Compilers not coming with the OS anymore

From: Huw Davies <H.Davies_at_latrobe.edu.au>
Date: Fri Jun 18 23:36:54 1999

At 16:25 18/06/99 -0700, Derek Peschel wrote:

>_Anyway_... do you know anything about BLISS? DEC people seem to have
>complained about it, as if it was unpopular within DEC.

What would you like to know? I started using BLISS on a DECsystem-10 in
1975 or 1976. BLISS was written under the guidance of William Wulf at CMU.
I've just looked at my copies of the source and the compiler I/O module
(written in Macro-10) is dated 1972.

I'm assuming BLISS was written to investigate new techniques in compiler
optimization and then became a tool used by the PDP-10/20 community. It was
certainly used within Digital to write parts of the new linker
(surprisingly called LINK) and also the FORTRAN compiler.

A new verson of BLISS called BLISS-11 was created (for the PDP-11) and a
cross-compiler written. This project is documented in the book "The Design
of an Optimizing Compiler" which makes interesting reading.

Eventaully BLISS-10 (the orginal version) became BLISS-36 and yet another
version BLISS-32 was created for the VAX. This required considerable
extensions to the language as BLISS-10 was designed for a hardware
architecture where the addressable unit (the word) could contain an
address, not quite the same for a byte oriented system like the VAX.

I've heard that Bliss was not liked within Digital. I'm assuming some
aspects of NIH along with support issues. Perhaps someone who was there
might like to comment?























                                                                          

 Huw Davies | e-mail: Huw.Davies_at_latrobe.edu.au
 Information Technology Services | Phone: +61 3 9479 1550 Fax: +61 3 9479
1999
 La Trobe University | "If God had wanted soccer played in the
 Melbourne Australia 3083 | air, the sky would be painted green"
Received on Fri Jun 18 1999 - 23:36:54 BST