10

86-DOS

 4 years ago
source link: https://en.wikipedia.org/wiki/86-DOS
Go to the source link to view the article. You can view the picture content, updated content and better typesetting reading experience. If the link is broken, please click the button below to view the snapshot at that time.
neoserver,ios ssh client
86-DOS BV3qeer.png!webiq67vaU.png!web

A sample 86-DOS session (simulated).

Developer Seattle Computer Products /Tim Paterson Written in 8086 assembly language OS family DOS Working state Historic, Unsupported Source model Closed source Initial release 1980 ; 40 years ago Latest release 86-DOS 1.10 / July 1981 ; 38 years ago Marketing target S-100-based micro-computers with 8086 processor Available in English Package manager N/A Platforms x86 Kernel type Monolithic kernel Defaultuser interface Command-line interface ( COMMAND.COM ) License Proprietary

86-DOS is a discontinuedoperating system developed and marketed by Seattle Computer Products (SCP) for itsIntel 8086-based computer kit. Initially known as QDOS ( Quick and Dirty Operating System ), the name was changed to 86-DOS once SCP started licensing the operating system in 1980.

86-DOS had a command structure and application programming interface that imitated that ofDigital Research'sCP/M operating system, which made it easy toport programs from the latter. The system was licensed and then purchased byMicrosoft and developed further asMS-DOS andPC DOS.

Contents

History [ edit ]

Origins [ edit ]

86-DOS was created because sales of the Seattle Computer Products 8086computer kit, demonstrated in June 1979 and shipped in November,were languishing due to the absence of an operating system. The only software that SCP could sell with the board was Microsoft's Standalone Disk BASIC-86 , which Microsoft had developed on a prototype of SCP's hardware.SCP wanted to offer the 8086-version ofCP/M that Digital Research had initially announced for November 1979, but it was delayed and its release date was uncertain.This was not the first time Digital Research had lagged behind hardware developments; two years earlier it had been slow to adapt CP/M for newfloppy disk formats andhard disk drives. In April 1980 SCP assigned 24-year-oldTim Paterson to develop a substitute forCP/M-86.

Using a CP/M-80 manual as referencePaterson modeled 86-DOS after its architecture and interfaces, but adapted to meet the requirements of Intel's 808616-bit processor, for easy (and partially automated) source-level translatability of the many existing8-bit CP/M programs;porting them to either DOS or CP/M-86 was about equally difficult,and eased by the fact that Intel had already published a method that could be used to automatically translate software from theIntel 8080 processor, for which CP/M had been designed, to the new 8086 instruction set.At the same time he made a number of changes and enhancements to address what he saw as CP/M's shortcomings. CP/Mcached file system information in memory for speed, but this required a user to force an update to a disk before removing it; if the user forgot, the disk would become corrupt. Paterson took the safer, but slower approach of updating the disk with each operation. CP/M'sPIP command, which copied files, supported several special file names that referred to hardware devices such asprinters andcommunication ports. Paterson built these names into the operating system asdevice files so that any program could use them. He gave his copying program the more intuitive nameCOPY. Rather than implementing CP/M'sfile system, he drew on Microsoft Standalone Disk BASIC-86's File Allocation Table (FAT) file system.

By mid-1980 SCP advertised 86-DOS, priced atUS$95 for owners of its US$1,290 8086 board and US$195 for others. It touted the software's ability to readZilog Z80 source code from a CP/M disk and translate it to 8086 source code, and promised that only "minor hand correction and optimization" was needed to produce 8086 binaries.

IBM interest [ edit ]

In October 1980,IBM was developing what would become the original IBM Personal Computer . CP/M was by far the most popular operating system in use at the time, and IBM felt it needed CP/M in order to compete. IBM's representatives visited Digital Research and discussedlicensing with Digital Research's licensing representative,Dorothy Kildall (née McEwen), who hesitated to sign IBM's non-disclosure agreement . Although the NDA was later accepted, Digital Research would not accept IBM's proposal of US$250,000 in exchange for as many copies as IBM could sell, insisting on the usualroyalty-based plan.In later discussions between IBM andBill Gates, Gates mentioned the existence of 86-DOS and IBM representative Jack Sams told him to get a license for it.

[ citation needed ]

Creation of PC DOS [ edit ]

Microsoft purchased a non-exclusive license for 86-DOS from Seattle Computer Productsin December 1980 for US$25,000.

[ citation needed ]

In May 1981, it hired Tim Paterson to port the system to the IBM PC,which used the slower and less expensiveIntel 8088 processor and had its own specific family of peripherals. IBM watched the developments daily,submitting over 300change requests before it accepted the product and wrote the user manual for it.

In July 1981, a month before the PC's release, Microsoft purchased all rights to 86-DOS from SCP for US$50,000.It met IBM's main criteria: it looked like CP/M,and it was easy to adapt existing 8-bit CP/M programs to run under it, notably thanks to theTRANS command which would translate source files from 8080 to 8086 machine instructions. Microsoft licensed 86-DOS to IBM, and it became PC DOS 1.0. This license also permitted Microsoft to sell DOS to other companies, which it did. The deal was spectacularly successful, and SCP later claimed in court that Microsoft had concealed its relationship with IBM in order to purchase the operating system cheaply. SCP ultimately received a US$1 million settlement payment.

Intellectual property dispute [ edit ]

When Digital Research founderGary Kildall examined PC DOS and found that it duplicated CP/M's programming interface, he wanted to sue IBM, which at the time claimed that PC DOS was its own product. However, Digital Research's attorney did not believe that the relevant law was clear enough to sue. Nonetheless, Kildall confronted IBM and persuaded them to offer CP/M-86 with the PC in exchange for a release of liability.

Controversy has continued to surround the similarity between the two systems. Perhaps the most sensational claim came fromJerry Pournelle, who said that Kildall personally demonstrated to him that DOS contained CP/M code by entering a command in DOS that displayed Kildall's name,but Pournelle never revealed the command and nobody has come forward to corroborate his story. A 2004 book about Kildall says that he used such an encrypted message to demonstrate that other manufacturers had copied CP/M, but does not say that he found the message in DOS;instead Kildall's memoir (a source for the book) pointed to the well-known interface similarity. Paterson insists that the 86-DOS software was his original work, and has denied referring to or otherwise using CP/M code while writing it.After the 2004 book appeared, he sued the authors and publishers fordefamation.The court ruled insummary judgment that no defamation had occurred, as the book's claims were opinions based on research, or were not provably false.

Versions [ edit ]

QDOS 0.10 1980-07 Roughly half completed version of the OS.It implemented the original form of Paterson's derivation of Microsoft'sFAT filesystem. In contrast to the earlier8-bit FAT filesystem found in Standalone Disk BASIC-86 , this variant already supported 12-bit table elements,reduced the number of FATs from 3to 2, redefined the semantics of some of the reserved cluster values, and modified the disk layout, so that the root directory was now located between the FAT and the data area. Paterson also increased the previous 9-character length limit to 11 characters in order to supportCP/M-style8.3 filenames andFile Control Blocks. This variant, however, still used 16-byte directory entries and therefore is not compatible with what became later known asFAT12 in MS-DOS/PC DOS. Possibly shipped already. QDOS 0.11 1980-08 QDOS/86-DOS 0.2 1980-08 EDLIN added. Renamed to 86-DOS by August 1980. 86-DOS 0.3 15 November 1980 First version licensed by SCP to Microsoft. 86-DOS 0.33 1980-12 First version distributed by SCP toOEMs and Microsoft as 86-DOS. 86-DOS 0.34 Cleaned up release for SCP OEMs including Microsoft. 86-DOS 0.42 At the indirect request of IBM (through Microsoft) size of directory entries changed from 16 to 32 bytes (similar toMDOS/MIDAS) in order to add support for a last-modified date stamp (2 bytes) and theoretical file sizes larger than 16 MB (4 bytes),thereby implementing the earliest form of the12-bit FAT file system logically fully compatible with what became known asFAT12 in MS-DOS/PC DOS later on. 86-DOS retained the capability to read volumes written under older versions of 86-DOS up toMS-DOS 1.14 at the least, whereas genericMS-DOS 1.20 /PC DOS 1.1 and higher no longer supported it. 86-DOS 0.56 Support for disk blocking/deblocking as requested since 1981-01-15. 86-DOS 0.60 The file DOSIO.ASM found in 86-DOS 1.00 mentions that it is the "I/O System for 86-DOS version 0.60 and later". 86-DOS 0.74 Addeddevice files. 86-DOS 0.75 Bug fix. 86-DOS 0.76 Bug fix. 86-DOS 0.80 Added system calls. 86-DOS 1.01 Bug fix. 86-DOS 1.10 Purchased by Microsoft and renamed to MS-DOS on 1981-07-27,therefore the last version with genuinely matching 86-DOS and MS-DOS version numbers. First version known to implement the 'hidden' attribute. 86-DOS 1.14 1981-10/11 According to Tim PatersonPC DOS 1.0 basically reflects 86-DOS 1.14,other sources find a match of PC DOS 1.0 with MS-DOS/86-DOS 1.10 as of 1981-07-21 more likely.

Features [ edit ]

Commands [ edit ]

The following list ofcommands is supported by 86-DOS.

Internal commands [ edit ]

External commands [ edit ]

EDLIN [ edit ]

By 1982, when IBM asked Microsoft to release a version of DOS that was compatible with ahard disk drive,PC DOS 2.0 was an almost complete rewrite of DOS, so by March 1983, very little of 86-DOS remained. The most enduring element of 86-DOS was its primitive line editor,EDLIN, which remained the only editor supplied with Microsoft versions of DOS until the June 1991 release ofMS-DOS 5.0, which included a text-based user interface editor calledMS-DOS Editor, based onQBasic. EDLIN can still be used on contemporary machines, since there is an emulated DOS environment up to Windows 10 (32 bit).

Supported disk formats [ edit ]

Seattle Computer Products' 86-DOS supported theFAT12 filesystem on a range of 8-inch and 5.25-inch floppy disk drives onS-100 floppy disk controller hardware manufactured byCromemco,Tarbell Electronics andNorth Star Computers. The Western Digital FD1771 -based Cromemco and Tarbell boards supported one-sided, single-density soft-sectored drives. A Tarbell double-density board utilizing theFD1791 was supported as well. Later, SCP offered advanced floppy disk controllers, like the Disk Master series.

86-DOS did not take advantage of aFAT ID byteorBIOS parameter block (BPB), as later DOS versions do, to distinguish between different media formats; instead different drive letters were hard-coded at time of compilation to be associated with different physical floppy drives, sides and densities. That meant, depending on its type, a disk had to be addressed under a certain drive letter to be recognized correctly. This concept was later emulated with more flexibility byDRIVER.SYS under DOS 3.x and later versions.

Two logical format variants of the 86-DOS 12-bit FAT format existed — the original format with 16-byte directory entries and the later format (since86-DOS 0.42) with 32-byte directory entries. Only the second one is logically compatible with the FAT12 format known since the release of MS-DOS and PC DOS. MS-DOS still cannot mount such volumes, as in absence of a BPB it falls back to retrieve the FAT ID in the FAT entry for Design of the FAT file system#CLUST 0cluster 0 to choose among hard-coded disk geometry profiles. In all formats of a volume formatted under MS-DOS that would otherwise be supported by both systems and typically also in all other formats, this ID is located in the first byte of logical sector 1 — that is, the volume's second sector with physicalcylinder-head-sector (CHS) address 0/0/2 or logical block addressing (LBA) address 1 — since MS-DOS assumes a single reserved sector, the boot sector. Under 86-DOS, the reserved sectors area is significantly larger (whole tracks), and therefore the prototypical FAT ID 0xFE (and 0xFF ) is located elsewhere on disk, making it impossible for MS-DOS to retrieve it, and even if it would, the hard-coded disk profile associated with it would not take this larger reserved sectors region under 86-DOS into account.

CP/M 2 floppy media were readable through.

86-DOS did not offer any specific support forfixed disks, but third-party solutions in form of hard disk controllers and corresponding I/O system extensions for 86-DOS were available from companies like Tallgrass Technologies , making hard disks accessible similar tosuperfloppies within the size limits of the FAT12 file system.

Size 8 inch 8 inch 8 inch 5.25 inch 5.25 inch Density SD DD DD SD SD Modulation FM MFM MFM FM FM Cylinders (CHS) 77 77 77 40 35 Physical sectors / track 26 8 8 18 10 Number of heads 1 1 2 1 1 Byte payload / physical sector 128 1024 1024 128 256 Bytes / logical sector 128 1024 1024 128 256 Logical sectors / cluster 4 1 1 2 1 Reserved logical sectors 52 (2 tracks) 1 1 54 (3 tracks) 30 (3 tracks) Number of FATs 2 2 2 2 2

Root directory entries

(á 32 bytes,

86-DOS 0.42 and higher only)

64 (16 sectors) 96 (3 sectors) 128 (4 sectors) 64 (16 sectors) 64 (8 sectors)

Root directory entries

(á 16 bytes)

64 ? ? 64 N/A Total logical sectors 2002 616 1232 720 350 Logical sectors / FAT 6? ? ? ? ? Hidden sectors 0 0 0 0 0 Logical sector order ? ? ? ? ? Sector mapping sector+/ track+ sector+/ track+ sector+/ head+/ track+ sector+/ track+ sector+/ track+ First physical sector 1 1 1 1 0 Sector index Soft Soft Soft Soft Hard (10+1) Controller type Cromemco/TarbellFD1771 Tarbell FD1791 Tarbell FD1791 Cromemco FD1771 NorthStar BPB presence No No No No No

Various OEM versions of MS-DOS 1.2x and 2.x supported a number of similar 8.0 in FAT12 floppy disk formats as well, although not identical to those supported by 86-DOS.

Disk formats supported by one of the last versions developed by Tim Paterson at Microsoft,MS-DOS 1.25 (March 1982) for the SCP Gazelle computer with SCP controller or Cromemco 16FDC controller (by default, this version only supported the MS-DOS-compatible variants of the 8.0 in formats with a single reserved sector but it could be built to provide two extra drive letters to read and write floppies in the previous SCP 86-DOS 8.0 in disk formats since 0.42 as well):

Size 8 inch 8 inch 8 inch 8 inch 8 inch 5.25inch 5.25 inch 5.25 inch Density SD SD DD DD DD SD DD DD Modulation FM FM MFM MFM MFM FM MFM MFM Formatted capacity (KB) 250.25 (SCP) 250.25 (MS) 616 1232 (SCP) 1232 (MS) 90 160 320 Cylinders (CHS) 77 77 77 77 77 40 40 40 Physical sectors / track 26 26 8 8 8 18 8 8 1 1 1 2 2 1 1 2 Byte payload / physical sector 128 128 1024 1024 1024 128 512 512 Bytes / logical sector 128 128 1024 1024 1024 128 512 512 Logical sectors / cluster 4 4 1 1 1 2 1 2 Reserved logical sectors 52 (2 tracks) 1 1 1 1 54 (3 tracks) 1 1 2 2 2 2 2 2 2 2

Root directory entries

(á 32 bytes)

64 (16 sectors) 68 (17 sectors) 96 (3 sectors) 128 (4 sectors) 192 (6 sectors) 64 (16 sectors) 64 (4 sectors) 112 (7 sectors) Total logical sectors 2002 2002 616 1232 1232 720 320 640 Logical sectors / FAT 6? 6 1 ? 2 4 1 1 0 0 0 0 0 0 0 Total number of clusters ? 497 ? ? 1227 ? 313 315 Logical sector order ? ? ? ? ? ? ? ? Sector mapping sector+/ track+ sector+/ track+ sector+/ track+ sector+/ head+/ track+ sector+/ head+/ track+ sector+/ track+ sector+/ track+ sector+/ head+/ track+ First physical sector 1 1 1 1 1 1 1 1 Sector index Soft Soft Soft Soft Soft Soft Soft Soft Controller type Cromemco / TarbellFD1771 Cromemco / Tarbell FD1771 Tarbell FD1791/FD1793 Tarbell FD1791 / FD1793 Tarbell FD1791 / FD1793 Cromemco 4FDC FD1771 SCP / Cromemco 16FDC SCP / Cromemco 16FDC FEh FEh FEh FEh FEh FEh FEh FFh BPB Presence No No No No No No No No

In 1984 Seattle Computer Products released an OEM version ofMS-DOS 2.0 for the SCP S-100 computer with SCP-500 Disk Master Floppy controller. It added support for 5.25 in DD/1S (180 KB) and DD/2S (360 KB) FAT12 formats and supported the older formats as well, although possibly with some of the parameters modified compared toMS-DOS 1.25.

See also [ edit ]

Notes [ edit ]

  1. Somewhat ironically in the context ofJerry Pournelle's claim thatGary Kildall demonstrated to him a way to display Kildall's name in DOS,Tim Paterson claims to have hidden aneaster egg, displaying his own name, in theFAT code ofMSX-DOS 1, when he developed thisZ80-basedCP/M clone resembling 86-DOS/MS-DOS 1.25 in 1983: Paterson, Tim (2014-02-17). "The History of MSX-DOS" . Jorito, Maggoo, John Hassink, MSX Resource Center . Retrieved 2014-05-31 .
  2. 8.0" 250.25 KB images formatted under 86-DOS 1.00 sport a FAT ID of FEh, however, in contrast to MS-DOS/PC DOS, 86-DOS does not seem to use this to detect the disk format, as this information is hard-coded into disk profiles associated to certain drive letters at compile-time. MS-DOS would not be able to mount such volumes as (in absence of a BPB) it expects the FAT ID in logical sector 1, assuming only one reserved sector of 512 bytes (the boot sector in logical sector 0) instead of the 52 reserved sectors á 128 bytes used by 86-DOS here. This works for MS-DOS, because the system files are not part of the reserved area under MS-DOS, while under 86-DOS there are no system files and the ca. 6 KB large DOS kernel is located in the reserved area.
  3. Executing the CLEAR command under 86-DOS 1.00 COMMAND.COM seems to initialize a volume's FAT ID byte to FEh regardless of disk drive and format used.
  4. DOS 1.x does not support aBPB, but this entry for the number of physical sectors per track corresponds with BPB offset 0x0D under DOS 3.0 and higher.
  5. DOS 1.x does not support aBPB, but this entry for the number of heads corresponds with BPB offset 0x0F under DOS 3.0 and higher.
  6. DOS 1.x does not support aBPB, but this entry for the bytes per logical sector corresponds with BPB offset 0x00 under DOS 2.0 and higher.
  7. DOS 1.x does not support aBPB, but this entry for the logical sectors per cluster (allocation units) corresponds with BPB offset 0x02 under DOS 2.0 and higher.
  8. DOS 1.x does not support aBPB, but this entry for the number of reserved logical sectors corresponds with BPB offset 0x03 under DOS 2.0 and higher.
  9. DOS 1.x does not support aBPB, but this entry for the number of FATs corresponds with BPB offset 0x05 under DOS 2.0 and higher.
  10. DOS 1.x does not support aBPB, but this entry for the number of root directory entries (á 32 bytes) corresponds with BPB offset 0x06 under DOS 2.0 and higher.
  11. DOS 1.x does not support aBPB, but this entry for the total number of logical sectors corresponds with BPB offset 0x08 under DOS 2.0 and higher.
  12. DOS 1.x does not support aBPB, but this entry for the number of logical sectors per FAT corresponds with BPB offset 0x0B under DOS 2.0 and higher.
  13. DOS 1.x does not support aBPB, but this entry for the number of hidden sectors corresponds with BPB offset 0x11 under DOS 3.0 and higher.
  14. For unknown reasons, some Microsoft documents give a value of 3 for this entry, where 0 seems correct technically. SCP MS-DOS 1.25 implicitly assumes 0 as well.
  15. DOS 1.x does not support aBPB, but this entry for theFAT ID corresponds with the media descriptor byte at BPB offset 0x0A under DOS 2.0 and higher.

References [ edit ]


About Joyk


Aggregate valuable and interesting links.
Joyk means Joy of geeK