Jump to content

Talk:RSX-11

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

Dave Cutler

[edit]

Removed the following from the article:

"Dave Cutler is credit as RSX-11 chief-designer, but there is some controversy on the subject."

I don't know where the that came from, but I was a tech writer on RSX and Cutler was right there beside me :=). The RSX sources were published on microfiche and on one occasion a customer at DECUS asked one of the engineers about the "group pseudonym" used in the code. "What group pseudonym?" "Oh, you know 'Dave Cutler'. Nobody could write that much code." Well, he did.

He was one foul-mouthed fellow as well. He insisted on calling the fork list the "fork queue" and also created the famous error message,

IE.NFW -69 Path lost to partner
I'd call him 'colourful' instead. Thank goodness not everyone is timid or cowardly, right?

He roasted a pig for us when he left the group, wearing a T-shirt with the subtle slogan "fork queue".

I wrote the RSX DCL Manual and was the project leader (engineering, doc, and packaging) for Micro/RSX. Once I gave the canned "history of RSX" speech at DECUS. Ortolan88 15:42 Jul 21, 2002 (PDT)

PS -- I contributed to the RSX-11M/M-PLUS error-message manual that contained the error message and wrote the first RSX-11M/M-PLUS Glossary for the Introduction to RSX-11M-PLUS (my first serious technical writing) and sought fruitlessly (if somewhat tongue in cheek) to include the expression "fork queue" in addition to "fork list" in the glossary because it still appeared here and there in the code, file names, messages and the like, and should therefore be identified. The product managers turned me down, as they had turned Cutler down before (a daunting proposition, to be sure) on the original terminology. Ortolan88 20:52 Aug 4, 2002 (PDT)

RSX-11/B

[edit]

I know it seems strange that /B would be developed / derived from /C, but that is indeed what I recall it saying in the manuals I had in the late 1970s. -- Rob Brown <brown@gmcl.com> —Preceding unsigned comment added by 68.149.194.250 (talk) 16:01, 25 August 2004

There was a predecessor to RSX-11/B in RSX-11A; as far as I can remember RSX-11/C was a diskless version of RSX-11/B, just like RSX-11/S was a diskless version of RSX-11/M.
Then there was of course RSX-15 that ran on the PDP-15 18-bit computers; it came out in the middle of 1971 (I was on the first European training course). It was followed by RSX-15 Plus and later by RSX-15 Plus/III. It had many features later found in RSX-11/M (1974). -- Rien Timmer <Rien.Timmer at wxs.nl> —Preceding unsigned comment added by 199.95.150.2 (talk) 08:49, 10 August 2005

RSX as papa

[edit]

"RSX was a separate path at DEC and the progenitor more than anything of VMS that went to NT via Dave Cutler."
I'm having a great deal of trouble parsing this quote. Does "that" refer to "progenitor" or to "VMS"? It's difficult to make sense of the statement either way. —Preceding unsigned comment added by Tabor (talkcontribs) 21:33, 8 February 2005

That statement seems pretty tortured, but the simple facts are these: Before Dave Cutler went on to participate in the VAX-A committee, he was the initiator and software lead for RSX-11/M and M+. IIRC, he, Peter Conklin, and Dick Hustvedt were the principal contributors to the design of Starlet, the software that went on to become VAX/VMS. Because of Cutler, VMS ended up looking a lot like a clean, 32-bit re-implementation of /M. Many software structures were brought forward and expanded: The ODS-1 disk structure became ODS-2, the QIO and AST mechanisms were brought forward and expanded, etc. (Thankfully, the taskbuilder eventually went away. :-) )
So now we just need a simple statement that sums all that up. ;-)
Atlant 01:08, 9 Feb 2005 (UTC)

RSX GPU

[edit]

RSX is also the name for a Graphics Processor in the Playstaion 3, isn't it? This needs to be added, maybe RSX needs a diambiguation page. —Preceding unsigned comment added by Steve a (talkcontribs) 18:27, 29 January 2006

RSX-15

[edit]

An OS very similar to RSX-11M that ran on the 18-bit PDP-15. I no longer have access to published sources, or I would add it to the article. BenBurch 14:16, 10 November 2006 (UTC)[reply]

Binaries and sources for XVM/RSX V1A and V1B (the last versions of RSX-15) are available at [1]. It will be possible to install them in SIMH 3.7 (the current version at this time is 3.6-2, and XVM/RSX will not run in it due to several bugs) once it is released. There will probably also be a pre-installed disk image of it available as well. 207.153.26.200 08:55, 26 November 2006 (UTC)[reply]

Article should be moved to RSX (operating system)

[edit]

This article should be moved to RSX (operating system) and RSX-11 redirected to it. RSX originated on the PDP-15, and RSX-15 is very similar to RSX-11, especially RSX-11D. 207.153.31.5 13:01, 30 November 2006 (UTC)[reply]

I disagree. Far, FAR more copies of the various flavors of RSX-11 were sold than RSX-15. RSX-11M(+) also has pride of place in being the system that Cutler developed first before he went on to apply most of the same principles to the second-generation VMS and third-generation WNT.
Atlant 15:53, 30 November 2006 (UTC)[reply]
Just because RSX-11 is better known does not mean that RSX-15 shouldn't be mentioned in the list of RSX variants, and RSX-15 is RSX, but not RSX-11. RSX-15 and RSX-11 (especially RSX-11D) are very similar. RSX-11D is in many ways an RSX-15 "clone" (it is different in some ways, like the filesystem, and the MCR, which was heavily extended) for the PDP-11. I know because I am running XVM/RSX and several variants of RSX-11 (M, M+, and IAS) in emulators. If RSX-15 were added, the article wouldn't be just about RSX-11 anymore, and it wouldn't make sense for the article to be titled RSX-11 when it also mentions RSX-15. 207.153.26.235 13:53, 1 December 2006 (UTC)[reply]

Object Based

[edit]

As a student of and trainer in VMS and NT for over fifteen years I have had to steer inept understudies clear time and again from mistakenly grabbing the epithet 'object oriented'. Neither system is object oriented - they're object based. Big difference and end of discussion.

Errors in article

[edit]

First of all, the status is not really "historic" yet. Mentec have continued to develop and sell RSX. This last year the status have become uncertain as the status of Mentec as a company is unknown.

Second, the trivia section talks about the rotating light patter. There is no "reverse rotation" when IND is running. I don't know where that comes from, but it's not true. I have a PDP-11/70 running RSX-11M+ right here next to me to verify this. Also, the source code for the OS is included in the distribution, so it's not exactly difficult to verify the normal idle pattern that way. Since the pattern is the bus data lines, it is very "random" data while the processor is busy, so if IND is running, the pattern will not be something "stable" anyway. /bqt (bqt@softjar.se) --213.65.173.92 (talk) 12:58, 26 July 2008 (UTC)[reply]

Actually, it is true. It just wasn't true by the time your 11/70 booted RSX-11M-PLUS. The "parent/offspring tasking directives" (spwn$ and exst$) were added to RSX-11M in version 3.0. Prior to that, IND could only start an external task and then had to wait for it by periodically examining the active task list to see if it was still running. In order to do this in a way that was non-destructive to system performance, IND would drop itself to be last in the active task list, and then never actually block. It did this by emulating the idle loop by itself, and to differentiate, rotated the lights in the other direction. (By the way, the direction is IN for the kernel idle loop and OUT for INDs idle loop). This whole scheme didn't work well for multiple users, and was replaced with spwn and wait. - Brian McCarthy (mccarthy.bsm@gmail.com)

"First appeared on"

[edit]

It's not correct that a given operating system "first appeared on the PDP-11/xx" if, in fact, it was available for the PDP-11/xx and the PDP-11/yy when it was first released; saying RSX-11D first appeared on the PDP-11/40 makes it sound as if it was initially ran only on the PDP-11/40. If it were announced at the same time that the PDP-11/40 was announced, one could say it was "first announced along with the PDP-11/40", but that's it. I'm fixing it to say "first appeared on the PDP-11/40 and PDP-11/45"; there's nothing at all special bout the 11/40 here. Guy Harris (talk) 18:09, 7 December 2017 (UTC)[reply]

What I (and the previous editors) meant is just that PDP 11/40 and RSX-11 were announced at the same time. 11/40 was announced after the /45 (June 1972), so late 1972 or very early 1973 (data varies slightly: DEC officially said "sometime in 1972 after the 11/45", PDP FAQ says January 1973). "Late 1972 or very early 1973" is the same time frame I got for the launch of RSX-11: in March 1973 there's already a nice printed set of manuals for 11A, so it was generally available at that time and no doubt there were customers who got it before that time. Lastly, RSX 11D came out in May 1973. Since the time frames are the same, I find no reason to doubt what previous editors have said: 11/40 and RSX were announced at the same time. Some editors who have written in this talk page seem to have first hand knowledge about the subject. Pity the page was a bit vandalized by someone misreading PDP-11 newsgroups. English is a second language for me, so please go on and correct the page yourself. I'm confident your writing is better than mine could ever be. Alchemist (talk) 23:02, 7 December 2017 (UTC)[reply]
In English, "first appeared on the PDP-11/40" would be appropriate if the first machine on which it could run on was the 11/40, which is not the case - 11D could run on the 11/40 and the 11/45 when it was introduced; for "first appeared on", what's appropriate is "first appeared on the PDP-11/40 and PDP-11/45", as it currently states. "...was first announced along with the 11/40" would be appropriate if they were both announced at the same time, but the set of machine(s) announced in the same time frame is less interesting and relevant than the set machines it supported when it was announced. Guy Harris (talk) 23:33, 7 December 2017 (UTC)[reply]
[edit]

This section in the article is more wrong than right.

First of all, RSX was/is distributed with sources, and the sources do not have comments stripped out. I believe this is the case for RT-11, but it has never been the case for RSX-11. There was a separate source kit product available as well, which held additional sources not needed to rebuild the system. Mostly tools and libraries.

Second, Mentec Inc. was not acquired by anyone. It went into bankrupcy. It's important to keep the distinction between Mentec Limited, and Mentec Inc. Mentec Inc was certainly owned by Mentec Limited, but it went to a different fate. The ownership of the PDP-11 software was then sold from the remains of Mentec Inc to a company called XX2247 Ltd. which is still the owner of the software today, and in theory you can still get a licence if you want to. However, in practice it might be hard because of paperwork and legal details that needs to be handled if this is needed. XX2247 is not actively doing anything with this software today, except keeping it around for some future day. /bqt@softjar.se 80.219.208.146 (talk) 13:49, 27 March 2020 (UTC)[reply]

Dubious

[edit]

I highly doubt that, as the article claims, "the current owner cannot be firmly established". I can't believe that HP would have split in two without firmly establishing which half owns which intellectual property. It's like saying they didn't decide who gets the office buildings, so they'll have to fight over them. Just because we don't know who got RSX-11 doesn't mean nobody knows.—Chowbok 19:25, 10 June 2020 (UTC)[reply]

If I were to guess, I'd guess that all the non-PC stuff Hewlett-Packard got from Compaq went with Hewlett Packard Enterprise. But a reliable source would be required here, so my guess doesn't count. Guy Harris (talk) 19:54, 10 June 2020 (UTC)[reply]
I think you're right, because I did find a couple old papers about it on support.hpe.com. But again, that's not really enough evidence. Anyway, I'm probably just going to take that sentence out unless somebody can point me to an RS that says its ownership is in doubt.—Chowbok 07:09, 11 June 2020 (UTC)[reply]
It's sadly enough even more complicated than that. DEC sold RSX-11M and its siblings (meaning -11S and -11M-PLUS) to Mentec. However, even though it was sold, DEC retained some kind of control. The exact formulation of that agreement has never been made public. So not only do we not know which part of HP this in the end leads to, we also don't know in which way this relates to the ownership that Mentec might have claimed.
However, you're absolutely right in that *someone* knows. But it's very doubtful HP (any part of it) have a clue. But there might be papers buried somewhere in a vault. And someone probably have the papers from the Mentec side as well. Sillbit (talk) 01:40, 13 August 2020 (UTC)[reply]
Either someone in HPE (or, less likely, HP) actually knows or they could know by actually reading the papers. However, observers outside of HP / HPE cannot tell anything about the legal status of these OSes. It is not publicly known. Alchemist (talk) 14:50, 7 April 2021 (UTC)[reply]

Relation to RSX-20F

[edit]

I noticed a "reverted" edit that claims that RSX-20F was based on RSX-11M. I know there are sources that claims this, but unfortunately it is still incorrect. Having access and have read through the sources of both, it is very clear that this is not the case. However, there is strong similarities to -11D. The question is what to do when there are available sources that claims one thing, but knowledge from being familiar with the code makes me know it is incorrect. Sillbit (talk) 01:40, 13 August 2020 (UTC)[reply]

Actually, if the reference is read, it do start by saying RSX-20F is based on RSX-11M and RSX-11D, but unfortunately, the rest of the text does not expand on this, and mostly just refers to RSX-11M. So the source does not really clarify how much of RSX-20F comes from one or the other. Not sure how to make the change so that it just don't get reverted. I can essentially assure that the major part of RSX-20F comes from -11D, but there are a few bits and pieces which might originate in -11M in there as well. But sources for any such claims would be very hard to come by, and even though the sources can be found, it will require quite some knowledge for anyone to verify the lineage based on this. It's all MACRO-11 sources, and quite a lot of it. Sillbit (talk) 02:01, 13 August 2020 (UTC)[reply]
To quote WP:RS, "Articles should be based on reliable, independent, published sources with a reputation for fact-checking and accuracy." Sometimes accuracy is lost somewhere between the engineering department and the technical publications department, whether because engineering didn't explain something properly or because technical publications didn't understand something properly, so even official documents may not provide sufficient "fact-checking and accuracy".
This may be one of those cases; perhaps we should just deem that document "insufficiently reliable" when it speaks only of RSX-11M, and just say RSX-20F is derived from "RSX-11D and RSX-11M" and leave it at that. Guy Harris (talk) 02:32, 13 August 2020 (UTC)[reply]
And add the quote about 11D and 11M to the reference, with quote=, to emphasize that particular quote. Guy Harris (talk) 02:46, 13 August 2020 (UTC)[reply]