Difference between revisions of "Talk:ARPANET Host-to-Host Protocol"
(→Additional opcodes: NXR should be 21.) |
(→ICP is hard, let's go shopping!: new section) |
||
Line 10: | Line 10: | ||
: They are octal. I would think the WAITS code is wrong here. ITS has RAP as 20 and NXR as 21. I'll fix the article now. [[User:Larsbrinkhoff|Larsbrinkhoff]] ([[User talk:Larsbrinkhoff|talk]]) 10:12, 5 January 2025 (CET) | : They are octal. I would think the WAITS code is wrong here. ITS has RAP as 20 and NXR as 21. I'll fix the article now. [[User:Larsbrinkhoff|Larsbrinkhoff]] ([[User talk:Larsbrinkhoff|talk]]) 10:12, 5 January 2025 (CET) | ||
+ | |||
+ | == ICP is hard, let's go shopping! == | ||
+ | |||
+ | The section about ICP isn't entirely correct. The server is allowed to send the new U,S RST and STR in parallel with the first W,U STR. (How is the client supposed to know about S then? you might ask. Good question!) | ||
+ | |||
+ | Also, the use of U+2 and U+3 seems to be mandated, not optional. Do you think otherwise? If so, that might be how a client can recognize and match the new RST/STR messages. [[User:Larsbrinkhoff|Larsbrinkhoff]] ([[User talk:Larsbrinkhoff|talk]]) 11:57, 9 January 2025 (CET) |
Revision as of 11:57, 9 January 2025
Additional opcodes
Are the numbers in octal or decimal? (NIC #8246 uses decimal.) If you've just cut-and-pasted PDP-10 source, they're probably in octal there.
Also, you have given both RAP and NXR as '20' - I assume there's an error there?
Please don't edit the article page momentarily; I have a huge wedge of text in preparation for it. Thanks! Jnc (talk) 17:45, 3 January 2025 (CET)
Done. (Although I still have to make sure content is corrected divided between here and Network Control Protocol.) Jnc (talk) 09:55, 5 January 2025 (CET)
- They are octal. I would think the WAITS code is wrong here. ITS has RAP as 20 and NXR as 21. I'll fix the article now. Larsbrinkhoff (talk) 10:12, 5 January 2025 (CET)
ICP is hard, let's go shopping!
The section about ICP isn't entirely correct. The server is allowed to send the new U,S RST and STR in parallel with the first W,U STR. (How is the client supposed to know about S then? you might ask. Good question!)
Also, the use of U+2 and U+3 seems to be mandated, not optional. Do you think otherwise? If so, that might be how a client can recognize and match the new RST/STR messages. Larsbrinkhoff (talk) 11:57, 9 January 2025 (CET)