So from here:

https://www.ibm.com/legal/dmca

I wrote to:

[email protected]

to ask them the below. But I got a reply of:

Address not found
Your message wasn't delivered to [email protected] because the address couldn't 
be found or is unable to receive email.
The response from the remote server was:
550 5.1.1 User Unknown


I'm currently in Ligao, Albay, Philippines. I'm reluctant to try to phone
my "local" IBM branch to try to explain the situation.

Oh well ...



Hi IBM legal team.

Sometime around 1977, IBM released DOS/VS R34.

I believe this entire platform is considered to be public
domain, or at a minimum it was freeware.

I recently discovered that when a DOS/VS executable
that does I/O is linked, the following modules are
automatically linked in to the executable itself:

IJJFCBZD
IJGUIZZZ
IJGUOZZZ

I'm only interested in the object code, not the source code -
but as I noted - I believe even the source code is considered
to be public domain.

(Note that I was also able to separately punch the object
code using RSERV).

Given that these modules - in object form - originate from
IBM source code, these are considered to be part of the
"runtime library" (even for assembler programs).

In modern times, companies like Microsoft still claim
copyright over their runtime libraries, but they are
normally much bigger. Either way, they spell out what
you can do with their runtime library that is linked in
to a customer's executable.

The (relevant) IBM runtime library is extremely small,
perhaps 2000 bytes of machine code (I can try to get
a more accurate byte count if you think it is relevant).

However, at the machine code level, even 2000 bytes
is not "trivial", except for the man who doesn't have to
write it himself.

I have used a disassembler to convert the object code
into (generated) source, to simplify the external
("pdld" on the PC) link procedure, and I am currently
trying to see if I can get that to work. (the disassembler
I used didn't do a perfect job).

Note - I used disasm.zip from here:
http://csg.uwaterloo.ca/sdtp/watcutil/index.html
and ran it on my Windows 2000 system.
(DOS/VS isn't the only ancient system I use)

Note that other environments like MVS, CMS,
MUSIC/SP and MTS do not link in any runtime library
like DOS/VS is doing.

In the short/medium term, I would like to include this
disassembled code in my own assembler code (which
is hand-written and explicitly public domain). And the
code would basically be unchanged for now, although
I might improve it over time for readability.

In the long term, I want to focus on the actual SVCs
that this (and other) code invokes, and I have my own
OS (PDOS-generic) that will be done in such a way
that C applications "directly", but conditionally (the
other condition would use a callback instead, thus not
require a real SVC or associated privilege) call the
raw SVC, so that I can eliminate this complicated
assembler. Something like int86x() on MSDOS, if
you're familiar with that.

In order to eliminate this 2000 bytes (estimate) machine
code (now disassembled assembler code), I will need to
decipher what it does (non-trivial for the man who has to
do it himself). You could potentially call this
reverse-engineering.

But since the original source code is available (I believe),
I don't really need to reverse engineer it.

(Note that I am NOT saying that I am not allowed to
reverse engineer this code - regardless of whether it is
necessary or not. I'm just pointing out that you MAY
wish to use that term and it MAY (or may not) be
technically correct).

I'm just doing it to keep away from the original IBM source
code, and focus on the object code/machine code.

But I don't want IBM to accuse me of stealing 2000 bytes
or whatever of their IP, and jeopardize 30+ years of effort
writing PDOS/PDPCLIB (see pdos.org), which is tens of
thousands of lines of C source code (mostly).

Could you please tell me what your position is on this code
that is being linked in to presumably almost all DOS/VS
executables since the 1970s? Presumably you're not going
to claim copyright over other people's executables for the
last 50+ years. Or perhaps 21st Century Software who
distribute VSEn would be the ones who would attempt to
assert copyright?

But I'd rather not be the one that is singled out, to test some
obscure theory of the law, to which I have no answer to, and
some judge won't have any clue whatsoever about it, and
will presumably rule in IBM's favor because IBM lawyers
sound professional, while I don't. Plus the fact that it is
undisputed that IBM wrote the actual (freeware back then) code.

So I'd like to know where I (and technically thousands upon
thousands of other DOS/VS to VSEn executables written
by non-IBM people) stand.

Thanks. Paul.

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to