XT: expanded memory. This was a card. These sorts of cards existed for CP/M systems. You didn't need DOS 4 the applications were responsible for mapping the card's memory into memory. The first standardized card, that is card with OS support for expanded memory was April 1985.
286: extended memory. You could run in 286 mode and use up to 16 megs. Also DOS would allow 286 mode programs. But you could also use extended memory in place of expanded. http://en.wikipedia.org/wiki/Qemm. This worked fine on DOS 3.31 and possibly earlier versions. I know Microsoft introduced their own EMS drivers with DOS 4.01 but I didn't use those so I have no idea how they worked.
DOS Protected Mode Interface. the 0.9 version was included in DOS from 1990 on (and of course you could get it via. FidoNET or other BBSes) and heavily used prior. This standardized Protected mode programs but Lotus didn't follow the standards.
The later version that was used inside Windows 3.0 made this even more popular, though it was a DOS function.
DRDOS 5 and MSDOS5 -- allowed real mode DOS applications to load some drivers into hi memory area using a virtualized memory map. Really really cool, really really useful. Nothing to do with protected mode and 32 bit DOS apps.
DOS32 I'm not sure what you mean by this. I think we are using the word to mean two different things since what I'm talking about would have been pointless by 1996.
What should be there is the ability to "apt-get install debian" on a Ubuntu system.
Well first off neither Fink nor Macports does anything remotely like that. Both of them require OSX to install their own initial layers and then only have a selection of packages. They don't offer guest OSes virtualized. Far less do they offer running alien software not virtualized. I think what you are asking for now is simply impossible. But even if I'm wrong what you are now demanding more and easier virtualization than even MVS (IBM Z-series) mainframes offer. So I'm going to start by saying, Linux package management doesn't suck because it fails to provide a system that no one has ever successfully implemented on any operating system ever designed.
Second since you want to repoint it you would be using dpkg not apt. dpkg does do a lot of what you want, but you do need to learn to use the Linux package management tools that exist.
Third Unix has never really offered this functionality at all The only system I know that's still sold that does a good job at this is IBM's System-Z. But even here Linux is on par with most Unixes and for example Red Hat Linux Advanced Server makes some attempt to offer basics. Canonical's desktop is not going to offer this sort of functionality.
Linux package management doesn't offer teleportation as a standard feature either.
I found a lower bound. March 13, 1988 there is an article in the NYTimes about the upcoming Lotus 1-2-3 3.0. So at this point we are down to somewhere between March and May of 1988 when DOS32 came into use.
They didn't have to crawl or beg, they just asked and Microsoft said yes. Microsoft was anxious to support Ubuntu since they don't want a repeat of the paranoia that surrounded Palladium.
It'd be a lot easier to accept if the task were granted to a company with no stake in the OS market, like Intel.
Most likely there are going to be about 6 signing authorities on the BIOS that ship. Microsoft, someone like Verisign, a few Asian ones, maybe the hardware vendors themselves (i.e. Dell signs for UEFI in Dell's and collects the check). There is no reason to believe Intel, Western Digital (which has played for open standards for decades) or someone unexpected like NVidia won't step forward. I could see IBM who is much more trusted by the Linux community doing it.
Lotus started playing around with expanded memory with 1-2-3 version 2.0 which ran on the XT. When the 286s came out with the capability for using expanded memory to emulate extended memory (cheaper hardware, better performance) this became popular. At the time of Windows 3.0 Lotus was focused on extending their DOS product. As far as when you could run 32 bit DOS apps, I can find the hard dates. But I was using DOS32 around Nov 88. DOS5 supported all sorts of 32 bit extensions and that was June '89. DRDOS was earlier had all sorts of additional support for DOS32 was May '88. So I think we are talking 87-89 not mid 90s.
I don't know what you mean by DOS32 in the mid 1990s unless you mean the NT version that was emulated.
The kinds of people that are going to like Linux are the sorts of people that would love screw with their BIOS. They are generally people who have been messing with settings for years prior to moving over.
Heck it might turn out to be a selling feature that "Linux lets you take back control of your hardware"
AC is right. They want both ways. But obviously their hypervisor only runs with Linux. Its VMWare or Xen or... that handle it if you are running Linux as the dominant system.
Now of course their real favorite is Windows underneath running Windows images.
I agree with you, there is a lot of paranoia about this. There are a couple of things that are getting mixed in people's minds:
1) A failure to understand signing authorities and how the system works. Thus often the difficulty between getting something signed and being an authority gets conflated. 2) The FSF concerns about moving towards a world of trusted computing, what's possible vs. what is being proposed. 3) An analogy with Apple and a failure to understand how Apple works leading to a view that Apple is more closed than it is. 4) A general mistrust of Microsoft
Of course they will they already do. The ribbon changes based on what sort of editing you are doing. The ribbon allows for more of that. That's called context sensitivity. It also allows for deeper embedding.
Its $80 to get Microsoft to sign a kernel for you. This is not out of reach of an individual. Despite all the paranoia Microsoft seems to (at least so far) not be doing anything tricky.
No it isn't a huge waste of time. 1) Secure boot loaders 2) hardware encryption 3) trusted computing capable kernels
are the 3 things you need for trusted computing. Consumer media is still dicey because of piracy. And lots of business would love secure documents. That's worth a bit of a pain in the ass to them.
Apple doesn't lock other OSes from Apple hardware. They in fact write a multi OS bootloader (Bootcamp) and give it away free to make it easy for people to install other OSes. They work with VMWare and Parallels for people who want to run OSes in VMs.
I hope you see the connection between still being on XP and Apple being in 1/2 the meetings. What's happened is IT is refusing to upgrade in a timely manner so end users are upgrading themselves.
However it seems clear to me the that the desktop as a common way of doing computing, is on the way out. It had a good run. I just hope kids in 10 years have a way to experiment with building and modifying their own computing power like I did.
No they won't. The golden age of kids modifying the computers was the late 70s when they were hobbyist systems. Even Linux today is too complex for kids to change. The computer languages for systems are less fun, though web languages and scripting languages work well.
1) Total size of the market. In which case most of those factors are irrelevant. As an aside the average phone lasts 11.5 months. What drives the difference in replacement cycle is much higher breakage for smartphones.
2) Size of the software market. 3) Relative rate of growth of the different markets. 4) Time spent using the device. 5) % of the population that owns or regularly uses a device.
As for gaming rigs, they aren't enough of a market at this point to be a huge influence. In a few years they may be.
In the times of WoW and before it generally wasn't an IT department that set them up. IT controlled the mainframe or the mini. They didn't have responsibility for other office equipment like typewriters or photocopiers. PCs were classified with other office equipment when they first showed up for most companies. IT started getting involved once people realized that PCs were fundamentally different than terminals in that the corporate data was on them and functionality was migrating off the mainframes entirely.
I agree with your analogy but your history is a bit screwed up.
Visicalc was popular with the late 70s era machines, CP/M and AppleDOS. It had made it over to PC's (PCDOS later MSDOS) but was underpowered. In 1983 Lotus came out with Lotus 1-2-3 and took the spreadsheet market. Lotus had banked heavily on DOS32 (running DOS apps in 32 bit mode) being the dominant speadsheet and had treated their Windows product as secondary. So during the transition to Windows Lotus fell behind Excel though their suite, the AmiPro suite, was excellent by the later Windows 3.1 days. Excel also beat Lotus on price.
While PCs had come into mid and large business in limited capacity the switch for the average corporate worker happened during the Windows for Workgroups days. By that point there was no visicalc at all.
How can that be? There are regulations that mandate certain controls for the radio. No one cares about the apps but the actual phone / data part is heavily regulated and OSes and carriers have to be certified. I've heard this before and I still don't buy it. Do you have any good information on what they mean by fully open bootloader?
First off Apple's share of the desktop market in the USA is 8-12% which is about where it was when Microsoft was considered a monopoly. Microsoft's defense at this point might be the existence of a tablet market where they have no presence. But even if one does include tablets Windows still far outsells iOS and OSX combined. Apple targets profitable customers not marketshare.
As for Apple restricting boot. No they don't. In fact they produce and support a multi-platform bootloader for their computers: http://www.apple.com/support/bootcamp/ They also work with parallels and VMware to help people load virtual images of windows. Apple doesn't mind in the slightest if you buy their hardware and then run someone else's OS on it.
On their iOS devices, iTunes allows you to put any BIOS image in you want.
Now you are wrong about the ribbon, only because you are not taking into account office jocks have been using office for over a decade and know it like the back of their hand. The ribbon blows muscle memory all to shit and I've watched as people that could fly on 2K3 were brought to a screeching halt thanks to the ribbon. Sure its great if you've never used office before, but that isn't their biggest demographic is it? IMHO they should have had a switch at install that let the user choose which layout to have along with a GPO so it could be deployed across the network in whichever config the IT dept wanted.
The old menu system wasn't fundamentally different than the ansi menus from CP/M apps that ran in 32k or ram. The only major upgrade had been: menu items that invoked setting's panes. They needed to force the switch since their intent at the time was to have people undergo the switching cost prior to them making the ribbon menus have more levels and be more context sensitive and thereby allowing for a huge increase in the feature set of office. They had to get their entire base unified on using ribbon before they started to really take advantage of ribbon.
Well first off most people capable of a server install will be capable of disabling UEFI or self signing so my inclination is no. Right now this is mainly being pushed as a desktop feature. On the other hand once implemented there is no reason that it couldn't be a server feature. Servers are always going to be more diverse hardware and server installs always more complex so people who make server class hardware are likely to offer better instructions for over riding.
I honestly think the Linux desktop people are too worried here.
Let me just say under your theory where Nokia's existing business was worth essentially nothing the best move for Nokia was not to go into Windows phones but sell their patents and sales relationship and exit the business. Under your theory 2 years ago Nokia was horribly overvalued.
I don't think the situation was nearly that desperate. I think people really did love Nokia phones and they really do have a loyal customer base in dumb phones. Conversely I don't think they have a customer base for smart phones. I think that they needed to grow and so Nokia like products, like the N9 were the right way to go. In other words that prior to Elop Nokia had an execution problem but not deep deep problems. That is the "burning platform" was false.
Conversely if the burning platform were true and the dumbphone business were dead, then the Windows strategy wasn't a terrible one. Though I don't understand why Elop wouldn't allow the N9 strategy at the same time. See which one worked.
Have you ever done Symbian development? Have you done Symbian development and iPhone or Android (or hell blackberry) development? Symbian is soul crushingly terrible. I did a Symbian project in 2008 and people were thrilled to be onto a Blackberry project by 2009 when everyone knew BB was doomed (and still is), because that was at least better to work with than Symbian. The developer money is in ease of development and the App store model, and with those, Apple (and then Google) blew Symbian away.
No I've never done Symbian development and I've heard bad things about it. Your phrasing was amusing but you aren't the only one whose said that sort of stuff. OTOH I've done QT development and QT is lots of fun.
e idea of distributions was never to make a choice and force it on it's user,
Well they can't force anything on the user but yeah that was the idea. The very first Linux distributions like MCC focused on backup. Yggdrasil Linux came configured, the goal was a bootable Linux not choice. Caldera's aim was to create a system for Novell there was no configuration. Then later to create a desktop and they most certainly did pick software in that desktop. Then a unified platform for VARs and embedded system.
There are distributions like Debian which just try and act like a large collection of packages.
____
The closest thing to what I want is actually outside of the Linux world: Fink and MacPorts.
Well Fink's interface is taken from Debian source. You can do what you do with Fink with Debian source. MacPorts is based on the BSD ports system, which is what Gentoo is based on.
And that's been my point, what you want exists in Linux.
Don't forget every app gets approved or not in 6 regions and those regional approvals often involve multiple country approvals. I don't know what it costs, but at least Apple's claim is that they aren't really making that much on it. But that was back a year or two. The numbers are getting too large for that to be plausible anymore.
Slashdot Mirror
You are getting confused here.
XT: expanded memory. This was a card. These sorts of cards existed for CP/M systems. You didn't need DOS 4 the applications were responsible for mapping the card's memory into memory. The first standardized card, that is card with OS support for expanded memory was April 1985.
286: extended memory. You could run in 286 mode and use up to 16 megs. Also DOS would allow 286 mode programs. But you could also use extended memory in place of expanded. http://en.wikipedia.org/wiki/Qemm. This worked fine on DOS 3.31 and possibly earlier versions. I know Microsoft introduced their own EMS drivers with DOS 4.01 but I didn't use those so I have no idea how they worked.
DOS Protected Mode Interface. the 0.9 version was included in DOS from 1990 on (and of course you could get it via. FidoNET or other BBSes) and heavily used prior. This standardized Protected mode programs but Lotus didn't follow the standards.
The later version that was used inside Windows 3.0 made this even more popular, though it was a DOS function.
DRDOS 5 and MSDOS5 -- allowed real mode DOS applications to load some drivers into hi memory area using a virtualized memory map. Really really cool, really really useful. Nothing to do with protected mode and 32 bit DOS apps.
DOS32 I'm not sure what you mean by this. I think we are using the word to mean two different things since what I'm talking about would have been pointless by 1996.
What should be there is the ability to "apt-get install debian" on a Ubuntu system.
Well first off neither Fink nor Macports does anything remotely like that. Both of them require OSX to install their own initial layers and then only have a selection of packages. They don't offer guest OSes virtualized. Far less do they offer running alien software not virtualized. I think what you are asking for now is simply impossible. But even if I'm wrong what you are now demanding more and easier virtualization than even MVS (IBM Z-series) mainframes offer. So I'm going to start by saying, Linux package management doesn't suck because it fails to provide a system that no one has ever successfully implemented on any operating system ever designed.
Second since you want to repoint it you would be using dpkg not apt. dpkg does do a lot of what you want, but you do need to learn to use the Linux package management tools that exist.
Third Unix has never really offered this functionality at all The only system I know that's still sold that does a good job at this is IBM's System-Z. But even here Linux is on par with most Unixes and for example Red Hat Linux Advanced Server makes some attempt to offer basics. Canonical's desktop is not going to offer this sort of functionality.
Linux package management doesn't offer teleportation as a standard feature either.
I found a lower bound. March 13, 1988 there is an article in the NYTimes about the upcoming Lotus 1-2-3 3.0. So at this point we are down to somewhere between March and May of 1988 when DOS32 came into use.
They didn't have to crawl or beg, they just asked and Microsoft said yes. Microsoft was anxious to support Ubuntu since they don't want a repeat of the paranoia that surrounded Palladium.
It'd be a lot easier to accept if the task were granted to a company with no stake in the OS market, like Intel.
Most likely there are going to be about 6 signing authorities on the BIOS that ship. Microsoft, someone like Verisign, a few Asian ones, maybe the hardware vendors themselves (i.e. Dell signs for UEFI in Dell's and collects the check). There is no reason to believe Intel, Western Digital (which has played for open standards for decades) or someone unexpected like NVidia won't step forward. I could see IBM who is much more trusted by the Linux community doing it.
Lotus started playing around with expanded memory with 1-2-3 version 2.0 which ran on the XT. When the 286s came out with the capability for using expanded memory to emulate extended memory (cheaper hardware, better performance) this became popular. At the time of Windows 3.0 Lotus was focused on extending their DOS product. As far as when you could run 32 bit DOS apps, I can find the hard dates. But I was using DOS32 around Nov 88. DOS5 supported all sorts of 32 bit extensions and that was June '89. DRDOS was earlier had all sorts of additional support for DOS32 was May '88. So I think we are talking 87-89 not mid 90s.
I don't know what you mean by DOS32 in the mid 1990s unless you mean the NT version that was emulated.
The kinds of people that are going to like Linux are the sorts of people that would love screw with their BIOS. They are generally people who have been messing with settings for years prior to moving over.
Heck it might turn out to be a selling feature that "Linux lets you take back control of your hardware"
AC is right. They want both ways. But obviously their hypervisor only runs with Linux. Its VMWare or Xen or... that handle it if you are running Linux as the dominant system.
Now of course their real favorite is Windows underneath running Windows images.
I agree with you, there is a lot of paranoia about this. There are a couple of things that are getting mixed in people's minds:
1) A failure to understand signing authorities and how the system works. Thus often the difficulty between getting something signed and being an authority gets conflated.
2) The FSF concerns about moving towards a world of trusted computing, what's possible vs. what is being proposed.
3) An analogy with Apple and a failure to understand how Apple works leading to a view that Apple is more closed than it is.
4) A general mistrust of Microsoft
Of course they will they already do. The ribbon changes based on what sort of editing you are doing. The ribbon allows for more of that. That's called context sensitivity. It also allows for deeper embedding.
Its $80 to get Microsoft to sign a kernel for you. This is not out of reach of an individual. Despite all the paranoia Microsoft seems to (at least so far) not be doing anything tricky.
Or you pay $80 and get signed. The signing authorities will sign.
No it isn't a huge waste of time.
1) Secure boot loaders
2) hardware encryption
3) trusted computing capable kernels
are the 3 things you need for trusted computing. Consumer media is still dicey because of piracy. And lots of business would love secure documents. That's worth a bit of a pain in the ass to them.
Apple doesn't lock other OSes from Apple hardware. They in fact write a multi OS bootloader (Bootcamp) and give it away free to make it easy for people to install other OSes. They work with VMWare and Parallels for people who want to run OSes in VMs.
None of what you are saying is true.
I hope you see the connection between still being on XP and Apple being in 1/2 the meetings. What's happened is IT is refusing to upgrade in a timely manner so end users are upgrading themselves.
However it seems clear to me the that the desktop as a common way of doing computing, is on the way out. It had a good run. I just hope kids in 10 years have a way to experiment with building and modifying their own computing power like I did.
No they won't. The golden age of kids modifying the computers was the late 70s when they were hobbyist systems. Even Linux today is too complex for kids to change. The computer languages for systems are less fun, though web languages and scripting languages work well.
It comes down to what you want to measure:
1) Total size of the market. In which case most of those factors are irrelevant. As an aside the average phone lasts 11.5 months. What drives the difference in replacement cycle is much higher breakage for smartphones.
2) Size of the software market.
3) Relative rate of growth of the different markets.
4) Time spent using the device.
5) % of the population that owns or regularly uses a device.
As for gaming rigs, they aren't enough of a market at this point to be a huge influence. In a few years they may be.
GP, johnw, is absolutely right.
In the times of WoW and before it generally wasn't an IT department that set them up. IT controlled the mainframe or the mini. They didn't have responsibility for other office equipment like typewriters or photocopiers. PCs were classified with other office equipment when they first showed up for most companies. IT started getting involved once people realized that PCs were fundamentally different than terminals in that the corporate data was on them and functionality was migrating off the mainframes entirely.
I agree with your analogy but your history is a bit screwed up.
Visicalc was popular with the late 70s era machines, CP/M and AppleDOS. It had made it over to PC's (PCDOS later MSDOS) but was underpowered. In 1983 Lotus came out with Lotus 1-2-3 and took the spreadsheet market. Lotus had banked heavily on DOS32 (running DOS apps in 32 bit mode) being the dominant speadsheet and had treated their Windows product as secondary. So during the transition to Windows Lotus fell behind Excel though their suite, the AmiPro suite, was excellent by the later Windows 3.1 days. Excel also beat Lotus on price.
While PCs had come into mid and large business in limited capacity the switch for the average corporate worker happened during the Windows for Workgroups days. By that point there was no visicalc at all.
How can that be? There are regulations that mandate certain controls for the radio. No one cares about the apps but the actual phone / data part is heavily regulated and OSes and carriers have to be certified. I've heard this before and I still don't buy it. Do you have any good information on what they mean by fully open bootloader?
First off Apple's share of the desktop market in the USA is 8-12% which is about where it was when Microsoft was considered a monopoly. Microsoft's defense at this point might be the existence of a tablet market where they have no presence. But even if one does include tablets Windows still far outsells iOS and OSX combined. Apple targets profitable customers not marketshare.
As for Apple restricting boot. No they don't. In fact they produce and support a multi-platform bootloader for their computers: http://www.apple.com/support/bootcamp/
They also work with parallels and VMware to help people load virtual images of windows.
Apple doesn't mind in the slightest if you buy their hardware and then run someone else's OS on it.
On their iOS devices, iTunes allows you to put any BIOS image in you want.
Now you are wrong about the ribbon, only because you are not taking into account office jocks have been using office for over a decade and know it like the back of their hand. The ribbon blows muscle memory all to shit and I've watched as people that could fly on 2K3 were brought to a screeching halt thanks to the ribbon. Sure its great if you've never used office before, but that isn't their biggest demographic is it? IMHO they should have had a switch at install that let the user choose which layout to have along with a GPO so it could be deployed across the network in whichever config the IT dept wanted.
The old menu system wasn't fundamentally different than the ansi menus from CP/M apps that ran in 32k or ram. The only major upgrade had been: menu items that invoked setting's panes. They needed to force the switch since their intent at the time was to have people undergo the switching cost prior to them making the ribbon menus have more levels and be more context sensitive and thereby allowing for a huge increase in the feature set of office. They had to get their entire base unified on using ribbon before they started to really take advantage of ribbon.
Secure boot is a hardware feature. They can't ignore Windows 8 and make it go away.
Well first off most people capable of a server install will be capable of disabling UEFI or self signing so my inclination is no. Right now this is mainly being pushed as a desktop feature. On the other hand once implemented there is no reason that it couldn't be a server feature. Servers are always going to be more diverse hardware and server installs always more complex so people who make server class hardware are likely to offer better instructions for over riding.
I honestly think the Linux desktop people are too worried here.
Let me just say under your theory where Nokia's existing business was worth essentially nothing the best move for Nokia was not to go into Windows phones but sell their patents and sales relationship and exit the business. Under your theory 2 years ago Nokia was horribly overvalued.
I don't think the situation was nearly that desperate. I think people really did love Nokia phones and they really do have a loyal customer base in dumb phones. Conversely I don't think they have a customer base for smart phones. I think that they needed to grow and so Nokia like products, like the N9 were the right way to go. In other words that prior to Elop Nokia had an execution problem but not deep deep problems. That is the "burning platform" was false.
Conversely if the burning platform were true and the dumbphone business were dead, then the Windows strategy wasn't a terrible one. Though I don't understand why Elop wouldn't allow the N9 strategy at the same time. See which one worked.
Have you ever done Symbian development? Have you done Symbian development and iPhone or Android (or hell blackberry) development? Symbian is soul crushingly terrible. I did a Symbian project in 2008 and people were thrilled to be onto a Blackberry project by 2009 when everyone knew BB was doomed (and still is), because that was at least better to work with than Symbian. The developer money is in ease of development and the App store model, and with those, Apple (and then Google) blew Symbian away.
No I've never done Symbian development and I've heard bad things about it. Your phrasing was amusing but you aren't the only one whose said that sort of stuff. OTOH I've done QT development and QT is lots of fun.
e idea of distributions was never to make a choice and force it on it's user,
Well they can't force anything on the user but yeah that was the idea. The very first Linux distributions like MCC focused on backup. Yggdrasil Linux came configured, the goal was a bootable Linux not choice. Caldera's aim was to create a system for Novell there was no configuration. Then later to create a desktop and they most certainly did pick software in that desktop. Then a unified platform for VARs and embedded system.
There are distributions like Debian which just try and act like a large collection of packages.
____
The closest thing to what I want is actually outside of the Linux world: Fink and MacPorts.
Well Fink's interface is taken from Debian source. You can do what you do with Fink with Debian source.
MacPorts is based on the BSD ports system, which is what Gentoo is based on.
And that's been my point, what you want exists in Linux.
btw fink defaults to /sw and macports to /opt/local
Don't forget every app gets approved or not in 6 regions and those regional approvals often involve multiple country approvals. I don't know what it costs, but at least Apple's claim is that they aren't really making that much on it. But that was back a year or two. The numbers are getting too large for that to be plausible anymore.