Re: HLAs
From: James E. Morrison (janus.astrolabe_at_att.net)
Date: 03/09/04
- Next message: Jeff: "Re: It really looks like the way to do things is API"
- Previous message: Jim Carlock: "It really looks like the way to do things is API"
- In reply to: Randall Hyde: "Re: HLAs"
- Next in thread: hutch--: "Re: x86 Assembly Language FAQ"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: Tue, 9 Mar 2004 05:09:28 +0000 (UTC)
"Randall Hyde" <randyhyde@earthlink.net> wrote in message news:<SQz2c.185$Cm3.46@newsread3.news.pas.earthlink.net>...
> "Scott Moore" <samiam@moorecad.com> wrote in message
> news:Rzu2c.71552$ko6.411009@attbi_s02...
> >
> > Yes, the 1950s saw the rise of the Macro assember, and many assembly
> > programmers felt it qualified as a HLL in an of itself. For example,
> > it was (and is) possible to write a complete system that didn't require
> > any assembly code at all:
> >
> > whil count
> > add myvar, thatvar
> > prt myvar
> > sub count, 1
> > ewhl
> >
> > All being macros, etc.
> (snip)
> Cheers,
> Randy Hyde
I'm not sure where this thread started, so my comments may be
off-topic or on a slightly different topic. My own observation is
that the definition of what constitutes an High Level Assembler is
very time dependent.
The first programming aid that I encountered that qualifies as an HLA
in my mind is SOAP for the IBM 650. The IBM 650 had a drum memory and
coding for it was very tedious since you had to place instructions on
the drum in specific locations in order to minimize the drum latency
between instruction fetches. SOAP was written by Stan Poley of IBM in
1955. I knew Stan well and was with him when he died. SOAP had none
of the structured programming constructs of later languages, but it
made programmers much more productive, and that would seem to qualify
it as an HLA for that era.
The FORTRAN Assembler Program (FAP) for IBM's line of floating point
systems in the 50's (704/709/7090/7094) had a macro generator and
built in super instructions for common data conversion tasks. I don't
believe it was widely used because FORTRAN had become fairly complete
by the time FAP was widely distributed, but it was used when really
tight code was needed.
The Autocoder assemblers for IBM's decimal machines (e.g., 1401/1440,
705/7080, 7070, etc.) had crude macro generators, but little else in
the way of programming tools and would not qualify.
However, the macro assemblers created for the IBM/360 line announced
in 1964, might qualify on the grounds that the built in macros
provided a relatively high level interface to operating system
functions. Some (Many? Most?) of the OS/360 and DOS system calls
required quite a few parameters to be passed. The macros provided
made it quite easy to write assembler system code that interfaced with
system functions, and many extraordinary system elements were written
using this platform. These assemblers seem to qualify if programming
efficiency and productivity is relevant to whether an assembler
qualifies as an HLA. As an example, think how hard it would be to
write an assembler program to interface with TCP/IP without this sort
of function.
Another element that I have not seen in this thread is any mention of
the role linkers play in the language capabilities. The linkage
editor provided with IBM's OS/360 was the first to allow complex jobs
to be created from source code from many programmers and was the first
facility that I know of that made the concept of reusable code
realistic. Unfortunately, IBM never released the library tools that
were used internally to get the most out of this technology.
Best regards,
Jim
James E. Morrison
janus.astrolabe@att.net
- Next message: Jeff: "Re: It really looks like the way to do things is API"
- Previous message: Jim Carlock: "It really looks like the way to do things is API"
- In reply to: Randall Hyde: "Re: HLAs"
- Next in thread: hutch--: "Re: x86 Assembly Language FAQ"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Relevant Pages
|
