Aucbvax.6324 fa.works utcsrgv!utzoo!decvax!ucbvax!works Sat Feb 27 15:41:59 1982 IBM PC Review >From COMSAT.SoftArts@MIT-Multics Sat Feb 27 15:34:16 1982 Original to: WORKS at MIT-MC Via: Mit-Mc; 27 Feb 82 17:27-EDT Via: Brl-Bmd; 27 Feb 82 17:32-EDT I have spent quite a bit of time using the IBM PC, and use it full time as a terminal on my host machine (with the PC running a vt100 (subset) emulator at 19200 baud). I have a vt100 to the right of my IBM PC; it is typically used only when I happen to need 132 mode. I find the keyboard to be one of the best that I have ever typed on, with the possible exception of one or two PLATO keyboards. The positive mechanical feedback is a dream. Yes, it took me several weeks to get used to the single-width shift keys (and thus the placement of the back-slash key), but now I can type circles around a vt100. I bind the keys on the various keypads to various EMACS functions, and use the right pad extensively (arrows and page keys). The left pad, in my opinion, suffers greatly because it was not designed with enough room to put a "template" around it. The IBM monochrome monitor suffers from a terrible lack of contrast. We have reduced this problem by putting a $200 Polaroid CP-70 Contrast Enhancement Filter over the display. I have not yet been exposed to the color monitor. I am not impressed with the processor speed. The programs that we have converted from z80 to the 8088 do not run much faster, primarily because of the 8-bit data path. The instruction set is obviously better than the 8-bit machines. As Seth noted, however, dealing with 64k segments is a tremendous pain. This is, of course, not IBM's fault, though the 8-bit bus IS. A MAJOR problem with the IBM is the lack of board slots available. This limits the amount of memory to 512k, or typically 256k. You must trade off peripherals for memory. The IBM sorely needs a hard disk. The IBM hardware beats the hell out of the previous generation 8-bit systems. All peripherals interrupt through a very handy 8259a programmable interrupt controller. There are three hardware clocks: one to refresh memory, one used for DOS timer services (and to time out the disk), and the last to provide a tone generator to play music. The machine provides a socket for the 8087 floating point processor. The color graphics card uses the (fairly flexible) 6845 CRT controller. The disk controller is a NEC uPD765. All in all, it is a very high quality piece of hardware, and very nice to work with. The ASYNC interface is satisfactory. The supplied driver in ROM is a polling driver, and is not much good for many applications. NOTE: If you attempt to write your own interrupt-enabled driver, you may get very frustrated. If you ever want an interrupt from the 8250, you must set bit 3 (OUT 2) of the Modem Control Register. This is documented as a "user-designated output", where the user is IBM, and the designation is to inhibit interrupts(!). The only place that you can find this out is by looking at the schematic on page D-48 of the Technical Reference Manual. IBM DOS is satisfactory. Disk error recovery is pretty good, but other error recovery is marginal. For example, "break"ing during program image activation will frequently cause DOS to crash. The function calls have all seemed to work "as documented". The "end but stay resident" function call is very handy for those of us who wish to write device drivers that stay around. It also enables one to write a very nice "histogramming" program that stays resident during user program execution. Last but not least, I must CHEER for IBM with regards to the $36 "Technical Reference Manual". It has EVERYTHING you ever wanted to know about the machine: schematics, chip write-ups, chip and connector pin-outs, configuration info, LISTINGS OF THE ROM, etc. IBM should be complimented for making all of this info available in an inexpensive, well-put-together manual. Ray Ozzie (Ozzie.SoftArts) ----------------------------------------------------------------- gopher://quux.org/ conversion by John Goerzen of http://communication.ucsd.edu/A-News/ This Usenet Oldnews Archive article may be copied and distributed freely, provided: 1. There is no money collected for the text(s) of the articles. 2. The following notice remains appended to each copy: The Usenet Oldnews Archive: Compilation Copyright (C) 1981, 1996 Bruce Jones, Henry Spencer, David Wiseman.