What keeps you motivated to work for so long on this project (no criticism intended) while there are so many fascinating newer technologies to work on instead?
After getting out of the mainstream DOS lived on, mainly in embedded systems with minimal hardware or low resources. Anything embedded today is dominated by ARM designs, running in hardware much more powerful than required by DOS back in the day.
With Ghz CPUs and GB RAM costing so little today, have DOS at long last lost its purpose?
by Anonymous Coward writes:
on Monday June 10, 2019 @07:44AM (#58738392)
FreeDOS, as I understand it, is meant to be compatible with MS-DOS 6. Is it tricky to maintain this legacy compatibility while also trying to improve, or modernize, FreeDOS? I see a lot of utilities are developed for the OS and I'm guessing you get a lot of requests to port FreeDOS to modern CPUs, or add advanced features. So how do you decide what can go into FreeDOS and what is kept out to maintain backward compatibility?
Thanks for making FreeDOS. It and related projects have kept a handful of classic DOS applications and games running for me for years.
DOS isn't that complex of an OS. There are no drivers per say only TSR which often add common controls for a mouse. But DOS mostly just handles the raw machine code, and depending if it is an exe or com file will split the memory command into chunks so larger programs can be run.
During the golden days of DOS Games, we went up VGA 320x200 265 color graphics because that was the last fully standard video card, SVGA started having variances in their standard making difficult for MSDOS to handle.
But trying to make DOS to support 64bit multi-core CPU's with Gigs of RAM, and Terabytes of disk space while possible, will cause problems later on. Unless it is designed to be more complex, then it wouldn't really be DOS anymore, and you would need DOSBox for your DOS environment ro run you old DOS Apps.
Multi-core I understand, since DOS is a single CPU OS. But why would a simple extension of the code from 16 to 32 or 64 bits be problematic? Keep the 16-bit part of it as is, so that it can still run on a PC from 1989, but add maybe a 32-bit version based on just, say, the 386SX instruction set (assuming here that DOS doesn't need a 387) and a 64-bit version based on the first Athlon-64 CPU (not using the instructions that are different b/w Intel and AMD). On these longer versions, have support for 4GB o
Add a few things like long filename support, Powershell commands and so on.
I like your thinking, with PowerShell, but there's a ton of.NET underneath each cmdlet that would make porting all but impossible. Sure, you could edit COMMAND.COM and alias gci and ls to dir, but that's about it, barring one heck of a DOS extender.
FreeDOS is still in use for some firmware updaters, or is a practical way to run some DOS-based updaters. So, did it bring you any nice corporate support, whether it was patches, man-days or just money to improve or maintain FreeDOS?
Back in the '90s and early millennium, I could see the need for FreeDOS and appreciated the passion you put into now.
Now, with Linux being very mature and having CLI only options (including WINE) for minimal system footprints, does it make sense to continue with FreeDOS?
If it does, could you explain why and when, if ever, you can see support ending for FreeDOS?
WINE is only a compatibility layer for Windows, for DOS applications you need something like FreeDOS. It works very well if you want to play games from the 1980s or early 90s. There are plenty of games on places like gog.com that actually ship with FreeDOS so they're playable on modern systems. It's also pretty useful for bootdisks to install firmware updates.
Now, with Linux being very mature and having CLI only options (including WINE) for minimal system footprints, does it make sense to continue with FreeDOS?
Statistically nobody is using DOS because it's small. They're using it for DOS compatibility. You could run Unixlikes on a 286! I used to have a 286@6MHz with 1MB of RAM and a 40MB RLL disk, and I ran Xenix 2.3.2 on it as a UUCP node. It was flawless, literally never giving me so much as a moment's grief.
Hmmmm.... sounds like an old AT&T6300 vintage computer I used to have. It actually had a problem running DOS because the timer chip wasn't generating interrupts at exactly the same number of interrupts per second that DOS expected every computer to have.
Since you started very early in the history of personal computing, have you noticed a degradation of coding 'cleverness' (for want of a better word), as hardware processing capabilities, network speeds, and power consumption increased ?
There is... there are plenty. Some have a small C layer, some compile to native code, the RVM e.g. is completely self hosting.No idea about GraalVM, but I guess they are close.
Point is: if you insisted to write an OS in Java, you just make sure that the relevant parts can be compiled to native machine code, problem solved. Nothing in Java or its toolchain prevents you doing that, see e.g. AviaVM.
All have a C or assembly layer or similar, there are certain kinds of machine code java can't emit and that the language can't handle. Specific memory locations, interrupts, register handling, protected mode...ain't happening in Java without calling low level language.
there are certain kinds of machine code java can't emit and that the language can't handle. That is nonsense. A compiler can emit what ever code the compiler writer wants...
Ideally if POSIX were properly updated, Java would be the way to go.
Maybe not Java now that Oracle has completely curb-stomped it, but some form of bytecode that can run in an LLVM, or can be coded in a machine-compilable language like C when it matters.
With Operating Systems likely to be moving to the Cloud (Microsoft has already indicated that they are going to do this), do you see locally run full OSes to have no future? Will most people migrate to a cloud-based operating system?
I've had to go in hiding, keeping the shades drawn on my homes and traveling with Groucho Marx disguise glasses and nose, which is difficult at airports. It's especially hard on the wives since they really don't like using the same disguise (women). The kids are dressed as young Bill Gates (even the girls) since it's assumed that Bill won't go after someone who looks like him, but then I remember what he did to Ballmer and cringe at the possibility. It's been a scary life to say the least. My best bud, J.O., w
by Anonymous Coward writes:
on Monday June 10, 2019 @09:06AM (#58738644)
DOS relies on code built into the BIOS, and BIOS is starting to become unavailable in newer hardware. I noticed the last motherboard I bought was UEFI only. The FreeDOS website says that FreeDOS cannot emulate BIOS.
So what's the long term plan for FreeDOS? Will FreeDOS die, or will it only run on older systems that support BIOS, or inside a VM?
If I had to guess, and I do because we don't have the answers yet, I'd guess that most use of FreeDOS is inside of VMs.
I'd further guess that most non-VM use of FreeDOS is on antiquated hardware, the kind of hardware they'll simply keep putting BIOS on into the future. You know, passive backplane motherboards, PC-104-bus embedded PCs, that kind of thing.
" What is the justification, the value to society, of a FreeDOS project in 2019?
You said it:
"I understand its history, personal significance, and emotional attachments."
What more do you need? Maybe it's fun. Why do I have a 55 year old tube-based oscilloscope? Because of its 3.9GHz sampling plug-in? Or because it's fun to tool around in living history?
We use FreeDOS in all of our systems. It is a deterministic operating system. We build industrial laser systems with high speed beam steering systems that controls the pulse duration, pulse period, and linear speed of the beam in real time while tracking the part through space. Our current systems have I7 processors and required a custom bios to allow for upper memory ( the remain memory above 640k below the 1mb mark). We do trig math in between motion updates that happen at 100khz. I have test many oth
I run FreeDOS on my P133. Why do people drive around 25 year old cars? Because not enough has changed in technology to justify throwing it away. Why replace my software library when what I have works? I mainly use my PC as a glorified typewriter anyways. If I need to do serious scientific research I'll let my employer pay for it or rent time on the cloud.
I've seen FreeDOS in elevators and drive-through order screens, but I imagine most of those will be migrating to Linux.
As long as you can get rugged PC hardware about as cheap (or cheaper!) than non-PC rugged hardware, DOS will continue to be viable for such simple tasks.
Have you tried to contact the companies that hold the rights to classic DOS tools like Norton Commander, Norton Disk Doctor, PC Tools, DiskDupe, Laplink etc? Maybe some would be willing to release their old DOS software as a goodwill gesture?
Just wondering, as hardware has advanced, have there been a lot of times where some new notionally-compatible CPU just totally breaks things and you have to tweak things to make FreeDOS run on it?
(Also, what's it like to have a Slashdot ID under 3,000?)
My question: Is there a way to keep the DOS tradition, in a hardware world that is increasingly divorced from the actual hardware?
Ex: An IBM PC in 1984 has BASICA in ROM, and you could write applications out of the box, to do actual hardware operations, even if they crashed. Now, there's is no standard hardware/software support included with any operating system that doesn't require elaborate drivers to function, and no impetus to write drivers for old or otherwise unsupported hardware. The latest OS with support for some of my cards was Linux, and that ended in 98 or so; and driver virtualization and timing skews don't really allow mutitasking type OS'es to work anyway.
FreeDOS still works fine, however.:)
I still run old hardware, mostly to use some data acquisition hardware that either no longer exists, or would be incredibly expensive to replace.
I have various setups doing research, that take data continuously, and have been running for years.
Some of the hardware will run under Win95 or 98, but for solutions that are timing sensitive, DOS is still the best.
I've found that different old DOS versions all have their quirks, so standardizing on FreeDOS makes things simpler, and if I have to help someone out with old hardware, there's no licensing embuggerences to deal with.
Setting up EISA hardware Requires DOS, of some flavor, as well as most Adapted SCSI raid setups, so this won't go away until the hardware does.
First, I want to thank the people who have kept FreeDOS going. FreeDOS is a great embedded platform, and with some basic C coding, it can be used to handle realtime applications with ease. For a very limited environment, writing code and using libraries to use FreeDOS as a RTOS is a useful thing.
One thing that would be nice to see. Something like DoubleSpace/DriveSpace/Stacker implemented. The patents on these things have long since expired, and it would be nice to have some volume compression, especially in some embedded environments where space is at a premium.
How much commercial support have to had for FreeDOS? By that I mean code contributions from businesses, financial support for yourself, that sort of thing.
There needs to be a very strong motivation, or even more than just one, among the freedos team to make this project so far and beyond. Please, explain them to us all.
One of the limitations of DOS back in the day was the 640kB limit. There were things like DOS extenders and so on, but essentially, since the OS was built on a 286, it was 16 bit and that was it: hence the memory limitations.
Has it ever been contemplated extending the OS to either 32 or 64 bits since we had several generations of x86 CPUs? Maybe a 32-bit version w/ just the 386 instruction set and maybe a 64-bit version w/ either the AMD64 or Intel 64 instruction sets? Something where one could still d
What's the biggest deployment or use for FreeDOS that you are aware of? And what would you say the current active use cases this is being deployed today, in 2019?
Long Filename support (Score:5, Interesting)
How hard was it to add support for Long Filenames (DOSLFN) and what were the lessons learnt?
Thanks for making FreeDOS!
Motivation (Score:5, Interesting)
Best use of FreeDOS (Score:5, Interesting)
Purpose in a ARM world (Score:1)
After getting out of the mainstream DOS lived on, mainly in embedded systems with minimal hardware or low resources. Anything embedded today is dominated by ARM designs, running in hardware much more powerful than required by DOS back in the day.
With Ghz CPUs and GB RAM costing so little today, have DOS at long last lost its purpose?
Have you looked at FDPP ? (Score:2, Interesting)
FDPP aims to be the DOS used for DOSEMU. Basically a protected mode way of running DOS. Have you looked at the project ?
Compatible versus modern (Score:5, Interesting)
FreeDOS, as I understand it, is meant to be compatible with MS-DOS 6. Is it tricky to maintain this legacy compatibility while also trying to improve, or modernize, FreeDOS? I see a lot of utilities are developed for the OS and I'm guessing you get a lot of requests to port FreeDOS to modern CPUs, or add advanced features. So how do you decide what can go into FreeDOS and what is kept out to maintain backward compatibility?
Thanks for making FreeDOS. It and related projects have kept a handful of classic DOS applications and games running for me for years.
Re: (Score:2)
MSDOS had later versions delivered with Windows 95/98/ME.
Re: (Score:2)
DOS isn't that complex of an OS. There are no drivers per say only TSR which often add common controls for a mouse. But DOS mostly just handles the raw machine code, and depending if it is an exe or com file will split the memory command into chunks so larger programs can be run.
During the golden days of DOS Games, we went up VGA 320x200 265 color graphics because that was the last fully standard video card, SVGA started having variances in their standard making difficult for MSDOS to handle.
But for many
Re: (Score:2)
But trying to make DOS to support 64bit multi-core CPU's with Gigs of RAM, and Terabytes of disk space while possible, will cause problems later on. Unless it is designed to be more complex, then it wouldn't really be DOS anymore, and you would need DOSBox for your DOS environment ro run you old DOS Apps.
Multi-core I understand, since DOS is a single CPU OS. But why would a simple extension of the code from 16 to 32 or 64 bits be problematic? Keep the 16-bit part of it as is, so that it can still run on a PC from 1989, but add maybe a 32-bit version based on just, say, the 386SX instruction set (assuming here that DOS doesn't need a 387) and a 64-bit version based on the first Athlon-64 CPU (not using the instructions that are different b/w Intel and AMD). On these longer versions, have support for 4GB o
Re: (Score:2)
Add a few things like long filename support, Powershell commands and so on.
I like your thinking, with PowerShell, but there's a ton of .NET underneath each cmdlet that would make porting all but impossible. Sure, you could edit COMMAND.COM and alias gci and ls to dir, but that's about it, barring one heck of a DOS extender.
Corporate support (Score:2, Interesting)
FreeDOS is still in use for some firmware updaters, or is a practical way to run some DOS-based updaters. So, did it bring you any nice corporate support, whether it was patches, man-days or just money to improve or maintain FreeDOS?
When will support for FreeDOS end & why? (Score:2)
Back in the '90s and early millennium, I could see the need for FreeDOS and appreciated the passion you put into now.
Now, with Linux being very mature and having CLI only options (including WINE) for minimal system footprints, does it make sense to continue with FreeDOS?
If it does, could you explain why and when, if ever, you can see support ending for FreeDOS?
Thanx.
Re: (Score:1)
WINE is only a compatibility layer for Windows, for DOS applications you need something like FreeDOS. It works very well if you want to play games from the 1980s or early 90s. There are plenty of games on places like gog.com that actually ship with FreeDOS so they're playable on modern systems. It's also pretty useful for bootdisks to install firmware updates.
Re: (Score:3)
Now, with Linux being very mature and having CLI only options (including WINE) for minimal system footprints, does it make sense to continue with FreeDOS?
Statistically nobody is using DOS because it's small. They're using it for DOS compatibility. You could run Unixlikes on a 286! I used to have a 286@6MHz with 1MB of RAM and a 40MB RLL disk, and I ran Xenix 2.3.2 on it as a UUCP node. It was flawless, literally never giving me so much as a moment's grief.
Re: (Score:1)
Hmmmm.... sounds like an old AT&T6300 vintage computer I used to have. It actually had a problem running DOS because the timer chip wasn't generating interrupts at exactly the same number of interrupts per second that DOS expected every computer to have.
Is coding less elegant with more powerful pcs (Score:2, Interesting)
Since you started very early in the history of personal computing, have you noticed a degradation of coding 'cleverness' (for want of a better word), as hardware processing capabilities, network speeds, and power consumption increased ?
What Next (Score:5, Interesting)
Commercial uses (Score:1)
What commercial uses (of any) of FreeDOS are you aware of? For example, driver update disks.
Languages (Score:5, Interesting)
It looks like FreeDOS is mostly C and assembler. Have you been tempted to move to C++, or any other language?
Re: (Score:2)
But with ancient concepts still built in to POSIX, you end up back in the old standards.
What "ancient" concepts in POSIX 1) do you think should go and 2) prevent using Java?
Re: (Score:2)
lolz false
java can't function as an operating system on bare metal, you always need some kind of shim which would be your real kernel.
where do you java-tards get your stupid ideas, your language is unsuitable for the realm where C and assembler rule
Re: (Score:2)
There are plenty of Java compilers and OSes around them, that generate machine code.
When do people like you finally grasp that a language is just a language ...
Re: (Score:2)
there is no java OS that runs on bare metal.
Java is useless for such applications
Re: (Score:2)
There is ... there are plenty. Some have a small C layer, some compile to native code, the RVM e.g. is completely self hosting.No idea about GraalVM, but I guess they are close.
Point is: if you insisted to write an OS in Java, you just make sure that the relevant parts can be compiled to native machine code, problem solved. Nothing in Java or its toolchain prevents you doing that, see e.g. AviaVM.
Re: (Score:2)
All have a C or assembly layer or similar, there are certain kinds of machine code java can't emit and that the language can't handle. Specific memory locations, interrupts, register handling, protected mode...ain't happening in Java without calling low level language.
Re: (Score:2)
there are certain kinds of machine code java can't emit and that the language can't handle. ...
That is nonsense.
A compiler can emit what ever code the compiler writer wants
Re: (Score:2)
Ideally if POSIX were properly updated, Java would be the way to go.
Maybe not Java now that Oracle has completely curb-stomped it, but some form of bytecode that can run in an LLVM, or can be coded in a machine-compilable language like C when it matters.
The Cloud (Score:1)
With Operating Systems likely to be moving to the Cloud (Microsoft has already indicated that they are going to do this), do you see locally run full OSes to have no future? Will most people migrate to a cloud-based operating system?
Microsoft (Score:4, Interesting)
How has Microsoft responded to your project over the years?
Re: (Score:1)
I've had to go in hiding, keeping the shades drawn on my homes and
traveling with Groucho Marx disguise glasses and nose, which is difficult
at airports. It's especially hard on the wives since they really don't like
using the same disguise (women). The kids are dressed as young Bill
Gates (even the girls) since it's assumed that Bill won't go after someone
who looks like him, but then I remember what he did to Ballmer and cringe
at the possibility. It's been a scary life to say the least. My best bud, J.O.,
w
What happens to FreeDOS when BIOS is dead? (Score:5, Interesting)
DOS relies on code built into the BIOS, and BIOS is starting to become unavailable in newer hardware. I noticed the last motherboard I bought was UEFI only. The FreeDOS website says that FreeDOS cannot emulate BIOS.
So what's the long term plan for FreeDOS? Will FreeDOS die, or will it only run on older systems that support BIOS, or inside a VM?
Re: (Score:3)
If I had to guess, and I do because we don't have the answers yet, I'd guess that most use of FreeDOS is inside of VMs.
I'd further guess that most non-VM use of FreeDOS is on antiquated hardware, the kind of hardware they'll simply keep putting BIOS on into the future. You know, passive backplane motherboards, PC-104-bus embedded PCs, that kind of thing.
Re: (Score:1)
" What is the justification, the value to society, of a FreeDOS project in 2019?
You said it:
"I understand its history, personal significance, and emotional attachments."
What more do you need? Maybe it's fun. Why do I have a 55 year old tube-based oscilloscope? Because of its 3.9GHz sampling plug-in? Or because it's fun to tool around in living history?
Re: (Score:2)
Re: Free DOS in 2019 (Score:1)
I run FreeDOS on my P133. Why do people drive around 25 year old cars? Because not enough has changed in technology to justify throwing it away. Why replace my software library when what I have works? I mainly use my PC as a glorified typewriter anyways. If I need to do serious scientific research I'll let my employer pay for it or rent time on the cloud.
Support (Score:3)
who are the big users? (Score:4, Interesting)
Who are the most numerous users of FreeDOS? Is it gamers, or industrial control, or some other group I haven't imagined?
Re: (Score:2)
Re: (Score:2)
I've seen FreeDOS in elevators and drive-through order screens, but I imagine most of those will be migrating to Linux.
As long as you can get rugged PC hardware about as cheap (or cheaper!) than non-PC rugged hardware, DOS will continue to be viable for such simple tasks.
Q: Will this be the year of the FreeDOS desktop? (Score:5, Funny)
Re: (Score:2)
Classic 3rd party tools - abandonware? (Score:4, Interesting)
Have you tried to contact the companies that hold the rights to classic DOS tools like Norton Commander, Norton Disk Doctor, PC Tools, DiskDupe, Laplink etc? Maybe some would be willing to release their old DOS software as a goodwill gesture?
Does new hardware break things? (Score:1)
Just wondering, as hardware has advanced, have there been a lot of times where some new notionally-compatible CPU just totally breaks things and you have to tweak things to make FreeDOS run on it?
(Also, what's it like to have a Slashdot ID under 3,000?)
As a frequent user of FreeDOS... (Score:5, Insightful)
I'd like to start with saying Thanks!
My question: Is there a way to keep the DOS tradition, in a hardware world that is increasingly divorced from the actual hardware?
Ex: An IBM PC in 1984 has BASICA in ROM, and you could write applications out of the box, to do actual hardware operations, even if they crashed.
Now, there's is no standard hardware/software support included with any operating system that doesn't require elaborate drivers to function, and no impetus to write drivers for old or otherwise unsupported hardware.
The latest OS with support for some of my cards was Linux, and that ended in 98 or so; and driver virtualization and timing skews don't really allow mutitasking type OS'es to work anyway.
FreeDOS still works fine, however. :)
I still run old hardware, mostly to use some data acquisition hardware that either no longer exists, or would be incredibly expensive to replace.
I have various setups doing research, that take data continuously, and have been running for years.
Some of the hardware will run under Win95 or 98, but for solutions that are timing sensitive, DOS is still the best.
I've found that different old DOS versions all have their quirks, so standardizing on FreeDOS makes things simpler, and if I have to help someone out with old hardware, there's no licensing embuggerences to deal with.
Setting up EISA hardware Requires DOS, of some flavor, as well as most Adapted SCSI raid setups, so this won't go away until the hardware does.
Some features of FreeDOS and a request (Score:3)
First, I want to thank the people who have kept FreeDOS going. FreeDOS is a great embedded platform, and with some basic C coding, it can be used to handle realtime applications with ease. For a very limited environment, writing code and using libraries to use FreeDOS as a RTOS is a useful thing.
One thing that would be nice to see. Something like DoubleSpace/DriveSpace/Stacker implemented. The patents on these things have long since expired, and it would be nice to have some volume compression, especially in some embedded environments where space is at a premium.
Commercial support (Score:2)
How much commercial support have to had for FreeDOS? By that I mean code contributions from businesses, financial support for yourself, that sort of thing.
Temptations (Score:3, Interesting)
Which motivation? (Score:2)
There needs to be a very strong motivation, or even more than just one, among the freedos team to make this project so far and beyond.
Please, explain them to us all.
32 or 64 bit? (Score:2)
One of the limitations of DOS back in the day was the 640kB limit. There were things like DOS extenders and so on, but essentially, since the OS was built on a 286, it was 16 bit and that was it: hence the memory limitations.
Has it ever been contemplated extending the OS to either 32 or 64 bits since we had several generations of x86 CPUs? Maybe a 32-bit version w/ just the 386 instruction set and maybe a 64-bit version w/ either the AMD64 or Intel 64 instruction sets? Something where one could still d
active use cases (Score:2)
What's the biggest deployment or use for FreeDOS that you are aware of? And what would you say the current active use cases this is being deployed today, in 2019?
Psychological challenges (Score:1)
Did you ever face burnout, or troubles relating to dealing with thousands of requestsnfrom users, and how did you cope?
How are the cats doing? (Score:1)
How are your cats doing, and how do you keep them off of your keyboard?
Questions for Jim Hall about FreeDOS (Score:2)
Hello,
A few questions, if I might:
Thank you for taking the time to