Dne Sat, 22 Jul 2000 07:04:24 +0200, [EMAIL PROTECTED] (arachne-digest) napsal: >> Linux multitasking is real pre-emptive multitasking, not semimultitaksing > I said semi-simultaneous. There can be no disagreement over the meaning > of simultaneous, while the meaning of multi-tasking is open to agrument. > I run Arachne in DOS. I run Autocad in DOS. That can easily be called > multitasking even if I do them a week apart.<G> I know that that isn't > the common understanding of the word but it works with my dictionary. I think most of the world as already agreed on meaning of the term "task": task is kernel structure, which has unique memory for program data allocated, and uniqued scheduling information set (like process ID), and task is what getting CPU time, if OS is not in kernel mode. Tasks are allso called "processes". Then there are also threads: threads have uniques sheduling information, but share memory with parent thread (process). In Linux, this is implemented in a way, when threads have unique record in process table, having its own process ID, etc., but share memory with parent thread (process). In DOS, we can (maybe ;) talk about multithreading: TSR programs can eventually run as separate threads, but their memory is not protected from other tasks, as well as "foreground process" memory is not protected from them. >> This is not real multitasking, as you have no control over which priority >> which task gets, you cannot easily kill tasks, etc... > Whatever the words REAL multitasking might mean to different observers isn't > important - the truth is that the Task Manager I speak of DOES assign the > priorities in a re-programmable manner, including the ability to suspend. I am not sure if it will be able to protect memory of one process from another process. And I am not sure if it will be pre-emptive - but this is in fact not impossible, you can install sheduler at hardware timer interrupt, and you can reset CPU registers to values corresponding to other process... >> and to be able to communicate withpreviously launched copy of the >> application, instead of trying to access network directly. > My last point was that I didn't even want to try to do THAT. <G> Hmmm, because so many people tried in DOS, and failed - yes, IMHO even Desqview was failure, because it was not possible to run 32bit DDOS apps in it... at least in version which I was using. Once it was using DOS4GW extender - CRASH! > But if the manager switches foreground jobs then the MANAGER should > reload the video buffer. Otherwise it's hardly transparant is it ? Video buffer is not an issue. In fact, in real operating system, input and output should be virtualized. The way Desqview was trying to keep persistance of graphics mode screen was pathetic, compared to advanced technologies like X11... >> I'm building a 486 from scratch just to run Arachne. This box will have no >> HardDrive and use a flash disk for booting with 16MB ram. I was wondering >> what DOS experienced users thought would be best to use. I'm not concerned >> with licensing issues at present. > 5.0 DR-DOS 7.whatever, with QEMM-whatever installed as memory manager. Make sure you remove DR-DOS native EMM386 and DPMI driver ;-) > CP does. > Just press the "d" key while music is playing and then run Arachne. >> Ahh... too bad they never considered browsing while using thier player! > Whether or not they considered it... it works. > I'm doing it right now. :-) > CP v2.51 is playing K_SHALAM.S3M as I type this. Yes, background mod/s3m/xm players were working quite nice even on bacgkround, but they were usually leaving something like 100 KB of free DOS memory left. If your player is different, then it is big progress ;-) BTW, anyone knows about TSR _MP3_ player ? ;) . . . . . >> Could someone tell me how to disassemble an APM into it's components for >> manual installation ? It is just ARJ archive with different extension ;-) So try "arj x NAME.apm" -- Michael Polak: [EMAIL PROTECTED] Arachne Labs: http://arachne.cz/ My mobile phone - up to 160 characters: [EMAIL PROTECTED]
