Talk:ITS 138

From Computer History Wiki
Jump to: navigation, search

Interesting items in the ITS 1.4 manual

I saw a couple of interesting things in the ITS 1.4 manual. The first is that what we now call a 'process' is referred to in it as a "procedure". The manual will be a little bit more comprehensible if this is remembered! :-)

The second is something I caught in Section 5.1, right at the bottom of the page: "A procedure's core image may be easily swapped out without <blah, blah>". This was interesting because you had said that early ITS didn't do swapping? I looked for a while to see if I could find more on this 'swapping' (since they apparently had their own private language, and words might mean something else than they normally mean now :-), but all I could find quickly was in Section 6.1, at the bottom of page 56 (printed page number, not PDF): "while not in actual use they [separate copies of the command processor] could be swapped out of the system." On the other hand, in Section 1, on page 4, we find: "(2) All user programs reside in core storage ... (The coming installation of additional disk storage will enable the swapping of programs". Note that the term "program", not 'procedure' is used - do they mean something different, or had they changed their terminology, and this text in the Introduction wasn't caught, and updated? Similarly for the "will enable .. swapping": maybe this was added, and they just never went back to update this text? I have this vague memory that the very earliest system only had DECtapes, and no disks, so not swapping probably made sense there. I suppose you will have to study the code to see for certain! Jnc (talk) 12:56, 29 November 2023 (CET)

Despite the wording in the manual, I maintain this early version had no swapping. Fred Wright, who seems to have a good memory of such things, has claimed that paging was considered for ITS early on. So my guess would be that the authors had this in mind as a distinct possibility, even though it was not implemented yet. Hence the weasel words "may" and "could". We'll know for sure when we have the full listing.
The story goes, the first two disk systems for ITS weren't really successful, and hence paging or swapping wasn't possible. In 1969, an interface to IBM 2311 disks were added, and a pager was implemented in 1970.
Unfortunately ITS 138 is very early, so there's a large gap in the evidence. All we will be able to say if the full listing confirms my theory, is that in 1967 there was no disks and no swapping, and in 1970 there was paging to disk.
Larsbrinkhoff (talk) 14:21, 29 November 2023 (CET)
Project MAC Progress Report IV from July 1967 says "The [moby] memory is utilized to avoid swapping user programs". The 1.4 manual was published after the Data Disc drives were added a year later, so possibly the authors were considering swapping to those. Since the capacity of the disks were not much larger than the moby, that didn't pan out. Larsbrinkhoff (talk) 17:20, 29 November 2023 (CET)
Or maybe I should say, the ITS programmers decided to use the small disks solely for user file storage instead. Larsbrinkhoff (talk) 17:23, 29 November 2023 (CET)
Yes, I did not pay enough attention to the 'weasel words' in the text passages I found.
Good research on the details - any you may be correct that swapping might have come and gone in a window where we now have no records. Still, it would be good to confirm that ITS 138 did or did not swap. Jnc (talk) 19:07, 29 November 2023 (CET)
Now that I got it running on a PDP-6 emulator, I can readily confirm:
* There is no support for any type of disk storage. The only mass storage is to DECtape.
* I have seen no trace of swapping. It would have been rather painful with DECtapes anyway.
Larsbrinkhoff (talk) 15:17, 30 December 2023 (CET)
True, but there's a big gap between 138 and the first post-disk one we have (671/2), which, I expect, is post the paging box. As you suggested, swapping could have come and gone, in that gap.
Still, 138 has been enormously educational, in terms of what was there, that was still there in later ITS - seeing how far back some of those things came from.
Speaking of the evolution of ITS, have you read ITS HISTRY? A 'must read', if not. It starts just before 946; it would be interesting to try and extend it back. Do you have any versions between 672 and 946, in the backup tapes?
Oh, the first version of CTSS (on the 709) swapped to mag-tapes! :-) Jnc (talk) 17:19, 30 December 2023 (CET)
There are many preserved ITS versions between 672 and 946, binary and/or source. See here: https://github.com/PDP-10/its/issues/672#issuecomment-510027187 Larsbrinkhoff (talk) 15:28, 2 January 2024 (CET)