Jump to content











Photo
* * * - - 1 votes

nx-dos


  • Please log in to reply
24 replies to this topic

#1 was_jaclaz

was_jaclaz

    Finder

  • Advanced user
  • 7100 posts
  • Location:Gone in the mist
  •  
    Italy

Posted 24 May 2008 - 03:06 PM

If anyone is interested, there is this open-source DOS:
http://sourceforge.net/projects/nxdos/

it's not clear if the release is 18 May 2007 or 18 May 2008, dates of files inside the archives suggest it is 2008 ;), it seems there like NO instructions or readme of any kind in the compiled download.....
....exception made for this FILEID.DIZ:

NX-DOS OS/3 (Alternative DOS) v8.01VI
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
An embedded MS-DOS 5.0 compatible
MINI RTOS for the IBM-PC Pentiums.
If you need a OS to run an embedded
appliance then this may suit you
needs.
-------------------------------------
UNIX name: nxdos
Descriptive name: NX-DOS Operating System Project
License: GNU General Public License (GPL)
Public description: NX-DOS Operating System
Registration description: The NX-DOS Project
is intended to be a clone of popular Disk
Operating Systems such as MS-DOS 5.0. I am using MS-DOS and PC-DOS as
design templates but decided that to go GPL and open it to others that
wish to work on it instead of just one programmer toiling away in his
room. Hopefully, with many minds this OS will be what I intend it to be. Currently, I have written a FDISK, SYS, COMMAND, KERNEL, BOOT LOADER, and other utilities. -ce, 4-1-2005
-------------------------------------
THIS SOFTWARE IS IN BETA DEVELOPMENT
READ ALL THE DISCLAIMERS PRIOR TO USE
(e.g. install on a crappy cyberdeck)
unarchive to a floppy and run SYS A:
-------------------------------------
Website: http://www.aotksc.com/
Email : tekno1911@yahoo.com


That appears to be remarkably out of date.

But some DOCS are in the "source" download, some are remarkably old, but the CHANGE.TXT confirms that development has continued recently:

; Change.log - reverse chronological order
MAY-12-2008 15:13 - ROMSCAN detect specific ROMs in computer
APR-27-2008 16:45 - Fixed the print() problem, ZSTART variable is now no longer boolean.
JUN-11-2007 - RTCupdate added to clock driver
APR-22-2007 17:10 - Finished nextmcb/firstmcb/minit/getlastarena/mmconfig in
- DOSMEM.ASM. Needs setup_arena(), malloc amd mfree.
APR-10-2007 22:00 - Using a new device driver design. wrote CLOCK$ handler,
now sets dos time values from INT 1A Real time clock from
BIOS.
MAR-22-2007 22:00 - added mbr loader for first IDE drive
MAR-05-2007 15:00 - USB4DOS code library added...
OCT-13-2006 20:00 - Added FS_RUN(), Still locks up on files > 512 bytes
OCT-02-2005 21:58 - Updates to graphics.asm
AUG-17-2006 13:16 - DOSMEM.ASM
JUN-29-2006 15:09 - Started on beta 9 design... Complete rewrite.


Any taker to try it and possibly write a few lines on how to build/assemble or whatever is needed to test it?

(the source archive seem to have a number of apps missing from the exscutable archive)

A bootable floppy image would be perfect!


jaclaz

#2 Nuno Brito

Nuno Brito

    Platinum Member

  • Team Reboot
  • 10441 posts
  • Location:boot.wim
  • Interests:I'm just a quiet simple person with a very quiet simple life living one day at a time..
  •  
    European Union

Posted 24 May 2008 - 10:02 PM

Do you have any idea where this would surpass the FreeDos effort? ;)

I also visited the author site and he kept a sort of blog for some time but no significant information about nxdos: http://www.aotksc.com/m-index.html

:thumbup:

#3 was_jaclaz

was_jaclaz

    Finder

  • Advanced user
  • 7100 posts
  • Location:Gone in the mist
  •  
    Italy

Posted 25 May 2008 - 10:29 AM

No, I don't even know if it actually boots/works. ;)

But being aimed to embedded systems it seems more "compact" than Freedos.

It could be useful to be used as a small dos for some peculiar things, like playing with hmload.com of grub4dos, or even integrate a tiny-tiny OS in it.

The COMMAND.COM is very compact, around 15Kb, and the NXBIO.SYS is around 62Kb, an elder version is 43Kb.

As a reference:
Win98SE (Dos 7.1):
COMMAND.COM 92 Kb
IO.SYS 218 Kb

FreeDOS:
COMMAND.COM 65 Kb
KERNEL.SYS 44Kb

Datalight ROM-DOS (NOT redistributable):
ROM-DOS.SYS 75Kb
COMMAND.COM 45 Kb
(I seem to vaguely remember a minicmd.com with reduced functions weighting about 4 or 5 kb)


RX-DOS is smaller, but not really working:
http://rxdos.sourceforge.net/
and not developed anymore.

DR-DOS (of which I still do not understand fully license and redistributability status):
DRBIO.SYS 33 Kb
DRDOS.SYS 35 Kb
COMMAND.COM 61 Kb

jaclaz

#4 wendy

wendy

    Frequent Member

  • Lady
  • 266 posts
  • Location:one mile from the QR main line
  • Interests:Operating systems, Weights and Measures, Geometry
  •  
    Australia

Posted 27 May 2008 - 07:33 AM

Of course:
MS-DOS: 5.02: io.sys 33.430 msdos.sys 37.394 command.com 47.845 sys.com 13.440
LZ-DOS: 7.10: io.sys 75.885 command.com 53.834 sys.com 9.329
PC-DOS: 7.10: ibmbio 43.709 ibmdos 42.550 command.com 53.525
DL-DOS: 6.00: rom-dos.sys 46.518 command.com 26.321

Where sys.com is included, this is the complete redistribution.

msdos 5.02 see eg http://support.microsoft.com/kb/105747

#5 MedEvil

MedEvil

    Platinum Member

  • .script developer
  • 7771 posts

Posted 27 May 2008 - 08:26 AM

LZ-DOS: 7.10: io.sys 75.885 command.com 53.834 sys.com 9.329
DL-DOS: 6.00: rom-dos.sys 46.518 command.com 26.321

Wendy, never heared of those. Who has build them?

;)

#6 was_jaclaz

was_jaclaz

    Finder

  • Advanced user
  • 7100 posts
  • Location:Gone in the mist
  •  
    Italy

Posted 27 May 2008 - 04:25 PM

@Medevil
LZDOS:
http://dos.nm.ru/

DL-DOS=Datalight ROM-DOS:
http://www.datalight...roducts/romdos/

As far as I know there is a "6.30" and a "7.10" version of ROM-DOS, if I recall correctly the latter has native FAT32 support.


@Wendy

http://support.micro...kb/105747/en-us
is titled:

LAN Manager Workstation Hangs with Sound Blaster Pro Card

though it does mention DOS 5.02 I am failing to see it's relevance. ;)

jaclaz

#7 MedEvil

MedEvil

    Platinum Member

  • .script developer
  • 7771 posts

Posted 27 May 2008 - 04:51 PM

DL-DOS=Datalight ROM-DOS:

If you explain it this way, then i know it too. ;)

Thanks! :thumbup:

#8 was_jaclaz

was_jaclaz

    Finder

  • Advanced user
  • 7100 posts
  • Location:Gone in the mist
  •  
    Italy

Posted 27 May 2008 - 06:58 PM

Well, I am far too young ;) to remember a ROM-DOS 6.00
:thumbup:
I got the above by actually reading what wendy posted, bolded below for your convenience:

.....
DL-DOS: 6.00: rom-dos.sys 46.518 command.com 26.321


:thumbup:

jaclaz

#9 MedEvil

MedEvil

    Platinum Member

  • .script developer
  • 7771 posts

Posted 27 May 2008 - 08:50 PM

I got the above by actually reading what wendy posted, bolded below for your convenience:

DL-DOS: 6.00: rom-dos.sys 46.518 command.com 26.321

;) So, the DOS containing io.sys would be IO-DOS? :thumbup:

:thumbup:

#10 was_jaclaz

was_jaclaz

    Finder

  • Advanced user
  • 7100 posts
  • Location:Gone in the mist
  •  
    Italy

Posted 28 May 2008 - 10:15 AM

;) So, the DOS containing io.sys would be IO-DOS? :thumbup:

:thumbup:


Not really, most probably it would be a version of dos developed on one of Jupiter's moons:
http://en.wikipedia.org/wiki/Io_(moon)

....ask a silly question... :thumbup:

jaclaz

#11 bobsobol

bobsobol

    Member

  • Members
  • 31 posts
  •  
    United Kingdom

Posted 28 May 2008 - 06:37 PM

* Note: This reply was supposed to go just after Wendys first post, but there was server maintainence so I had to leave it in Notepad for a bit. ;)

Awww. Never show DOS OS related source code to the old fogie. :thumbup:

Okay... firstly... RX DOS looked so promising for a while, shame it never happened... well, I think there was a commercial off-shoot of it that was reasonably successfully for a while. I wouldn't be at all surprised if it's used in document scanners (mark sheets) and such.

Second of all, RTFM!!! some of what you ask is included in the source documentation for the OS.

You should type the following instructions to produce the
binaries, (assemble and link) and to install on a 1.44MB floppy disk:

make boot nxdos sys command label ...
sys A:

The .C and .CPP sources can be compiled usign Borland C and
or Turbo C 2.01.

Taken from "nxdos-0.9.6.1-src-i386-20080517.zip\DOCS\NXDOS.TXT"

I'm slightly confused by some of the miss conceptions this developer has. He clearly has a better understanding of how to write an OS than I do (it's something I've tried several times, and the closest I get is a command line boot sector) and yet he makes statements about legacy hardware (often using outdated branding hype) which is actually temporally mixed terminology.

fe.

The project began in 1992 when I received my first IBM-PC
an 286XT Monochrome w/ dual 360K floppy drives where I disassembled
MSDOS 2.1 and 3.1 using nothing but debug program.

Well *my* first "IBM Compatible" came with DOS 3.1 on a 360K Floppy, and I did the same. Funny thing is that MicroSofts DOS handbook used to encourage you to disassemble the OS giving examples of how to do so. The only thing I learned from that was that MS Debug was a heck of a lot harder to operate than a Spectrum Multiface 1, and crashed about as often as Unix VI before Linux upgraded it to VIm (VI Improved).

A 286, is however an AT machine. XT specifically meant 8086/8080, only the 80286 processor had the "Advanced Technology" to provide 24-bit addressing via the LIM (Lotus Intel Microsoft) EMS system, and provide additional memory protection required for Microsofts new OS/2. (Basically Windows 1, before Microsoft and IBM went their separate ways and the v2 NT version became IBM OS/2 v2 and the DOS version Microsoft Windows v2)

He also writes that it should run on an "80x286 C/ISC" where the / is supposed to be an italic I in C/SC for CISC (Complex Instruction Set Computer) as opposed to the new RISC (Reduced Instruction Set Computers) which were proving far more powerful for less power input and lower heat expenditure. Hence why many Universities used Acorn Archimedes systems Emulating an IBM compatible to get results from their scientific computations quicker and cheaper than if they ran the same software on a true PC. The C/SC logo was supposed to promote the Complex Instruction sets as being the real genuine thing, not some jumped up upstart who ran twice as fast on half as much fuel but goddamnit it's a FAKE!!! Just like the "Intel Inside" logos which proved you had a slow expensive buggy Intel processor not some cheep fast and stable Cirix or AMD chip. :thumbup:

ARM (Acorn RISC Machines) aficionados had as much fun with the C/SC logo as Motorolla (especially Amiga) fans had with the iNTEL inside logs.

Additionally (of course) there is no such thing as an 80x286. The 80286 processor is an early variant of the 80x86 range, where the x represents the model number.

with minimum of 640KB of RAM and atleast 4MB of secondary storage space.


hmm. In OS kernel terms the 640K Base RAM in DOS is defiantly secondary storage space. CPU Registers are considered the Primary storage, anything on the data or address lines is Secondary and Floppies, CDs, Hard Disks and any other "Long Term" storage is a Tertiary storage system. So I don't know where he expects the 4Meg to be, or if he expects the 4Meg to be on top of the 640K or the 640K to come out of the 4Meg. I'm also not sure where he intends to find 286 based Embedding processors when even the COAC (Computer on a Chip) type processors made by Zilog et el are 386 based.

To answer Nunos question, again the answer is in the docs:-

Generally, an RTOS provides a preemptive multitasking kernel and middleware utilities such as TCP/IP, FOSSILS, and FILE SYSTEMS etc... this OS should be geared for the embedded computing needs... small, robust, realtime clock, and a tcp-ip/fossil stack that run on any I/O port IRQ interface. OS need to be reprogrammable on the fly via scripts such as PHP interpreter support in command with hardly any hardcoding


Okay so we're looking at something that aspires to be very like RX DOS. Should be Multitasking natively, and he mentions that he has done the leg work of making DOS API calls re-entrant (the biggest stumbling block of any other attempt to do this so far) already. Should be capable of network communication natively in the kernel which is very impressive, even Windows and Linux don't do that, except via Kernel extensions... but then he is only talking about 1 network driver built into the kernel, the NE2K driver. (If you don't know, NE2K is to Ethernet what VBE is to Graphics)

So what he's mapped out so far, and what he claims to have achieved is incredibly impressive. I can only suggest that this is an incredibly talented, entirely self trained individual who is better at doing what he does than communicating his intentions.

Unfortunately I fear that as a GPL group project, his lack of educated industry standardized terminology will probably kill it off. In my experience, this sort of thing leads to so many arguments about what is being done that work is actually slowed the more people get involved. But I'd like to think I'm wrong.

Oh and BTW @Wendy.
Don't believe everything that Microsoft says is true for all DOS variants. I'm pretty sure that even some of the older MS versions of DOS the SYS command only copied the Boot sector from the present system disk to a new one. This sometimes led to Boot Sector virus being moved from one floppy to another via SYS.COM. Around DOS 3.x there was an upgrade to SYS.COM which allowed it to modify the boot sectors parameters from a 160K floppy to a 320K floppy to a 720K floppy etc as different size floppies were becoming available, but it didn't do much more than alter the Sides, Tracks, Sectors Per Track information according to the size disk inserted in the drive from the boot sector you had on your boot disk.

I'm sure that if I used SYS.COM from MS DOS 3 in DR DOS 6 to a new disk, the new disk booted DR DOS 6 not MS DOS 3. :thumbup:

And the legality of distributing DR DOS is that if it is branded DR DOS then you can't re-distribute it (any more than you can re-distribute MS DOS, but nobody's gonna be beating down your door) if it's labeled Novel DOS the same applies. If it's Caldera Open DOS then that's fine, do what you like with it.

#12 was_jaclaz

was_jaclaz

    Finder

  • Advanced user
  • 7100 posts
  • Location:Gone in the mist
  •  
    Italy

Posted 29 May 2008 - 07:36 AM

Second of all, RTFM!!! some of what you ask is included in the source documentation for the OS.Taken from "nxdos-0.9.6.1-src-i386-20080517.zip\DOCS\NXDOS.TXT"


Well, since you probably know how to do it, why don't you actually try and help? ;)

In the binaries distribution there is NO readme, no SYS.COM, there is however a bunch of MBR's and bootsectors, some SYSLINUX files, that makes me think that the "new" version uses Syslinux to boot.

Reading inside the MAKE.BAT it seems like a SETUP.EXE file is compiled and run to actually create a bootable disk, SETUP.EXE that is missing in the binaries archive.

Why don't you build a bootable floppy out of it and make it available?

jaclaz

#13 wendy

wendy

    Frequent Member

  • Lady
  • 266 posts
  • Location:one mile from the QR main line
  • Interests:Operating systems, Weights and Measures, Geometry
  •  
    Australia

Posted 01 June 2008 - 10:31 AM

In relation to bobsobol's message.

I have been using computers at home since 1663 (a tandy model 100 laptop!), and generally computer proficient since the acquisition of proggies like 4dos and os/2 2.1 in 1673. With little to turn to, one glarks the code. On the first computer i used, i wrote the necessary code directly to core in raw binary code. Dates in twelfty: today is 1688.

One notes that neither OS/2 or Windows NT have a SYS command. Here, you can prepare a bootable diskette for these operating systems, (eg by SYSINSTX), but it is not by means of the sys command. Nor is it easily accessable.

With DOS, however, DOS is not just an OS, but a user interface, with things like block-devices with single letter names (c: d: ...), and other notions like command.com etc. A programmer who is trying to ape dos does well to emulate the standard functions that are found in the main DOS versions (DR/PC/MS dos). All versions of DOS i have played with feature a sys command.

You can of course manufacture a boot diskette in other ways, eg (where the nxdos files are in a subdirectory nxdos.

bfi -f=nxdos.dsk -b=nxdos\boot.bin -o=nxbio.sys -o=nxdos.exe nxdos\

bfi = http:\www.nu2.nu\

This does indeed produce a bootable nxdos.dsk image. Like the distribution, it stops with a string of characters in the VPC window.

One can manufacture DOS boot diskettes from the likes of rexx-scripts. It's just a matter of reading the boot sector to decide how long to make the fats and the root directory. It's not hard, You just set up an array of boot sectors, and make the command accordingly.

On the other hand, the rexx script makes a raw diskette, berift of files.

While one does not read everything MS writes, there are some rather interesting slip-in distributions, such as MS-DOS 5.02 and MS-DOS 6.30, LZ-DOS 7.10, which are variously slip-ins or string hacks of MS-DOS 5.00, 6.22 and 7.10. The particular distributions for 5.02 and 7.10 are slip-ins with just an additional file sys.com.

The 6.30 distro is in fact a rather poor string-hack over 6.22.

Of 5.02 and 7.10, the matter is different. The kernels do not agree with microsoft's files (even if they did, what the matter: one needs to boot 5.02 to install it, so even if the msdos.sys were identical to 5.00, it would not be different to pc.dos 6.00/6.10/6.30, for example. So we need to look closer to what makes these DOS versions tick.

#14 was_jaclaz

was_jaclaz

    Finder

  • Advanced user
  • 7100 posts
  • Location:Gone in the mist
  •  
    Italy

Posted 01 June 2008 - 11:05 AM

Dates in twelfty: today is 1688.


Interesting approach, why not in base 60? ;) (divisions would be easier):
http://en.wikipedia....ative_notations

Or are you from Nigeria? :thumbup:


jaclaz

#15 wendy

wendy

    Frequent Member

  • Lady
  • 266 posts
  • Location:one mile from the QR main line
  • Interests:Operating systems, Weights and Measures, Geometry
  •  
    Australia

Posted 02 June 2008 - 05:37 AM

The six-score hundred is known from ancient times in England, and much of the germanic peoples.

One needs to look up something like 'long hundred', and 'long thousand', in any decent dictionary. Such words, along with teenty and elefty, are known from old english.

I am not awares of the twelve-based system in nigeria.

When i sat down to decide on what is the most useful system to represent the sorts of fractions i was dealing with, taking into cognance all of the experience of previous number systems (such as 8, 10, 112, 14, 16, 18, 20, 22, 26, 28, 30, 40, 42, 70, 80, 90, 108, &el)., and all the various alternatives to repeating bases, in all sorts of circumstances, and with the full effect of culture and so, the final selection was to use an alternating system, dividing into 12, and multiplying by 10's.

When the existance of an extensive toolset for the system of sixty was taken to account, this failed to come close to the advantages to be found in the first multiple of six, surrounded by composite numbers. In practice, divisions are considerably easier in twelfty than they are in sixty.

Accordingly, it was RECCOMENDED that certian numbers ought be expressed exclusively in this number system, such as the orders symmetry groups, and figures thus derived, and an ever growing number of other things.

Wendy

#16 was_jaclaz

was_jaclaz

    Finder

  • Advanced user
  • 7100 posts
  • Location:Gone in the mist
  •  
    Italy

Posted 02 June 2008 - 10:10 AM

Hmmm, ;) I only use base 13 for my dates and time:
http://en.wikipedia.org/wiki/Base_13
and I'm serious about it:

Nobody writes jokes in base 13 [...] I may be a pretty sad person, but I don't make jokes in base 13.


I am afraid it will be rather difficult to synchronize our watches :thumbup: , and, as it became evident in this thread, to be able to communicate.


jaclaz

#17 bobsobol

bobsobol

    Member

  • Members
  • 31 posts
  •  
    United Kingdom

Posted 02 June 2008 - 08:02 PM

Sounds like Wendy has an easier way of making a boot floppy. I doubt the string of characters is an error in the floppy build, but rather a symptom of timing issues in NX DOS. Such problems have been reported to occur in Bochs with various DOS systems and various Timing setups on different machines. It's not possible to remove the issue from Bochs, but you can correct it for different DOS version on a specific machine by changing the CPU timing... That sounds to me like what Wendy is describing.

If I get time to go through my archives (or hunt the Web) I may see if I can find a TCPP / TASM install set and try to build directly from the source... but I think this system is probably (at the moment) very specific to one particular PC... The developers PC.

I haven't used TCPP or TASM since my college days, and even then, preferred to work on 680x0 assembler and Basic than C++ and 80x86. I am supposed to be trained to do it, and occasionally include in-line assembler in my modern Windows programs (plug-in dlls mostly for speed) but I don't really have the interest to become a member of a "development team" on this project.

I certainly couldn't make the changes necessary to make it work on a more readily available / modern compiler. VCPP or gcc etc. Some of the old Turbo C compilers are available for free from Borland now tho. If anyone has more time / interest, I would recommend trying to build using this in DOSBox. I find it useful for more than just running old games, and since it uses your native directory structure (in NTFS or WHY) old DOS programs can modify files on your main PC with less effort than installing DOS on a more generalized VM.

This is only a suggestion. YMMV. ;)

#18 tekno1911

tekno1911
  • Deactivated
  • 2 posts
  •  
    United States

Posted 24 June 2008 - 12:27 AM

There were 286XT clone machines made, the board was based on the IBM 5162 (PC/XT 286) is simply an XT upgraded to accept a 286 processor running 6-8MHZ (12 MHZ in turbo mode) 8 bit XT style bus slots, dip switches for rom settings, XT keyboard. Hand me down hardware from USC.

I did successfully port the RXDOS dos memory manager over.

Ignore the docs for NXDOS, way out of date and should be deleted. The project is in need of a working filesystem, the one that is there is not working right.

--chris
http://bbx.flnet.org/

A 286, is however an AT machine. XT specifically meant 8086/8080, only the 80286 processor had the "Advanced Technology" to provide 24-bit addressing via the LIM (Lotus Intel Microsoft) EMS system, and provide additional memory protection required for Microsofts new OS/2. (Basically Windows 1, before Microsoft and IBM went their separate ways and the v2 NT version became IBM OS/2 v2 and the DOS version Microsoft Windows v2)



#19 bobsobol

bobsobol

    Member

  • Members
  • 31 posts
  •  
    United Kingdom

Posted 25 June 2008 - 09:35 PM

There were 286XT clone machines made
http://bbx.flnet.org/


Thanks Chris, I did not know that, and in this country at least, the expense of a 286 processor and the simplicity of fabricating a motherboard with AT and XT card slots with off the shelf memory and control chips it would not be practical to make a 286 based XT only machine. (Yes, some hardware struggled with AT clock cycles, but most was okay. IMS some boards had clock halving dip switch for the XT slots. ISA/EISA. They didn't help much. lol)

It's hard to confirm your information, but you seem to speak as though you have historical knowledge which is little documented, so I'm happy to take your word on that. It definitely makes some sense of the authors writings. And it would explain why (s)he seems so knowledgeable and yet speaks with mixed terminology.

I wish the original author would write here. I'd love to hear the argument for this project as anything more than historical curiosity.

IMHO, a DOS-like OS (with DOS design principals) that was 32/64 native, with access to the full memory buses and such would be interesting. But to make that compatible with DOS programs would require a virtualisation sub-system similar to VDM / WOW / DOSEMU.

The nice things about DOS (those principals I spoke of) where that essentially every program was granted full kernel control of the system. Each application was granted 100% of the CPU with (not including NMI routines) nothing running in the background.

The use of Interrupts for API calls is fantastically simple, allows for simple API patching and upgrade of routines. Parameter passing via CPU registers is simple, fast and effective. Actually Digital Research GEM followed this tradition on in a single tasking GUI environment.

All of this makes it completely impractical for a day-to-day OS, but very useful as an emergency maintenance system. Modern equivalence in simplicity and control include SYSLINUX (and derivatives) and Visopsys.

Visopsys is quite capable in it's present form, but is also unfinished, and so writing applications for it may not retain compatibility with future versions. The APIs and build process was not hugely well documented to my mind, but the documentation and source are available. There is no non-GUI environment in Visopsys, so fast simple fsck type operations without the faff of a GUI is not possible.

SYSLINUX is quite mature now, and it's APIs are well documented and system control is total with a true 32-bit environment. There are significant differences between versions designed for booting from Hard Disk, CD, Network or other devices/bootloaders however, and writing an application (COMBOOT / COM32) which will work flawlessly in any of the above environments is somewhat tortuous at present. I gather this is an area under current development. It is essentially console based, but has some simplistic graphical capabilities, mostly not exposed to the API.

Neither is compatible with DOS applications by default, but I imagine a DOSEMU/DOSbox/BOCHS build could be made for either. SYSLINUX can execute DOS .com files (original < 32K binaries with no executable header, CPM style) natively in some of the boot formats. Note that some ".com" files are actually misnamed ".exe" files, usually for reasons of backward compatibility, like "command.com" which sometimes is a COM file and is sometimes an EXE, not that I can remember which versions of DOS used which. It's not hard to check... just look for the "MZ" signature in any good Hex editor. Also of note, if you can find a copy of EXE2BIN, it will attempt to strip the header and relocation information from any EXE which will fit (code and data segments) in < 32k, so that small ".exe" files could be used as ".com" files in SYSLINUX.

IMS PKUNZIP.EXE could do this but PKZIP.EXE could not. DOS Compilers usually build directly to EXE, but not all programs (especially simple commands that take a parameter and return producing a result) require the complex executive loading methods of an ".exe" file. Any ".com" can be built as an ".exe" not all ".exe" files will work in ".com" format. It's rather like building CWSDPMI into an application that never uses more than 50K of combined code and data, except that most compilers didn't build in DPMI in any form unless you told them to, where most produced an EXE even if you knew that a COM was all that was needed. ".com" files are smaller, (on disk) because of the striped header and relocation data. ".exe" files play nicer with any fancy memory management the OS may do. Stripping ".exe" files to ".com" files used to be a nice way to save some space without introducing a lot of compression on a 360/720k boot floppy. :)

I'd like to know what NX-DOS could do that these systems can't. Aside from the fact that it may be possible to build NX-DOS command line and memory management as a COM32 for SYSLINUX, but you could also do that with FreeDOS or FreeDOS32. What could you do in an NX-DOS environment that you couldn't do elsewhere? Or what could you do better in an NX-DOS environment than anywhere else? Right now, I just can't think of anything.

#20 MedEvil

MedEvil

    Platinum Member

  • .script developer
  • 7771 posts

Posted 25 June 2008 - 10:41 PM

All of this makes it completely impractical for a day-to-day OS

:) Those kids today crack me up!

:)

#21 tekno1911

tekno1911
  • Deactivated
  • 2 posts
  •  
    United States

Posted 09 July 2008 - 04:24 AM

>Thanks Chris, I did not know that, and in this country at least, the expense of a 286 processor and the simplicity of >fabricating a motherboard with AT and XT card slots with off the shelf memory and control chips it would not be practical >to make a 286 based XT only machine. (Yes, some hardware struggled with AT clock cycles, but most was okay. IMS >some boards had clock halving dip switch for the XT slots. ISA/EISA. They didn't help much. lol)

If I recall I think i replaced the clock chip got 15mhz out of it :) I threw the board out after a botched attempt to repair the keyboard connection with solder.

>It's hard to confirm your information, but you seem to speak as though you have historical knowledge which is little >documented, so I'm happy to take your word on that. It definitely makes some sense of the authors writings. And it would >explain why (s)he seems so knowledgeable and yet speaks with mixed terminology.

Self taught thats all..

>I'd like to know what NX-DOS could do that these systems can't. Aside from the fact that it may be possible to build >NX-DOS command line and memory management as a COM32 for SYSLINUX, but you could also do that with FreeDOS or >FreeDOS32. What could you do in an NX-DOS environment that you couldn't do elsewhere? Or what could you do better in >an NX-DOS environment than anywhere else? Right now, I just can't think of anything.

Right now it would classify it as an experiment in making a DOS. I still haven't figured out a working FAT driver for it.



#22 scrumpyjack

scrumpyjack

    Newbie

  • Members
  • 20 posts
  •  
    United Kingdom

Posted 07 September 2008 - 01:02 PM

I tried to get nx-dos going but been unsuccessful virtual or non virtual.

There is talk of MSDOS 5.02 using small command.com and io.sys, true but DOS 6.22 and previous use a MSDOS.SYS of size 37K which makes it bigger

#23 was_jaclaz

was_jaclaz

    Finder

  • Advanced user
  • 7100 posts
  • Location:Gone in the mist
  •  
    Italy

Posted 06 November 2009 - 12:37 PM

There is a newish release of NX-DOS:
http://sourceforge.n...ts/nxdos/files/

September 20, 2009

jaclaz

#24 bobsobol

bobsobol

    Member

  • Members
  • 31 posts
  •  
    United Kingdom

Posted 06 November 2009 - 07:36 PM

Thanks jaclaz.

I'm impressed with the fact that the docs in this edition actually address the issue of "what is it good for" with a level of mission statement.

For those too lazy to download just to read the docs, that statement basically sets out that he intends it to be a modular Multi-tasking RTOS with DOS compatibility and a new driver system.

There is a limited TCP-IP stack implemented against the Serial Port of your 80286, but with the driver support you can map devices as your needs require.

The target platform and application is remote or embedded systems. :dubbio:

#25 Wonko the Sane

Wonko the Sane

    The Finder

  • Advanced user
  • 13576 posts
  • Location:The Outside of the Asylum (gate is closed)
  •  
    Italy

Posted 02 July 2013 - 06:49 PM

New version of Nx-dos:

http://tawhakisoft.com/nxdos.html

http://sourceforge.n...ts/nxdos/files/

26-06-2013

 

:cheers:

Wonko






0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users