INMC 80 News |
October–December 1981 · Issue 5 |
Page 39 of 71 |
---|
attempt is made to keep track of where data was stored. It being up to the user to keep a record of that. D-DOS is rather like using a cassette tape recorder and keeping notes as to where data went by noting the numbers on the tape counter. Make a mistake, and it will lose something. Its advantage is that it is easy to understand, easy to modify for your ow purposes if you are thinking of writing your own DOS, and, of course, being very simple, it’s incredibly fast. The next level of cleverness comes with the new NAS-DOS, DCS-DOS, and the new Polydos. These work like D-DOS internally, but instead of you keeping notes on paper, the notes are recorded on the disk in a reserved area known as the directory. The only input to the routines required are whether a read or write is required, the name you chose for the program, and in the case of some writes, the length of the program. The rest is automatic. The DOS looks at the directory for the name required, and the information about it is recorded after it.
CP/M and the Floppy. Much has been written about CP/M in this newsletter of late, but no-one has explained what it is. CP/M stands for Control Program for Minicomputers. When that name was coined, minicomputers were pretty dumb animals by todays standards, and micros are now more powerful that minis then, so for minicomputers, read microcomputers. It was originally written to take the chore of handling DOS away, and make the disk system as transparent as possible. This end is achieved quite well, although CP/M isn’t the friendliest piece of software around. Anyway, CP/M is an operating system much as NASSYS is, but revolves around disks, particularly floppy ones. CP/M is divided into 3 parts CBIOS CCP and FDOS The cbios does all the I/O and is written for each system, as disk I/O and keyboards differ. Now NAS-SYS and CBIOS are much the same, and indeed could be the same if the entry points were looked at closely.
We all know how NAS-SYS calls routines, so how does CP/M do it?
CP/M has a number of routines, 36 for revision 2.0, which are called by passing a routine number in C and an address in DE. Results are returned in A with double bytes in HL.
As CP/M only uses RST 0H, and for the same as NAS-SYS, it should be possible to simply move NAS-SYS up to F000H to F800H and the video to FC00H to FFFFH and incorporate the CBIOS, plus initialise the RST 1 to 7H addresses to jump into NAS-SYS. The only problem then is to modify the VRAM, or rather calls to it. ZEAP is the biggest offender here, mainly because I don’t have the source.
The NAS-SYS routines require about 2K, and the disk routines require 1K, NAS-SYS workspace is 128 bytes, so its a bit of a push, but no doubt it can be done. The work space could be put somewhere else. That way it would provide the best of both worlds.
CP/M reads the disk by using a directory, that relates the logical name (thats what you call the file) and its physical position on the disk. CP/M does not use sector and track as in DDOS but uses block numbers, each block may be, say, 8 sectors, and may therefore be 1024 bytes. The directory entry has 16 bytes set aside to keep the block numbers for that file. If the block size is 1K, then each file can have up to 16k bytes per directory entry, this is know as an extent. Files longer than 16K require two or more consecutive directory entries, the first telling CP/M that it extends to the next extent, etc.
Page 39 of 71 |
---|