOS and setup issues easier.
(Also, IBM gets points for including a "PS2.EXE" DOS/OS2 commandline program to make BIOS settings. Perfect for batch files) --
That would be a great advertisment of how crappy Netscape on Linux is. People would just draw the conclusion that Linux is a worse operating system than Windows. --
Considering that the CuMine CPUs are apparently incompatible with existing motherboards, the fastest your system is ever going to get is duel-600Mhz. (Worth upgrade to from 2x 500Mhz? Probably not.)
If you want system longevity, do yourself a favor and don't scrape on components. Otherwise your just signing over your limited budget directly to Intel with the hope that a fast CPU will overcome all your problems. Build the duel 500 system, and when you get money, upgrade to SCSI and a fast disk, give it as much RAM as possible, and get a video card with fast 2D under Linux, like a Matrox. (Only worry about 3D if you play games.)
(Footnote1: I have a P-133 that's perceptualy faster than low end PPros and Celerons for everything except for booting and hardcore number crunching. The reason is that system has virtually no bottlenecks except for the CPU. Footnote2: I'm getting a duel PII-400 next week.) --
Actually, IBM probably should start giving away OS/2. It isn't like they make much of their profit on it.)
Maybe not from OS/2 itself, but the corporations who still run OS/2 tend to also buy millions of dollars of other IBM hardware and software. These people are also more than happy to pay exorbitant prices for OS/2. Resulting situation: OS/2 will never die, but will never be free either. --
Unisys has been trying to enforce their patents since the BBS days -- I've been reading these Unisys-is-getting-evil-with-GIFs threads since back when I had a 1200 bps modem.
A brief bit of history - Compuserve published the GIF format in the 1980s without regard to the fact that it used patented technology. The "minicomputer giant", Unisys (formerly Sperry Univac) was so out of it that they didn't realize that their algorythm was all over the online world until the late eighties. Then they started to demand that authors fork over money. The big guys (Adobe, Corel) have been paying Unisys for over a decade. --
Only when the format was a strongly entrenched de facto standard...
You mean only when it entrenched as a de facto BBS standard. Unisys has been trying to enforce it patents since the late 80s, long before the WWW days.
Most BBS standards died along with the medium -- GIF unfortunately survived. Blame Mosaic and Netcape for foisting a dubious standard on to the World Wid Web at it's inception. There was an opportunity there to introduce a new image format, and it was balked. --
Correction - Microsoft has announced that Windows 2000 "DataCenter" will support 32 CPUs, and journalists are speculating that this product will ship Q2 or Q3 2000. (Of course, they were speculating that Win2000 would ship last month.)
I would like to "me too" your comment about the datasets. The governement distribution model is stuck in the nine track tape era, when this public data could and should be delivered for the cost of bandwidth. --
Note that your argument also applies to Microsoft and others using the BSD TCP/IP stack. (Thinking back to flame wars in the past....)
In many of the license wars here on slashdot, it's often forgotten that certain software was developed either directly or indirectly with government funding, usually as part of the national defense infrastructure, but also for the good of the public as a whole (commercial and noncommercial interests included). --
One more thing, Apple maintains a central registry of type/creator codes, so there is never going to be a conflict.
This generally isn't a problem, but then recently I noticed our Lotus Domino gateway was assigning *.DOC files the MIME type "X-LotusManuscript" -- a program that has been produced in 10 years! --
Well, I always thought that MacOS's Type/Creator system was a feature. For example, if you were last working with a JPEG in Photoshop, double-clicking it should not launch JPEGView. Admittedly, not having a good integrated UI to change the types is a big problem. (I used to use a shareware extention to the Get Info box that gave nice dropdown menus.)
Anyway, it's certainly better than Windows' last-installed-program-wins battle of file type associations, and Linux's who-knows-what-might-happen approach. --
What usually matters is that your program's preferences are in some dot-rc file in your home directory or in/etc, and these things can't be moved around without telling the program.
Which is exactly my point. By the time all the "dot-rc"'s and the "/etc"'s and so-on have been removed/dumb-down, it Just Wouldn't Be Unix Any More (although it might be POSIX compliant).
While the problem might not be totally intractable, I have yet to see a Unix distribution that didn't have hard coded paths (and symlink workarounds) all over the place to get it working. Really not any better than the Windows registry, except you have better parsing tools at hand.
Meanwhile the only hardcoded paths on the Mac are in the System Folder. (And yes, a Mac app will start with default settings if you delete it's prefs file. Of course a Mac application isn't mearly an app that runs on the Macintosh -- it must follow the Mac Laws.)
As for your driver loading example, my feeling is that as easy as the Mac way is, dragging files around can be left behind. A nice GUI frontend for/etc/conf.modules is probably all that's needed (and probably already exists.)
That doesn't mean that there is some deep ``UNIX principle'' if you will which rules out the possibility of making graphical interfaces for system configuration and other things.
The key to the MacOS' usablility is not the titlebars or the button shapes. It's features built into the system that make it act more intutively to regular dumb users.
+ On MacOS you can drag programs around your hard drive, and they will still work (Except MS Office 4.2, but that's a Microsoft product!).
+ You can design your own directory structure in a way that makes sense to you. The parts of the system that you can't move or rename (System Folder) contain files that have understandable names.
+ Installing a driver on MacOS is as simple as putting a file in a folder. Uninstalling works in reverse.
+ Double-clicking and Drag-and-Drop are more likely to produce the desired effect because that's the primary interface into the system.
+ *All* configuration is done in the GUI.
The bottom line is that these kinds of features will never be in Unix DE, because they're contrary to the way Unix operates. It's likely that KDE or Gnome will equal the 'ease-of-use' of Windows, but without radical changes in the core operating system, it's unlikely that Linux will ever have an environment that equals the Mac in accessiblity for the 'dumb user' population.
(Of course Apple could be throwing out all of these features with OS X, but we'll see.....)
BIOS support for large IDE boot partitions is a pretty new thing in the PC world, and NT 4.0 is getting pretty long in the tooth. When NT4 shipped, even SCSI BIOSes generally only supported a 2GB boot partition.
(I'd be happier if they just had shipped an NT4.1 rather than pretending that Windows 2000 was just about read for the last 2 years. That would have at least resolved the stupid install issues like large IDE disks.) --
So you want an easy way to store state on the client, but don't want cookies? This is actually planned for Netscape 6.0 -- it's called Oreos. Microsoft has also been whispering about "Project TollHouse".
Seriously, you can maintain state over HTTP without cookies or URL goofyness, but it involves some pretty complicated javascript, which probably wouldn't be cross-browser compatible. --
I've been hearing whispers from Microsoft that "Next year we'll ship a Notes killer" for five years now. Right now, Microsoft can't even decided what "Tahoe" is exactly, much less when it will ship. Anyways, wake me up when it happens.
Other than that, your post is entirely full of astroturf crap.
Millions of Notes/Domino users are crying for better MS Office integration, and are going to the point of "masochistic wrangling" in the form of OLE Automation to get it.
Lotus solves the problem, but packages as another expensive product, obscurely marketed and likely to be dropped within a year or two, just like dozens of other interesting but unpopular Domino add-ons they've produced over the years.
No real point here - just the perils of dealing with Lotus as a vendor.
As a late follow-up, the ZDNet story has this mysterious addendum:
Sun reports no change in its strategy toward Linux -- Linux applications run on Solaris -- but the company does include an Open Source Software Toolkit with Solaris 7 that offers tools and applications optimized for ISPs running Solaris on Intel.
Domino doesn't automatically render MS Word documents into HTML - you would need to convert them to the Notes storage format.
Although the Office 2000 server stuff isn't very popular here, considering if that's where your data is, it might be the best bet. Within the next few years, Microsoft is planning to build an entire web groupware system more akin to Domino based on the Office server extentions. Not that you can wait for them, just something to think about.
Slashdot Mirror
One thing Apple learned is that they'd conditioned their customers to expect a good quality floppy drive (with mechanical eject). The Clone machines with standard floppies didn't sell well to mac fans.
Unlike $10 PC Floppies, the replacement part value of a Mac Floppy drive was $50-$75. I don't know what the wholesale costs are, but I think it's safe to say that Apple was losing at least $20 per computer shipped relative to PC companies.
And, while I'm blabbering on the subject, has anyone here ever used a floppy drive from an older IBM brand machine? Wonderful devices -- fast, quiet, and stable. IBM even standardized on 2.88MB for a while. Of course, who wanted to pay for a nice floppy drive? People settle every day for the ones that make mysterious and random 'crunk' sounds while Windows 95 is booting. The things are screaming "unreliable" at you.
But that's the general rule of consumer computing hardware - don't expect any real quality unless it translates directly into higher Quake benchmarks.
--
The jist of your post is that PC BIOSes suck in multifaceted ways, the largest two being that they're completely retarded when dealing with multiple OSes on the same computer, and there's no configuration/updating interface from any OS more modern than DOS.
Unfortunately, the fish rots from the head down, and it's probably going to have to be Intel/(Microsoft)/Compaq/IBM/Dell that will have to solve that problem, not "PC Chips".
That having been said, it would be nice if a FreeDOS boot disk creator was included with Linux distributions to make handing BIOS and setup issues easier.
(Also, IBM gets points for including a "PS2.EXE" DOS/OS2 commandline program to make BIOS settings. Perfect for batch files)
--
That would be a great advertisment of how crappy Netscape on Linux is. People would just draw the conclusion that Linux is a worse operating system than Windows.
--
Considering that the CuMine CPUs are apparently incompatible with existing motherboards, the fastest your system is ever going to get is duel-600Mhz. (Worth upgrade to from 2x 500Mhz? Probably not.)
If you want system longevity, do yourself a favor and don't scrape on components. Otherwise your just signing over your limited budget directly to Intel with the hope that a fast CPU will overcome all your problems. Build the duel 500 system, and when you get money, upgrade to SCSI and a fast disk, give it as much RAM as possible, and get a video card with fast 2D under Linux, like a Matrox. (Only worry about 3D if you play games.)
(Footnote1: I have a P-133 that's perceptualy faster than low end PPros and Celerons for everything except for booting and hardcore number crunching. The reason is that system has virtually no bottlenecks except for the CPU. Footnote2: I'm getting a duel PII-400 next week.)
--
Actually, IBM probably should start giving away OS/2. It isn't like they make much of their profit on it.)
Maybe not from OS/2 itself, but the corporations who still run OS/2 tend to also buy millions of dollars of other IBM hardware and software. These people are also more than happy to pay exorbitant prices for OS/2. Resulting situation: OS/2 will never die, but will never be free either.
--
Actually there might have been a SALT II, but it was never ratified by either side.
--
Unisys has been trying to enforce their patents since the BBS days -- I've been reading these Unisys-is-getting-evil-with-GIFs threads since back when I had a 1200 bps modem.
A brief bit of history - Compuserve published the GIF format in the 1980s without regard to the fact that it used patented technology. The "minicomputer giant", Unisys (formerly Sperry Univac) was so out of it that they didn't realize that their algorythm was all over the online world until the late eighties. Then they started to demand that authors fork over money. The big guys (Adobe, Corel) have been paying Unisys for over a decade.
--
Only when the format was a strongly entrenched de facto standard ...
You mean only when it entrenched as a de facto BBS standard. Unisys has been trying to enforce it patents since the late 80s, long before the WWW days.
Most BBS standards died along with the medium -- GIF unfortunately survived. Blame Mosaic and Netcape for foisting a dubious standard on to the World Wid Web at it's inception. There was an opportunity there to introduce a new image format, and it was balked.
--
Correction - Microsoft has announced that Windows 2000 "DataCenter" will support 32 CPUs, and journalists are speculating that this product will ship Q2 or Q3 2000. (Of course, they were speculating that Win2000 would ship last month.)
--
I would like to "me too" your comment about the datasets. The governement distribution model is stuck in the nine track tape era, when this public data could and should be delivered for the cost of bandwidth.
--
Note that your argument also applies to Microsoft and others using the BSD TCP/IP stack. (Thinking back to flame wars in the past....)
In many of the license wars here on slashdot, it's often forgotten that certain software was developed either directly or indirectly with government funding, usually as part of the national defense infrastructure, but also for the good of the public as a whole (commercial and noncommercial interests included).
--
One more thing, Apple maintains a central registry of type/creator codes, so there is never going to be a conflict.
This generally isn't a problem, but then recently I noticed our Lotus Domino gateway was assigning *.DOC files the MIME type "X-LotusManuscript" -- a program that has been produced in 10 years!
--
Well, I always thought that MacOS's Type/Creator system was a feature. For example, if you were last working with a JPEG in Photoshop, double-clicking it should not launch JPEGView. Admittedly, not having a good integrated UI to change the types is a big problem. (I used to use a shareware extention to the Get Info box that gave nice dropdown menus.)
Anyway, it's certainly better than Windows' last-installed-program-wins battle of file type associations, and Linux's who-knows-what-might-happen approach.
--
Hmmm, what about Solaris?
--
What usually matters is that your program's preferences are in some dot-rc file in your home directory or in /etc, and these things can't be moved around without telling the program.
/etc/conf.modules is probably all that's needed (and probably already exists.)
Which is exactly my point. By the time all the "dot-rc"'s and the "/etc"'s and so-on have been removed/dumb-down, it Just Wouldn't Be Unix Any More (although it might be POSIX compliant).
While the problem might not be totally intractable, I have yet to see a Unix distribution that didn't have hard coded paths (and symlink workarounds) all over the place to get it working. Really not any better than the Windows registry, except you have better parsing tools at hand.
Meanwhile the only hardcoded paths on the Mac are in the System Folder. (And yes, a Mac app will start with default settings if you delete it's prefs file. Of course a Mac application isn't mearly an app that runs on the Macintosh -- it must follow the Mac Laws.)
As for your driver loading example, my feeling is that as easy as the Mac way is, dragging files around can be left behind. A nice GUI frontend for
That doesn't mean that there is some deep ``UNIX principle'' if you will which rules out the possibility of making graphical interfaces for system configuration and other things.
No, just 20 years of "UNIX Practice"!
--
The key to the MacOS' usablility is not the titlebars or the button shapes. It's features built into the system that make it act more intutively to regular dumb users.
+ On MacOS you can drag programs around your hard drive, and they will still work (Except MS Office 4.2, but that's a Microsoft product!).
+ You can design your own directory structure in a way that makes sense to you. The parts of the system that you can't move or rename (System Folder) contain files that have understandable names.
+ Installing a driver on MacOS is as simple as putting a file in a folder. Uninstalling works in reverse.
+ Double-clicking and Drag-and-Drop are more likely to produce the desired effect because that's the primary interface into the system.
+ *All* configuration is done in the GUI.
The bottom line is that these kinds of features will never be in Unix DE, because they're contrary to the way Unix operates. It's likely that KDE or Gnome will equal the 'ease-of-use' of Windows, but without radical changes in the core operating system, it's unlikely that Linux will ever have an environment that equals the Mac in accessiblity for the 'dumb user' population.
(Of course Apple could be throwing out all of these features with OS X, but we'll see.....)
--
Agreed. The only thing that Exchange and it's long shutdown prove is that the NT Services Control Manager is retarded.
This is an old problem on NT, and it would be nice if it was solved.
--
The long reboot problem has existed in Exchange since version 1.0 (err I mean 4.0).
If they finally fixed it with 5.5 SP2, I'd be amazed that they actually got around to bothering.
--
BIOS support for large IDE boot partitions is a pretty new thing in the PC world, and NT 4.0 is getting pretty long in the tooth. When NT4 shipped, even SCSI BIOSes generally only supported a 2GB boot partition.
(I'd be happier if they just had shipped an NT4.1 rather than pretending that Windows 2000 was just about read for the last 2 years. That would have at least resolved the stupid install issues like large IDE disks.)
--
So you want an easy way to store state on the client, but don't want cookies? This is actually planned for Netscape 6.0 -- it's called Oreos. Microsoft has also been whispering about "Project TollHouse".
Seriously, you can maintain state over HTTP without cookies or URL goofyness, but it involves some pretty complicated javascript, which probably wouldn't be cross-browser compatible.
--
I've been hearing whispers from Microsoft that "Next year we'll ship a Notes killer" for five years now. Right now, Microsoft can't even decided what "Tahoe" is exactly, much less when it will ship. Anyways, wake me up when it happens.
Other than that, your post is entirely full of astroturf crap.
Millions of Notes/Domino users are crying for better MS Office integration, and are going to the point of "masochistic wrangling" in the form of OLE Automation to get it.
Lotus solves the problem, but packages as another expensive product, obscurely marketed and likely to be dropped within a year or two, just like dozens of other interesting but unpopular Domino add-ons they've produced over the years.
No real point here - just the perils of dealing with Lotus as a vendor.
As a late follow-up, the ZDNet story has this mysterious addendum:
Sun reports no change in its strategy toward Linux -- Linux applications run on Solaris -- but the company does include an Open Source Software Toolkit with Solaris 7 that offers tools and applications optimized for ISPs running Solaris on Intel.
The performance leader for Domino on x86 has always been Solaris/Intel. I would expect that on Linux, performance should be similar or even better.
Domino doesn't automatically render MS Word documents into HTML - you would need to convert them to the Notes storage format.
Although the Office 2000 server stuff isn't very popular here, considering if that's where your data is, it might be the best bet. Within the next few years, Microsoft is planning to build an entire web groupware system more akin to Domino based on the Office server extentions. Not that you can wait for them, just something to think about.