Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uunet!zaphod.mps.ohio-state.edu!samsung!emory!ox.com!yale.edu!cmcl2!kramden.acf.nyu.edu!brnstnd
From: brnstnd@kramden.acf.nyu.edu (Dan Bernstein)
Newsgroups: comp.lang.misc
Subject: Re: Will this *thread* ever halt?
Message-ID: <5979.Jun2519.38.4991@kramden.acf.nyu.edu>
Date: 25 Jun 91 19:38:49 GMT
References: <OZ.91Jun23145702@ursa.ccs.yorku.ca> <12358.Jun2400.03.5891@kramden.acf.nyu.edu> <OZ.91Jun25114153@ursa.ccs.yorku.ca>
Organization: IR
Lines: 28

In article <OZ.91Jun25114153@ursa.ccs.yorku.ca> oz@ursa.ccs.yorku.ca (Ozan Yigit) writes:
>    > Those volumes are not out of favor, but they are out-of-date by a
>    > decade and a half,
>    Really?
> Yes, really.

So why don't you back up your claim?

Here, let me give you a head start. Knuth does spend quite a bit of time
on asymptotically fast matrix multiplication. Today we not only know
methods that are asymptotically even faster, but we have a somewhat
better understanding of *why* the methods work. That portion of that
section is clearly out of date, albeit by only a few years.

I can name a few other examples, each deserving an exercise or two and
perhaps a mention in the text. None of them are important, the way that
each new sorting algorithm has found its niche, or that the spectral
test caught on, or that so many memory allocators depend on the buddy
system. None of them could be discussed at Knuth's level for more than
half a page.

C'mon, Oz. You said the books are out of date by a decade and a half.
That means that there have been discoveries throughout the last 15 years
which, had they been known at the time, would have been included in the
books. What are they? I'm perfectly willing to believe that they exist,
but you have to prove it by example rather than rhetoric.

---Dan