Re: Old Compiler book - Compiler techniques, circa 1972

rpw3@rpw3.org (Rob Warnock)
08 Aug 2015 08:57:12 GMT

          From comp.compilers

Related articles
Old Compiler book - Compiler techniques, circa 1972 ed_davis2@yahoo.com (Ed Davis) (2015-08-06)
Re: Old Compiler book - Compiler techniques, circa 1972 rpw3@rpw3.org (2015-08-08)
| List of all articles for this month |

From: rpw3@rpw3.org (Rob Warnock)
Newsgroups: comp.compilers
Date: 08 Aug 2015 08:57:12 GMT
Organization: Rob Warnock, Consulting Systems Architect
References: 15-08-010
Keywords: books, history, comment
Posted-Date: 08 Aug 2015 09:26:43 EDT

Ed Davis <ed_davis2@yahoo.com.dmarc.email> wrote:
+---------------
| I was doing a book search, and ran across this:
| Compiler techniques, by Bary W. Pollack, 1972.
|
| However, I have not been able to find a description of this book.
| Since it is pretty old, I'm curious about the state of Compiler writing
| in 1972. Can anyone supply some information regarding this book?
+---------------


I just so happen to have a copy on the shelf... *there*!! ;-}


First, it's "Compiler Techniques, *edited* by Bary W. Pollack"
[emphasis added]. That is, it's a collection of articles by 30+
contributors -- some written for the book and some republished from
other sources [e.g., CACM, JACM] -- grouped by topic into eight
chapters, with a short overview for each chapter by Pollack. Some of
the names you will surely know -- e.g., Don Knuth, David Gries, Robert
Floyd, Doug McIlroy, Bill McKeeman, Alan Perlis, Saul Rosen, Ravi
Sethi, Jeff Ullman -- while others will probably not be so familiar.
The articles vary a lot in quality, as you might imagine. But Pollack
did a fairly good job of assembling them into more or less a coherent
whole, which you might find of some historical interest if you care,
as you said, "about the state of compiler writing in 1972".


It's 558 pages long, but the index is terribly skimpy [more peoples'
names than topics!!]. As far as I can see, the index makes *no*
reference to the terms "register", "graph" [though "binary trees" get
one ref], "color" or "coloring", "loop", "unroll", nor even "branch".
And of course no mention of "SSA" or threads or parallelism or caches.
[Indeed, many machines back then *had* no caches!!] Instead, in
common with the concerns of the day, there's an awful lot of space
spent on parsing.


+---------------
| [Never heard of it. But you can buy a used copy for $4 on
| Amazon so why not just get one and see if it's any good? -John]
+---------------


John's suggestion of just buying a used copy is a good one,
though shipping may cost you more than the book! ;-}


It also seems to be listed on Google Books, so you might look there to
see if any of the content is available, e.g., the Table of Contents.
[Sorry, I don't feel up to typing it in myself.]


Hope that helps,


-Rob


-----
Rob Warnock <rpw3@rpw3.org>
627 26th Avenue <http://rpw3.org/>
San Mateo, CA 94403
[Google Books has it, but only shows two pages from the front matter.
The $4 price at Amazon is 1c for the book, $3.99 for shipping. What a
deal. -John]


Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.