Re: Questions regarding Octave

Scott Seidman <namdiesttocs@xxxxxxxxxxxxxx> wrote in
news:Xns9A6FA6FFBFEB4scottseidmanmindspri@xxxxxxxxxxx:

Karl Molnar <karl.molnar@xxxxxxxxxxxx> wrote in news:fsjb3j$kom$1
@news.al.sw.ericsson.se:

Maybe the Mathworks should just come out and say they don't want to
change it, that it wouldn't be worth their while to fix it for the 3 or
4 people who care, and, in all likelihood, who wouldn't buy it anyway
if they did.

You would think that if it were all that important it would find a fix
in a gnu licensed open source package like octave!




This argument has been going on for a long time. Two years ago, Grant
Griffin and I posted the following (It was supposed to be April Fool's Day
but I was a little late):

Press - Release

April II, MMVI
(April 1, 2006 for those barbarian 0 index DSP guys)

After careful consideration, Mathwork's Cleve Moler had this announcement
to make:

It's been brought to our attention that many of our DSP customers are
unhappy with our choice of one-based indexing. We were relieved when dsp-
guru, Grant Grifffn made such a fine argument supporting our long held
convention. Since Mr Griffin is clearly the leader of all the DSP
community, being the dsp-guru and all, we believe we can now put this
matter to rest.

We also did some historical research. The Romans created many engineering
marvels during the height of their empire and they did not use 0 in their
numbering system. It is our belief that DSP barbarians (with their
heretical 0 based numbering notions), invaded from the north and set back
Western Civilization for over a thousand years.

Therefore to protect our current civilization from a similar fate,
Mathworks has decided that we have a responsibility to change our numbering
system to respond only to Roman Numerals. We know that this might be a
small inconvenience to a few of you but we are sure that you will all go
along when you consider the bigger issues.

Supporting Argument Follows:


Grant Griffin <nospam@xxxxxxxxx> wrote in news:4a503$442e975f$4088dbc7$1304
@EVERESTKC.NET:

Am I the only person who's noticed that C's zero-based indexing is
nothing short of...well...horrible? And it doesn't end there: C has
been so influential that its insanity has spread to tons of other
languages: C++, Java, C#, Python, Perl--heck, even LISP.

Am I the only one here who's suffered a needless off-by-one bug as a
result of C's horrible zero-based indexing? Let's see a show of hands...

But more important, to those of us in DSP, is the fact that C's indexing
doesn't conform to Matlab--which wisely employs one-based indexing in
accordance with the centurys-old convention of linear algebra. Now, one
might ask, "who died and left Matlab in charge?" Well, nobody, I guess.
But if the Matlab folks hadn't gotten this right, why would so many
DSP folks shell out *thousands* for Matlab, when zero-based alternatives
like SciPy (http://scipy.org/) are absolutely, 100%, free? There can be
only one answer: one-based indexing.

OK, I can already hear some of you about to cite the old "legacy code"
chestnut as a reason C can't be changed. True, there's a lot of C code
out there. But there's absolutely no reason that C can't switch to
one-based with a zero-based compatibility mode. OK, I know, I know: C++
programmers can easily create their own one-based array object. But
what sense does it make for countless programmers to home-brew a
solution when Dennis Ritchie could just get off his duff and fix the
problem once-and-for-all? After all, Grace Hopper did it for FORTRAN.

Sound radical? I don't think so: let's face it, we live in a one-based
world. For example, ask a child to count their fingers, and they'll
invariably assign "one" to the first (er, "zeroth") finger.

Still not convinced? OK, what about the calendar?--what if today was
"April 0th"?

=g2
_____________________________________________________________________

Grant R. Griffin
Publisher of dspGuru http://www.dspguru.com
Iowegian International Corporation http://www.iowegian.com
See http://www.iowegian.com/img/contact.gif for e-mail address

.

Relevant Pages

  • Re: Questions regarding Octave
    ... Since Mr Griffin is clearly the leader of all the DSP ... Am I the only person who's noticed that C's zero-based indexing is ... But if the Matlab folks hadn't gotten this right, ... solution when Dennis Ritchie could just get off his duff and fix the ...
    (comp.dsp)
  • Re: horrible C indexing
    ... unhappy with our choice of one-based indexing. ... Since Mr Griffin is clearly the leader of all the DSP ... heretical 0 based numbering notions), invaded from the north and set back ... But if the Matlab folks hadn't gotten this right, ...
    (comp.dsp)
  • Re: Duplicate Results/Catalog corruption?
    ... Forcing an incremental rescan through the MMC seemed to fix the problem - ... >> We're experiencing duplicate results from an index server catalog. ... >> This problem has persisted past restarting indexing service, ...
    (microsoft.public.inetserver.indexserver)
  • Re: Event ID: 3014 Source: Microsoft Search, Category:Gatherer
    ... what if it can't fix the issue, but leaves the indexing in a modified state? ... but when I look at file path files are there physically. ...
    (microsoft.public.exchange.setup)
  • Re: Control login or automatic service startup
    ... >> party DSP if the user logs in while the DSPs are loading. ... > Then you would know where to put the patch (fix, workaround, etc.). ... The app should launch an exe like ActualWinlogon.exe. ... > Setup service dependencies. ...
    (microsoft.public.windowsxp.embedded)