Difference between revisions of "Talk:TENEX"

From Computer History Wiki
Jump to: navigation, search
(Remove comment about category, it looks good now.)
(Hierarchical file system?: Soon to be tested on a running system.)
 
(4 intermediate revisions by 2 users not shown)
Line 4: Line 4:
  
 
Also, according to "TENEX and TOPS-20", by Dan Murphy (IEEE Annals 37-1, pp. 75-82), BBN actually for a number of years had a 940 with the Berkeley OS on it; so they would have had personal experience with it. [[User:Jnc|Jnc]] ([[User talk:Jnc|talk]]) 17:28, 29 March 2018 (CEST)
 
Also, according to "TENEX and TOPS-20", by Dan Murphy (IEEE Annals 37-1, pp. 75-82), BBN actually for a number of years had a 940 with the Berkeley OS on it; so they would have had personal experience with it. [[User:Jnc|Jnc]] ([[User talk:Jnc|talk]]) 17:28, 29 March 2018 (CEST)
 +
 +
: Thanks.  I did check the Murphy papers, but it doesn't explicitly say that parts could be a close translation, which is what I find interesting.  Hopefully, the 940 sources will be available soonish. [[User:Larsbrinkhoff|Larsbrinkhoff]] ([[User talk:Larsbrinkhoff|talk]]) 18:23, 29 March 2018 (CEST)
 +
 +
==Hierarchical file system?==
 +
 +
I'm not sure, but I have the impression there was no directory hierarchy.  [https://opost.com/tenex/tenex72.txt Murphy's paper] says "We are currently considering the feasibility of implementing a full tree directory structure similar to MULTICS".  [https://www.ultimate.com/phil/pdp10/tops-20 Budne's TOPS-20 Evolution] says it was added in TOPS-20 V3; I have verified this running V3A. [[User:Larsbrinkhoff|Larsbrinkhoff]] ([[User talk:Larsbrinkhoff|talk]]) 10:09, 21 April 2022 (CEST)
 +
 +
: Oh, could be. I assumed there were because of my use of TOPS-20 (on MIT-XX) which had them, and since TOPS-20 was based on TENEX... Let me check it out in the TENEX docs (Dan Murphy's site has the ACM paper - I think I have a photocopy somewhere I made 4 decades ago! :-), and then I'll fix the page. [[User:Jnc|Jnc]] ([[User talk:Jnc|talk]]) 18:51, 21 April 2022 (CEST)
 +
 +
: Yeah, I think you're right. [https://opost.com/tenex/tenex72.txt TENEX, A Paged Time Sharing System for the PDP-10] "1.0 INTRODUCTION" says "File system .. built on multi-level symbolic directory structure", but then "5.0 THE TENEX FILE SYSTEM" says "A symbolic name for TENEX files consists of up to five fields [below - JNC] and thus conceptually represents a tree of maximum depth five. .. This scheme was chosen rather than a full tree to simplify the problem of compatibility with existing DEC PDP-10 software .. We are currently considering the feasibility of implementing a full tree directory structure similar to MULTICS", so it sounds like not. (The 5 fields are "device name, directory name, file name, extension, and version number".) [[User:Jnc|Jnc]] ([[User talk:Jnc|talk]]) 22:02, 21 April 2022 (CEST)
 +
 +
:: Thanks!  I hope to verify this empirically soon.  We have the SUMEX-AIM monitor, based on TENEX 1.31, running on a PDP-10 emulator, and are working on bringing up the rest of the system. [[User:Larsbrinkhoff|Larsbrinkhoff]] ([[User talk:Larsbrinkhoff|talk]]) 07:04, 22 April 2022 (CEST)

Latest revision as of 06:04, 22 April 2022

SDS940 connection

The first/main TENEX paper ("TENEX, A Paged Time Sharing System for the PDP-10", by Bobrow, Burchfield, Murphy and Tomlinson) specifically calls out the 940 as one of the systems they used as a model. Jnc (talk) 17:17, 29 March 2018 (CEST)

Also, according to "TENEX and TOPS-20", by Dan Murphy (IEEE Annals 37-1, pp. 75-82), BBN actually for a number of years had a 940 with the Berkeley OS on it; so they would have had personal experience with it. Jnc (talk) 17:28, 29 March 2018 (CEST)

Thanks. I did check the Murphy papers, but it doesn't explicitly say that parts could be a close translation, which is what I find interesting. Hopefully, the 940 sources will be available soonish. Larsbrinkhoff (talk) 18:23, 29 March 2018 (CEST)

Hierarchical file system?

I'm not sure, but I have the impression there was no directory hierarchy. Murphy's paper says "We are currently considering the feasibility of implementing a full tree directory structure similar to MULTICS". Budne's TOPS-20 Evolution says it was added in TOPS-20 V3; I have verified this running V3A. Larsbrinkhoff (talk) 10:09, 21 April 2022 (CEST)

Oh, could be. I assumed there were because of my use of TOPS-20 (on MIT-XX) which had them, and since TOPS-20 was based on TENEX... Let me check it out in the TENEX docs (Dan Murphy's site has the ACM paper - I think I have a photocopy somewhere I made 4 decades ago! :-), and then I'll fix the page. Jnc (talk) 18:51, 21 April 2022 (CEST)
Yeah, I think you're right. TENEX, A Paged Time Sharing System for the PDP-10 "1.0 INTRODUCTION" says "File system .. built on multi-level symbolic directory structure", but then "5.0 THE TENEX FILE SYSTEM" says "A symbolic name for TENEX files consists of up to five fields [below - JNC] and thus conceptually represents a tree of maximum depth five. .. This scheme was chosen rather than a full tree to simplify the problem of compatibility with existing DEC PDP-10 software .. We are currently considering the feasibility of implementing a full tree directory structure similar to MULTICS", so it sounds like not. (The 5 fields are "device name, directory name, file name, extension, and version number".) Jnc (talk) 22:02, 21 April 2022 (CEST)
Thanks! I hope to verify this empirically soon. We have the SUMEX-AIM monitor, based on TENEX 1.31, running on a PDP-10 emulator, and are working on bringing up the rest of the system. Larsbrinkhoff (talk) 07:04, 22 April 2022 (CEST)