Talk:Dope vector

Prior to the invention of the linked-list...

That statement goes beyond the level of "citation needed": an old-timer like me can tell you it is completely implausible, citation or no citation.

Linked lists were well-established by 1961; indeed, they are the stock in trade of LISP. I never heard of a "dope vector" until IBM used that terminology in argument passing in PL/1 for the 360, which came out in 1965. The suggestion that there was a time before the invention of the linked list when dope vectors were used instead is, to say the least, temporally implausible. 72.94.107.197 (talk) 22:48, 4 September 2013 (UTC)[reply]

I think "dope vector" originated with ALGOL somewhat before PL/I, but that doesn't negate your argument. I'd assume linked lists existed from day one. Peter Flass (talk) 23:18, 4 September 2013 (UTC)[reply]

Eliminated the statement, the article still seems confused.Peter Flass (talk) 23:28, 4 September 2013 (UTC)[reply]

call by ...

I am used to call by value and call by reference, and there should be a name for calls that use dope vectors, more generally, descriptors. The VAX calling conventions (defined along with the VAX hardware, not VMS as might be more obvious) uses %VAL(), %REF(), and %DESCR(), to override the language default, allowing for cross language calls. But there is no call by descriptor page. The actual reason for writing this, is that PL/I is described as using call by reference, which it does for numeric scalars, but uses dope vectors for arrays, strings, and structures. What is the name for a calling convention that uses dope vectors? (Strangely, calling convention seems to be the low-level, system specific page, not the one that is a level up from call by value, call by reference, and call by name. Gah4 (talk) 19:55, 7 May 2019 (UTC)[reply]