Loading... |
Reply to author |
Edit post |
Move post |
Delete this post |
Delete this post and replies |
Change post date |
Print post |
Permalink |
Raw mail |
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Does SPSS use COBOL as a programming language? Or did SPSS use COBOL as a programming language in the past?
|
Loading... |
Reply to author |
Edit post |
Move post |
Delete this post |
Delete this post and replies |
Change post date |
Print post |
Permalink |
Raw mail |
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
When I started using SPSS in 1972, there was documentation on how to use FORTRAN to contribute. This allowed one to use all the data definition etc without reinventing it. There wee a skillion subroutines one could call.
One had to be careful when using DEC FORTRAN which had many features that were not available on other platforms. For example, DEC FORTRAN always cleared memory before assigning it to an individual user. On some other platforms it was imperative to initialize memory. I do not recall if it was early, but by 1977, DEC FORTRAN had a PARAMETER statement, for example in adapting INDSCAL on the DEC ten to handle by 210 10x10 similarity matrices for my dissertation I could change DIMENSION statements something like this in a file Parameter NSTIM =10 Parameter NSUBJ =210 The main code would have some statement to include the small file and then Dimension NDATA (NSUBJ, NSTIM, NSTIM) One also had to avoid using some kinds of DIMENSION statements like Dimension IPCT (-100, 100) Dimension IYEAR (1776, 1980) One also could not rely on automatic checking for a subscript being out of bounds for example if a value of IYEAR were 1770 when the array was Dimension IYEAR (1776, 1980)
Art Kendall
Social Research Consultants |
Loading... |
Reply to author |
Edit post |
Move post |
Delete this post |
Delete this post and replies |
Change post date |
Print post |
Permalink |
Raw mail |
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
This post was updated on Sep 25, 2022; 7:14pm.
In reply to this post by Manfred
AFAIK, SPSS never used any COBOL code. However, there are some input formats such as zoned decimal available for use with DATA LIST that would suit COBOL files.
|
Loading... |
Reply to author |
Edit post |
Move post |
Delete this post |
Delete this post and replies |
Change post date |
Print post |
Permalink |
Raw mail |
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
In reply to this post by Art Kendall
I never used a Fortran that used Dimension for ranges, like
Dimension IPCT (-100, 100) Dimension IYEAR (1776, 1980) Maybe after 1990? If that was added, I would expect maybe a slash instead of comma - so that the old performance would work. In old-style Fortran, when SPSS was new, when you gave Fortran's Dimension two values (always positive), you got a matrix. -- Rich Ulrich |
Loading... |
Reply to author |
Edit post |
Move post |
Delete this post |
Delete this post and replies |
Change post date |
Print post |
Permalink |
Raw mail |
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
<head slap> you are right it was a slash. My only excuse is that it a long time since 1971 or 1972 when I learned DEC FORTRAN
Since the default was to leave index checking turned on one would be notified if and index was out of range. On DEC machines one could have matrices where each of dimensions could have specific values something like this to make code that was easily adaptable to specific problems parameter minpct(-100) maxpct(100) minyr (1776) maxyr (1980) double precision mytab (minpct/maxpct, minyr (minyr/maxyr) integer pct year 2 read(7, 10, end=15) ICASE pct year dv 10 format(2I,F) mytab(pct,year) = mytab(pct,year) +dv go to 2 15 continue DEC Fortran always passed clean memory. This was not the case on some other brands. On some other brands one could do something like this and see all kinds of things Dimension x (100000) write (5,100) x 100 format (100000E18.16) write(6,200) x 200 format (100000A8)
Art Kendall
Social Research Consultants |
Free forum by Nabble | Edit this page |