On Fri, 9 Feb 2018 17:39:00 -0800, retired mainframer wrote:
>Since pointer arithmetic is performed in terms of object size, you want to
>do the arithmetic on a char pointer because sizeof(char) is guaranteed to be 1.
>
For (some?) members of the Scientific Data Systems (SDS) Sigma Family
http://s3data.computerhistory.org/brochures/sds.sigma.1967.102646100.pdf
...
When indexing is used, the index register automatically scales itself
according to the size of the data involved for easier programming.
For example, if the selected index register contains a 27, an indexed
Load Byte instruction automatically loads the 27th byte; an indexed
Load Word instruction automatically loads the 27th word, etc. ...
FSVO "easier". THe scheme seems peculiarly ill-suited for accessing various
sized members of structures in arrays. But, then FORTRAN was the only
programming language you'd ever use. And I faked arrays of structures by
using parallel static arrays of simple values, even for some list processing.
-- gil
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
