Lost Publications
Computing is a young field, but not nearly as young as the
cited references in many technical papers would have one believe.
The problem is that old papers can be very hard to find.
Very few appear online, and the first papers in each field
often predate the standard places to publish in that field.
Companies also have a nasty tendency to go out of business
or change in various ways that turn old technical documents
into shredded landfill.
It even can be very difficult to know there was earlier work
that one might want to find.
With an impending move to a new building that has less storage
space, it was necessary to go through Professor Dietz's files
to triage which documents to keep and how close to keep them.
Thus, we have begun digitizing some of the publications that
we reference but have been "lost" to the world at large. We
think we have the right to post them at least for "personal
use" within our group -- if you have rights to any of them and
don't want them linked here, please let us know.
-
John Cocke and J. T. Schwartz, "Programming Languages and Their
Compilers, Preliminary Notes, Second Revised Version,"
New York University Courant Institute of Mathematical Sciences,
April 1970.
(PDF)
I believe this is the first relatively complete treatment
of compiler technology, including automated construction of
parsers and methods for analysis and optimization of generated code.
The hand-written mark-up is partly original to the document, but
includes a number of corrections that I made when this document
was used in a compilers course at Brooklyn Poly taught by David
R. Doucette (then an Adjunct whose day job was at Grumman) --
and that course was how I got my copy.
I obtained formal permission from NYU to freely redistribute this
shortly after becoming an Assistant Professor at Purdue back in 1986,
so I believe that I do have the explicit right to post this.
-
AL Hartmann, "iAPX286 Compiler Writer's Guide, Version #1,"
Intel Corporation, May 1983.
(PDF)
I believe this contains one of the earliest treatments
of register allocation with overlapping register classes.
This was given to me by Ken Aupperle shortly after it was
written. At the time, Ken was a friend and an Intel employee
(although he is best known for later founding Hauppauge Digital, Inc.)
and I was teaching a compilers course at Brooklyn Poly
(which is now part of NYU). Many times since, I have asked
Intel for permission to copy/redistribute this and the answer
always has been that "no such document ever existed."
Anyway, here it is. ;-)
The only thing set in stone is our name.