Path: utzoo!mnetor!uunet!husc6!mit-eddie!ll-xn!ames!oliveb!amiga!jimm From: jimm@amiga.UUCP (Jim Mackraz) Newsgroups: comp.sys.amiga Subject: Re: Where's CAOS (was: Re: Are you playing with (enough) power?) Message-ID: <1981@amiga.amiga.UUCP> Date: 5 Jan 88 00:24:24 GMT References: <4124@watdragon.waterloo.edu> <4459@pyr.gatech.EDU> <1150@sugar.UUCP> <15998@watmath.waterloo.edu> <19975@amdahl.amdahl.com> <16076@watmath.waterloo.edu> Reply-To: jimm@amiga.UUCP (Jim Mackraz) Organization: Commodore-Amiga Inc., 16795 Lark Ave. Suite 106, Los Gatos CA 95030 Lines: 90 In article <16076@watmath.waterloo.edu> ccplumb@watmath.waterloo.edu (Colin Plumb) writes: )I've heard various people say that Intuition isn't "really" a device. )Can someone describe how it's not? I seems that AmigaDOG opens it to )provide a CLI window, at which point it opens the WB screen and the )input.device and grabs sprite 0 and the copper. Is any of this )forbidden to a device? How could (as I have heard) another six months )of software development time have fixed this? Can it be retrofitted? ) ) )I've been exploring the innards of AmigaDOG and )grown to loathe the BPTR. I've been dreaming about a replacement for )it. Since the most complex thing AmigaDOG seems to do is overlays )(please correct me if I'm terribly misguided - other than that, it )seems to do less than Minix), which isn't *that* hard, I've also been )toying with the idea of starting to consider perhaps beginning to )contemplate the possibility of implementing this hypothetical DOS. ) ) )Can anyone at C-A add some information, even to explain why you can't )say anything? ) -Colin (ccplumb@watmath) Well, I've got a little insight on this stuff. Intuition is not a device, it is a library. You interface to it by function calls, not by sending a request block. So there is no standard way to choose whether you want to wait until it's done or check later if it's done. It also means that it was a hard job (a VERY hard job) in 1.2 to go back and make the parts of it which weren't re-entrant safe (since two tasks can be calling the functions at once). You describe some of things that happen in rapid order, when Intuition library is initialized and when the first workbench window is opened. Very little of this has to do with being a library or device. It's more of a synchronization and serialization issue, as I described. I don't know if another six months would have mattered. I think it took more than that to realize that a device might provide a better model. It is very likely not retrofittable, at least not without breaking a lot of programs. Now that the grunt work has been done and it is reasonably safe, there isn't a lot to be gained making it a device, especially if you ask the old "how many Amigas would it sell?" Don't expect it. As for rewriting DOS, it's a concept, but here are two things to think about: 1) Lots of programs handle BPTRs, so even if you replace all of the BCPL code, we'll be stuck with the screwy data structures. 2) Dos does a fair bit more than handle overlays. Among other things, it is multi-threaded (more or less), handles volumes well, and is no small piece of change, from a technical standpoint. The worst parts (file handler) are being enhanced. In all of these discussions, I sense that people don't appreciate the "compatibility nightmare" that systems programmers must live with if they are to provide backward compatibility. The real trick is to provide support for a future that one cannot even imagine yet, which is a fair step harder than supporting every creative thing you can think of. The modules of the Amiga OS vary in the acheivement of this goal. The ARP people are beginning to appreciate this kind of problem, as the discussions on BIX illustrate. Of course, if you think compatibility is not an issue, and that Minix or anything like it is preferable, port away. I can't see that such a thing makes sense for Commodore to spend time on. Clearly (now), the key ingredient to a successful computer is application software. The world doesn't need a lot of new DOS's, as I'm sure we'll be seeing again with OS/2 (half-OS). And for you CAOS lovers, Andy Finkel is reportedly writing an article on CAOS, including its short length limit on path names and all. The road not taken is often not worth lamenting. So, we all live with what we've got, and the compatible enhancements that there are time for. The ARP project has some promise to relieve some of the nonsense, and Steve Beats' work on the file system will make a big difference. Try again in six months. Maybe I'll come in nights (hah) and make Intuition a device for practice. Cheers jimm -- Jim Mackraz, I and I Computing amiga!jimm BIX:jmackraz Opinions are my own. Comments regarding the Amiga operating system, and all others, are NOT to be taken as Commodore official policy.