Newsgroups: comp.lang.apl
Path: watmath!watserv2.uwaterloo.ca!torn!cs.utexas.edu!usc!sol.ctr.columbia.edu!The-Star.honeywell.com!umn.edu!csus.edu!sfsuvax1.sfsu.edu!emclean
From: emclean@sfsuvax1.sfsu.edu (Emmett McLean)
Subject: Re: Matrix Inverse in APL2
Message-ID: <1993Jan20.044439.10917@csus.edu>
Sender: news@csus.edu
Organization: San Francisco State University
References: <1993Jan16.105448.26408@csus.edu> <1993Jan18.191141.6303@csi.jpl.nasa.gov> <1499@kepler1.rentec.com>
Date: Wed, 20 Jan 1993 04:44:39 GMT
Lines: 46

In article <1499@kepler1.rentec.com> andrew@rentec.com (Andrew Mullhaupt) writes:
>>

>It should also be pointed out that Hilbert's matrix is a Hankel matrix,
>which means that there is a general class of special methods (e.g. Shur
>decomposition) which will also compute the inverse of this matrix. This
>is also unlike most ill-conditioned matrices. Finally, Hilbert's matrix
>is symmetric, which means that the eigenvalues are all real. This can
>also be exploited to avoid general case difficulties in inversion.
>
 Remarkable! I hadn't heard of Hankel matrices!

>>Hmm, are you forgetting to take absolute values? Or am I just blind.
>
>...Or is J harder to penetrate than its practitioners will readily admit?
>

 Well, Sam is right in that the primary source of my error is inattention
 to the mathematics. Having made the error I find it difficult to blame
 my mistake on the penetrability of J. My thoughts are :

 1. APL my be easier to learn and understand than J.  For example, I think
    APL's bracket indexing into matrices (nouns) is much more straight 
    forward than J's amend.

 2. Even now, but especially when I was learning J, I spend (spent)
    what feels like *too* much time using interactive hit and miss -
    creating verbs which do straight forward things. At different
    times I find this process either boring, frustrating, or brainless 
    fun, like a video game. Comparatively, developing programms in C,
    FORTRAN, or Modula feels like hard work, and can be exasperating.
    I especially hate having to work with dbx and step through a nested
    for loop and inquire about pointer variables, or debug some one
    else's 2000 line C program.  I'd rather go to the dentist -:)

 3. I've been programming in J for 2+ hours each day since mid
    July `92. If J was really *inpenetrable* I'd of given up
    on it. I've spent much more time writting in Modula than in
    J but I am already far more proficient in J.


    (BTW, I have to think about the rest of Andrew's post.)

    Cheers,
  
    Emmett
