Re:Debian should come with warning stickers...
on
Debian Retail on CNN
·
· Score: 2
I kind of like the idea of seeing a nice shrinkwrapped box in CompUSA with a big warning label on the front:
Warning - Debian Linux is for advanced Linux users only. If "MS Windows"-like simplicity is what you want, we suggest you try one of the many other fine commercial Linux distributions. However, if you are looking for the best application packaging system and the finest level of control and quality assurance in a Linux distribution, Debian Linux is for you!
Debian is a not-for-profit organization dedicated to free software
If anything, such a warning would probably increase sales.
The Apple ][, besides being first-to-market with a complete solution, was also much cheaper than your average CP/M machine in those days.
(*80/S100/CPM machines were primarily kit machine or put together by local retailers. I think the Apple ][ debuted at $2500, while CP/M machines were double that price once you got all the parts.
The CP/M machine were also very fractured from a hardware standpoint, with a number of competing video and disk drive standards. This problem wasn't really solved until IBM introduced it's own CP/M 'clone', the PC. Before then, the #1 selling CP/M computer was actually the Apple ][ with an optional Z80 card made by Microsoft of all people.)
For those who might not have experienced AU/X, it was a full blown System V UNIX that ran some Mac II and Quadra hardware. '
The cool thing about it was that MacOS 7 booted on top of AU/X as a UNIX process, giving you full access to Mac programs along side your Unix command shell. Note that there was no GUI for the Unix programs - you couldn't run X Windows with the MacOS running. This meant either Unix command programs or un-modern MacOS programs, but no nice Mac GUI for Unix programs.
The uncool thing about AU/X was that GNU banned it, so no gcc meaning lots of software was not accessible.
Anyways, as you said, Apple *could* have ported A/UX to PowerMacs, but that would mean big licence fees to UNIX, inc. Much better to salvage the cool part (MacOS-on-UNIX , aka "Blue Box" or "Classic"), and port that to an open platform like Mach/BSD.
I'll give Compaq and IBM credit for high-end PC servers. However, they both stopped making their own motherboards for mainstream PCs and even 'workstations' many years ago.
(I have one of the last desktops that Compaq designed in house, and it's a Pentium 133.)
The first USB ports I ever saw were on a Pentium Pro system from about 1995-6 (I remember Win95 had just been released). They were covered with a little sticker that said "Operating System Support Pending".
Intel shipped lots of USB chipsets that no one could really use until 1998. Amazing in the as-cheap-as-possible world of PC hardware.
Could Dell be just a facade for Andy Groves' operation?
All PC manufactures are facades for Intel. They screw together the parts, take a small profit margin, and hope that an ongoing support relationship with the customer will lead to some real money.
Even, back in the day, Novell Netware 3.x had patch lists several pages long.
How much you want to bet there's quite a few of these PC Week editors and "IT Managers" with an old Novell CNE tie-tack somewhere in their desk drawer. They know the routine - they've just forgotten.
IBM has an army of a sales force, and they have there fingers in many diverse industries
Except that IBM has much less influnce on the computing market than in 1981. Even in many shops with IBM datacenters, IBM doesn't have a lot of pull with the people who run the 'PC network', especially after the PS/2-OS/2 fiasco ten years ago.
And where IBM does have influence, it's dollars-to-donuts that the sales guys are getting a far larger commission for Lotus Domino than they ever would for Linux implementations.
However, for the true-blue IBM guys in the datacenter, there's a market for Linux. Generally, these guys have no concept of TCP/IP networking, and would be more than happy to have their local IBM field engineer drop a Netfinity running Linux into place to handle the DNS, DHCP, e-mail, etc.
most ISVs still consider Win* or an older Unix their primary platform, but they are becoming aware the Linux MAY replace said platforms
Yes, but IBM ain't "most ISVs". Most of their revenue comes from the S/390 and AS/400 systems running accounting systems and the like. (They've claimed that the AS/400 division is larger and more profitable than all of Sun Microsystems, and the theory is the S/390 is even more profitable than AS/400.) Linux is absolutely no threat to this space, technically or in terms of potential revenue.
If anything, Linux does nothing more than allow IBM to build an end-to-end solution, giving them cheap HTTP and DNS servers to front their heafty e-business infrastructure.
At the rate things are going, graphics cards will soon be the most expensive component in every system
Maybe the most expensive component in every *gaming* system. Most business PCs (which are most PCs) have pretty crappy graphics - stuff you could buy retail for $10-$20.
The video card is practically the last point of differentation between systems - most of which ship with similar CPUs, the same Intel-based motherboard, similar EIDE disks, and similar sound hardware.
Will Intel finaly take the gloves of with M$ and show them that the world does not revolve around Redmond?
Both Microsoft and Intel hate the concept of "wintel", even though Microsoft is 99% dependant on Intel systems and Intel is 90% dependant on Microsoft systems.
Another thing they both have in common is the lust for the high profit datacenter market owned by IBM, Sun, and HP. The difference is, since Microsoft dropped Alpha, they are totally dependant on Intel to get them there. (Not that NT-on-Alpha was selling well, which may also will be the case for NT-on-I64.) Intel, on the other hand, looks to be betting heavily on Unix to sell the super-high-margin systems.
Lotus Notes 4.x certainly deserves several assholes ripped for it's UI, but their "hall of shame" article doesn't pull it off very well. (Notes 5 is much better, but still has a share of problems.)
Much of their criticism is not aimed at any built-in 'features' of Notes, but instead at a poorly designed custom mail application that is only used at one corporation. All of their complaints about misplaced and confusing buttons don't apply to Notes itself. It's like saying that VisualBasic or KDevelop has a crappy interface because people build programs with crappy interfaces with using the tool. (Their excuse is something along the lines of "Our criticism is valid because our company installed it that way!")
In general, while I like the idea of the site, their site, the snotty attitude of the authors are pretty off-putting. What exactly is the point of ripping on a bunch of small time VB apps or some bizarre IBM CD playing program that no one uses? Also, they hold special hatred for "mouseover" buttons as are found in Netscape and IE -- I've seen quite a few stupid users, but never one that was so stupid that they couldn't figure out where their back button is!
Re:you must see this KDE app
on
KDE Looks Ahead
·
· Score: 1
A MS Outlook clone for Linux - very interesting.
What's disappointing is that there's no back-end yet. However -
later versions will be split into "open client" and "open server" and will include data sharing, such as ability to share mail folders, addressbooks etc. The long term purpose is to transform Magellan into a document-handling client and server.
Sounds alot like "Groupware"... Hmmm. (Although, I'm not sure if something ambitious is needed, or will every really stack-up in the big enterprises that use Notes and Outlook applicaitons. What is needed desperately is a good open standard calendar server.)
Maybe there should be a Moore's variation which states that compute-speed doubles every X
One variant I've heard is that the Performance/Price ratio doubles every 18 months. A few years back, I saw some data that indicated that Intel's marketing department took this seriously anyway - either trying to double speeds or cutting prices when speeds weren't increasing that fast. A corrolary is that "Moore's Marketing Law" could be met with SMP systems, if Intel can't scale the CPU speed fast enough over an extended period of time. (See the duel Celery people.)
(I know there are some problems with this, namely that performance/price doesn't scale linearly at any given point in time. Xeons and 600Mhz CPUs have a pretty huge premium )
"Mozilla" has always been the internal name for Netscape Navigator since version 1. mozilla.org is about opening up the internal process to external developers. There probably never will be a released product called "Mozilla", only a source tarball. In short Mozilla == Netscape, and if in doubt, check your HTTP headers or about:Mozilla
Yup. I think you and I are barking up the same tree.
Desktop OSes are a zero-sum game - and that fact creates alot of harsh will and frustration among Linux advocates facing a situation that's de facto "Windows everywhere". Note that I don't include myself in the "Linux Advocate" column, just making an observation about the Linux advocacy culture.
The original Oracle NC was a *486 running *BSD with a fluffy X-based GUI on top. What's going to make the 'new' Oracle NC any different? A *686 and *Linux?
IBM came the closest to the NC concept, with a box that actually ran a Java-based OS and had some Lotus Java productivity applications. I think the goal was to provide a replacement to their (apparently still profitable) mainframe terminal business.
As a point of trivia, the old Oracle Network Computer division has been spun off as "Liberate" (big building on US101), and no longer makes NCs, but instead set top boxes and webtv things. The article was unclear if Liberate whould make the new Oracle NCs.
My point was that it's not about individual choice. It's about the choice of corporations/organizations. That's why this battle often gets phrased in all-or-nothing terms, because right now it is All Windows-and-Nothing Else.
Most users have no input on what operating system they run, and therefore have no choice. (And if they have a home computer, they probably just want it to be compatible with their work computer.)
Very early on, Microsoft figured out the power of personal computing revolution. They understood the restrictive nature of the host-terminal model and relative poor price-performace and arrogance inherit in the 'mini-computer' space. (Where the vendors have always been more interested in high profit accounting systems than in the user's ability to work with the numbers the accounting system spits out.)
This realization has lead to an enormous amount of desktop functionality crammed into Windows and Office. This has lead to enormous profits for MS and other desktop vendors. However, it's also lead to an enormous expensive mess of unmanagible systems and inaccessible data.
Enter the web revolution. To a large degree, the ABM folks back in 1995 were right - the web does make Windows obsolete because it allows corporations to push systems off of desktops and back into the datacenter where they belong, and at a price point which makes PCs look expensive.
And Microsoft is aware of this trend, but can't really address it straight on, because it's a direct attack on their profit base. So, they come up with basterized strategies such as "Windows DNA", where corporations are supposed to implement a network centered architecture that's all tied together by MS Office ActiveX components installed on each machine and a "web store" running on MS Exchange Server. It's a bastardized strategy, but it's their only hope to embrace (and extend) network-centric computing while enforcing the predominance of the desktop.
What worries me is that many Linux users aren't really aware of the "management" problem inherit in PCs. While Unix is certainly more network oriented than Windows, it seems that KDE/Gnome/etc are trying to address the "eye-candy" problem much more than they are trying to address the flaws in the Personal Computing model itself. This could be because Linux developers are often students or outside of mainstream MIS organizations.
The bottom line is that there's alot of tail chasing going on. By the time a "free" Office Suite gets to an acceptable function point and can import MS Office documents, it will already be irrelvant because those Office documents won't be "documents" any more, and instead will be locked up in Microsoft DNA-based network systems and will only be accessible by special network protocols and MS presentation software.
I don't think anybody has a real good solution to this problem. Bits and pieces are out there (CORBA, XML, etc.), but someone has to tie them together in a way that cheap enough and accessible enough for the desktop user. Unfortuantly, that someone is probably going to be Microsoft, and folks aren't going to be happy with the results.
Well, for the vast majority of users, the operating system is whatever their employer says it is. This makes desktop operating systems somewhat of a zero-sum game. Either they are "the standard" or they are only in place for nitch uses. (And if you are a 'nitch' Linux user in an all-Windows desktop shop, more power to you - just realize that you are 1/1000th of the user base.)
If "world domination" is ever reached, it means that thousands of Windows computers are going to have to be reformatted and have Linux installed. This assumes that Linux is an adequate replacement for Windows on the desktop. Meanwhile, Microsoft has been working for years to promote Windows as a general solution So, you can see where the contention arises ("Windows Everywhere" versus "World Domination")
And, yes, I realize that World Domination was originally somewhat of a joke, but thousands of/. posts indicate that people are taking it very seriously.
(A) I don't play games, so I can't and didn't say what's better. (B) Microsoft can't force game companies to do anything. Do you really think John Carmack cares what MS's policy is? Windows games are attractive because the burden of writing device drivers is off the game developers' backs. (C) You can't spell idiot. Nice try.
I would actually argue that retrofitting a game machine into a computer did work for the Amiga, as it did for Atari with both the Atari 800 and ST lines.
These systems eventually died, but that was due a bad business model more than anything. (Sell at a thin margin, invest no money in R+D.) Sony can afford to subsidize these things with both game/software sales, and the 'pro' video-editing applications.
Sure, few "academics and scientists" took the Amiga seriously. It wasn't marketed at them - it was marketed to gamers and video people, who did take it very seriously. On the other hand, Windows PCs were designed as machines to run Lotus 1-2-3 and MS Word, yet they are constantly being kludged with things like DirectX and Windows 98 to use the great gaming hardware available.
NTSC to HDTV converters are going to be in such demand that hopefully Radio Shack will be selling them for $20 for a long time to come. (After all, they still sell 8 Track stuff!)
Well at least I hope so. There's an enormous investment out there in NTSC beyond Ataris and Playstations, even if it's not being broadcast.
At some point you have to tell people to fix their broken code because the bugs, compatibility hacks and obsolete features their software depends on have been deprecated and will disappear in the next major release.
Except that time and time-again the PC userbase has chosen backwards-compatibility over a better product. See OS/2. See Windows NT. Both products had pretty good backwards-compatibilty, but not good enough for people to flock to them in great numbers on the desktop. Microsoft has to sell upgrades, and therefore backwards compatiblity is job #1. (Prediction: Watch Win2000 beta get watered down over the next few months to try to get ye olde software working.)
The root problem is the broken applications. There are just too damn many 'business critical' applications in use for which there is no longer vendor support or perhaps no longer a vendor. (And no source code, either, so forget about fixing the problem.)
Slashdot Mirror
I kind of like the idea of seeing a nice shrinkwrapped box in CompUSA with a big warning label on the front:
Warning - Debian Linux is for advanced Linux users only. If "MS Windows"-like simplicity is what you want, we suggest you try one of the many other fine commercial Linux distributions. However, if you are looking for the best application packaging system and the finest level of control and quality assurance in a Linux distribution, Debian Linux is for you!
Debian is a not-for-profit organization dedicated to free software
If anything, such a warning would probably increase sales.
The Apple ][, besides being first-to-market with a complete solution, was also much cheaper than your average CP/M machine in those days.
(*80/S100/CPM machines were primarily kit machine or put together by local retailers. I think the Apple ][ debuted at $2500, while CP/M machines were double that price once you got all the parts.
The CP/M machine were also very fractured from a hardware standpoint, with a number of competing video and disk drive standards. This problem wasn't really solved until IBM introduced it's own CP/M 'clone', the PC. Before then, the #1 selling CP/M computer was actually the Apple ][ with an optional Z80 card made by Microsoft of all people.)
(As a former AU/X user, I just have to chip in.)
For those who might not have experienced AU/X, it was a full blown System V UNIX that ran some Mac II and Quadra hardware. '
The cool thing about it was that MacOS 7 booted on top of AU/X as a UNIX process, giving you full access to Mac programs along side your Unix command shell. Note that there was no GUI for the Unix programs - you couldn't run X Windows with the MacOS running. This meant either Unix command programs or un-modern MacOS programs, but no nice Mac GUI for Unix programs.
The uncool thing about AU/X was that GNU banned it, so no gcc meaning lots of software was not accessible.
Anyways, as you said, Apple *could* have ported A/UX to PowerMacs, but that would mean big licence fees to UNIX, inc. Much better to salvage the cool part (MacOS-on-UNIX , aka "Blue Box" or "Classic"), and port that to an open platform like Mach/BSD.
I'll give Compaq and IBM credit for high-end PC servers. However, they both stopped making their own motherboards for mainstream PCs and even 'workstations' many years ago.
(I have one of the last desktops that Compaq designed in house, and it's a Pentium 133.)
The first USB ports I ever saw were on a Pentium Pro system from about 1995-6 (I remember Win95 had just been released). They were covered with a little sticker that said "Operating System Support Pending".
Intel shipped lots of USB chipsets that no one could really use until 1998. Amazing in the as-cheap-as-possible world of PC hardware.
Could Dell be just a facade for Andy Groves' operation?
All PC manufactures are facades for Intel. They screw together the parts, take a small profit margin, and hope that an ongoing support relationship with the customer will lead to some real money.
(Even the AMD systems often have Intel chipsets.)
Even, back in the day, Novell Netware 3.x had patch lists several pages long.
How much you want to bet there's quite a few of these PC Week editors and "IT Managers" with an old Novell CNE tie-tack somewhere in their desk drawer. They know the routine - they've just forgotten.
IBM has an army of a sales force, and they have there fingers in many diverse industries
Except that IBM has much less influnce on the computing market than in 1981. Even in many shops with IBM datacenters, IBM doesn't have a lot of pull with the people who run the 'PC network', especially after the PS/2-OS/2 fiasco ten years ago.
And where IBM does have influence, it's dollars-to-donuts that the sales guys are getting a far larger commission for Lotus Domino than they ever would for Linux implementations.
However, for the true-blue IBM guys in the datacenter, there's a market for Linux. Generally, these guys have no concept of TCP/IP networking, and would be more than happy to have their local IBM field engineer drop a Netfinity running Linux into place to handle the DNS, DHCP, e-mail, etc.
most ISVs still consider Win* or an older Unix their primary platform, but they are becoming aware the Linux MAY replace said platforms
Yes, but IBM ain't "most ISVs". Most of their revenue comes from the S/390 and AS/400 systems running accounting systems and the like. (They've claimed that the AS/400 division is larger and more profitable than all of Sun Microsystems, and the theory is the S/390 is even more profitable than AS/400.) Linux is absolutely no threat to this space, technically or in terms of potential revenue.
If anything, Linux does nothing more than allow IBM to build an end-to-end solution, giving them cheap HTTP and DNS servers to front their heafty e-business infrastructure.
HURD is not Unix. Linux is not Unix. What was your point?
At the rate things are going, graphics cards will soon be the most expensive component in every system
Maybe the most expensive component in every *gaming* system. Most business PCs (which are most PCs) have pretty crappy graphics - stuff you could buy retail for $10-$20.
The video card is practically the last point of differentation between systems - most of which ship with similar CPUs, the same Intel-based motherboard, similar EIDE disks, and similar sound hardware.
Will Intel finaly take the gloves of with M$ and show them that the world does not revolve around Redmond?
Both Microsoft and Intel hate the concept of "wintel", even though Microsoft is 99% dependant on Intel systems and Intel is 90% dependant on Microsoft systems.
Another thing they both have in common is the lust for the high profit datacenter market owned by IBM, Sun, and HP. The difference is, since Microsoft dropped Alpha, they are totally dependant on Intel to get them there. (Not that NT-on-Alpha was selling well, which may also will be the case for NT-on-I64.) Intel, on the other hand, looks to be betting heavily on Unix to sell the super-high-margin systems.
Lotus Notes 4.x certainly deserves several assholes ripped for it's UI, but their "hall of shame" article doesn't pull it off very well. (Notes 5 is much better, but still has a share of problems.)
Much of their criticism is not aimed at any built-in 'features' of Notes, but instead at a poorly designed custom mail application that is only used at one corporation. All of their complaints about misplaced and confusing buttons don't apply to Notes itself. It's like saying that VisualBasic or KDevelop has a crappy interface because people build programs with crappy interfaces with using the tool. (Their excuse is something along the lines of "Our criticism is valid because our company installed it that way!")
In general, while I like the idea of the site, their site, the snotty attitude of the authors are pretty off-putting. What exactly is the point of ripping on a bunch of small time VB apps or some bizarre IBM CD playing program that no one uses? Also, they hold special hatred for "mouseover" buttons as are found in Netscape and IE -- I've seen quite a few stupid users, but never one that was so stupid that they couldn't figure out where their back button is!
A MS Outlook clone for Linux - very interesting.
What's disappointing is that there's no back-end yet. However -
later versions will be split into "open client" and "open server" and will include data sharing, such as ability to share mail folders, addressbooks etc. The long term purpose is to transform Magellan into a document-handling client and server.
Sounds alot like "Groupware"
Maybe there should be a Moore's variation which states that compute-speed doubles every X
One variant I've heard is that the Performance/Price ratio doubles every 18 months. A few years back, I saw some data that indicated that Intel's marketing department took this seriously anyway - either trying to double speeds or cutting prices when speeds weren't increasing that fast. A corrolary is that "Moore's Marketing Law" could be met with SMP systems, if Intel can't scale the CPU speed fast enough over an extended period of time. (See the duel Celery people.)
(I know there are some problems with this, namely that performance/price doesn't scale linearly at any given point in time. Xeons and 600Mhz CPUs have a pretty huge premium )
"Mozilla" has always been the internal name for Netscape Navigator since version 1. mozilla.org is about opening up the internal process to external developers. There probably never will be a released product called "Mozilla", only a source tarball. In short Mozilla == Netscape, and if in doubt, check your HTTP headers or about:Mozilla
Yup. I think you and I are barking up the same tree.
Desktop OSes are a zero-sum game - and that fact creates alot of harsh will and frustration among Linux advocates facing a situation that's de facto "Windows everywhere". Note that I don't include myself in the "Linux Advocate" column, just making an observation about the Linux advocacy culture.
The original Oracle NC was a *486 running *BSD with a fluffy X-based GUI on top. What's going to make the 'new' Oracle NC any different? A *686 and *Linux?
IBM came the closest to the NC concept, with a box that actually ran a Java-based OS and had some Lotus Java productivity applications. I think the goal was to provide a replacement to their (apparently still profitable) mainframe terminal business.
As a point of trivia, the old Oracle Network Computer division has been spun off as "Liberate" (big building on US101), and no longer makes NCs, but instead set top boxes and webtv things. The article was unclear if Liberate whould make the new Oracle NCs.
My point was that it's not about individual choice. It's about the choice of corporations/organizations. That's why this battle often gets phrased in all-or-nothing terms, because right now it is All Windows-and-Nothing Else.
Most users have no input on what operating system they run, and therefore have no choice. (And if they have a home computer, they probably just want it to be compatible with their work computer.)
Very early on, Microsoft figured out the power of personal computing revolution. They understood the restrictive nature of the host-terminal model and relative poor price-performace and arrogance inherit in the 'mini-computer' space. (Where the vendors have always been more interested in high profit accounting systems than in the user's ability to work with the numbers the accounting system spits out.)
This realization has lead to an enormous amount of desktop functionality crammed into Windows and Office. This has lead to enormous profits for MS and other desktop vendors. However, it's also lead to an enormous expensive mess of unmanagible systems and inaccessible data.
Enter the web revolution. To a large degree, the ABM folks back in 1995 were right - the web does make Windows obsolete because it allows corporations to push systems off of desktops and back into the datacenter where they belong, and at a price point which makes PCs look expensive.
And Microsoft is aware of this trend, but can't really address it straight on, because it's a direct attack on their profit base. So, they come up with basterized strategies such as "Windows DNA", where corporations are supposed to implement a network centered architecture that's all tied together by MS Office ActiveX components installed on each machine and a "web store" running on MS Exchange Server. It's a bastardized strategy, but it's their only hope to embrace (and extend) network-centric computing while enforcing the predominance of the desktop.
What worries me is that many Linux users aren't really aware of the "management" problem inherit in PCs. While Unix is certainly more network oriented than Windows, it seems that KDE/Gnome/etc are trying to address the "eye-candy" problem much more than they are trying to address the flaws in the Personal Computing model itself. This could be because Linux developers are often students or outside of mainstream MIS organizations.
The bottom line is that there's alot of tail chasing going on. By the time a "free" Office Suite gets to an acceptable function point and can import MS Office documents, it will already be irrelvant because those Office documents won't be "documents" any more, and instead will be locked up in Microsoft DNA-based network systems and will only be accessible by special network protocols and MS presentation software.
I don't think anybody has a real good solution to this problem. Bits and pieces are out there (CORBA, XML, etc.), but someone has to tie them together in a way that cheap enough and accessible enough for the desktop user. Unfortuantly, that someone is probably going to be Microsoft, and folks aren't going to be happy with the results.
Why does Microsoft have to lose for us to win?
/. posts indicate that people are taking it very seriously.
Well, for the vast majority of users, the operating system is whatever their employer says it is. This makes desktop operating systems somewhat of a zero-sum game. Either they are "the standard" or they are only in place for nitch uses. (And if you are a 'nitch' Linux user in an all-Windows desktop shop, more power to you - just realize that you are 1/1000th of the user base.)
If "world domination" is ever reached, it means that thousands of Windows computers are going to have to be reformatted and have Linux installed. This assumes that Linux is an adequate replacement for Windows on the desktop. Meanwhile, Microsoft has been working for years to promote Windows as a general solution So, you can see where the contention arises ("Windows Everywhere" versus "World Domination")
And, yes, I realize that World Domination was originally somewhat of a joke, but thousands of
(A) I don't play games, so I can't and didn't say what's better.
(B) Microsoft can't force game companies to do anything. Do you really think John Carmack cares what MS's policy is? Windows games are attractive because the burden of writing device drivers is off the game developers' backs.
(C) You can't spell idiot. Nice try.
I would actually argue that retrofitting a game machine into a computer did work for the Amiga, as it did for Atari with both the Atari 800 and ST lines.
These systems eventually died, but that was due a bad business model more than anything. (Sell at a thin margin, invest no money in R+D.) Sony can afford to subsidize these things with both game/software sales, and the 'pro' video-editing applications.
Sure, few "academics and scientists" took the Amiga seriously. It wasn't marketed at them - it was marketed to gamers and video people, who did take it very seriously. On the other hand, Windows PCs were designed as machines to run Lotus 1-2-3 and MS Word, yet they are constantly being kludged with things like DirectX and Windows 98 to use the great gaming hardware available.
NTSC to HDTV converters are going to be in such demand that hopefully Radio Shack will be selling them for $20 for a long time to come. (After all, they still sell 8 Track stuff!)
Well at least I hope so. There's an enormous investment out there in NTSC beyond Ataris and Playstations, even if it's not being broadcast.
At some point you have to tell people to fix their broken code because the bugs, compatibility hacks and obsolete features their software depends on have been deprecated and will disappear in the next major release.
Except that time and time-again the PC userbase has chosen backwards-compatibility over a better product. See OS/2. See Windows NT. Both products had pretty good backwards-compatibilty, but not good enough for people to flock to them in great numbers on the desktop. Microsoft has to sell upgrades, and therefore backwards compatiblity is job #1. (Prediction: Watch Win2000 beta get watered down over the next few months to try to get ye olde software working.)
The root problem is the broken applications. There are just too damn many 'business critical' applications in use for which there is no longer vendor support or perhaps no longer a vendor. (And no source code, either, so forget about fixing the problem.)