Reading Group?

David Klann dklann at grunch.org
Fri Mar 27 19:16:42 UTC 2020


Great suggestion Brian! And a helpful, informative reply Timothy! I have 
interest in participating, and limited time to do so...

On 3/27/20 1:02 PM, Timothy J. Salo wrote:
> On 3/26/2020 7:59 PM, Brian H wrote:
>> I was wondering if there would be any interest in a reading group for
>> the book "The Art of Computer Programming" starting with volume 1. I
>> am not a programmer by trade, but work with code very closely for
>> work. I was hoping to read this to further my understanding of
>> computer science in the academic sense and I was hoping others would
>> join me on this journey so I am not alone.
> 
> I'm not sure I would start with Knuth.  It's a bit dated and a bit
> peculiar.  And, I don't think that I have seen it used as a textbook
> for quite some time.  I don't see my copies of Knuth (I think they are
> in my storage locker).  But, for example, Knuth defines an
> imaginary machine architecture, the MIX machine, and uses that for
> a bunch of his examples.  Today, algorithms are more likely to
> be described in Java, Python, or just pseudo code.  In my view,
> getting familiar with the MIX machine seems sort of pointless.
> Or, maybe this was dropped from the current editions.
> 

I'm replying with this info not to be pedantic, simply to set the record 
straight and to lobby *for* Knuth. :) And full disclosure, I am not an 
academic, nor did I even study computer science in college. I am an 
amateur computer scientist (at best). This may be, at least in part, an 
emotional response: Knuth has been a personal hero of mine for quite 
some time.

My sweetie gifted me a current copy of Knuth last Fall. There are four 
volumes in the series on my shelf: Volumes 1 and 2 are in their third 
editions, Volume 3 its second, and Volume 4A its first edition. In the 
preface to Volume 1 Knuth states that the complete set will have five 
volumes with a total of ten chapters. He also says he wants to update 
Volume 1 with a fourth edition ... in fifteen years or so (here's to his 
long life!).

I think I understand your point about not learning algorithms using 
Knuth's generic MIX machine. But it also seems there might be value 
learning in a popular-language-agnostic way. It sort of frees your mind 
from the idiosyncrasies of the language du jour (of course the MIX 
machine has its own).

> If you want to [really] study algorithms, I recommend "Introduction
> to Algorithms" by Cormen, et al.  But, Cormen probably expects some
> background in discrete math and probably even in algorithms.  There
> are also another one or two widely used book on algorithms.  I think
> I have pointers to these somewhere.
> 
> A reading group in discrete math might be a good topic for those
> who want some background in academic computer science.  I think
> that Rosen's "Discrete Mathematics and its Applications" is still
> used, but there are several good texts available.  Every computer
> science curriculum has (or should have) a course in discrete
> math.  You can check and see what departments are using as a
> textbook.  Or, you can dig around Amazon and find something that
> people like, but isn't priced as a textbook.
> 
> You might also consider "cracking the Coding Interview" by
> McDowell.  I have described this as an introduction to
> computer science for non-majors.  In particular, it doesn't
> require the math background that an introductory computer
> science course would.  Plus, it's a lot cheaper than a
> textbook.
> 
> There are undoubtedly other good intro to computer science
> books.  But, I would consider alternatives to Knuth.
> 

These are wonderful suggestions and recommendations. I favor Knuth 
simply because of his long view of the subject and the "art" of computer 
science. I guess it depends on ones goal. In my case, I want to learn 
about math and computer science; I don't need this to advance my career. 
I think Knuth presents the foundations in a more-or-less "timeless" way.

> By the way, it may not be in the current editions, but in
> the introductions to the early editions, Knuth stated that
> he intended to cover all of computer science in 17 volumes.
> I don't think he got past volume 3.
> 
> -tjs

Thanks again for your views and suggestions! I intend to look at some of 
these as well.

   ~David Klann


More information about the Friends mailing list