MathGroup Archive 2011

[Date Index] [Thread Index] [Author Index]

Search the Archive

Re: How to write a "proper" math document

  • To: mathgroup at smc.vnet.net
  • Subject: [mg119990] Re: How to write a "proper" math document
  • From: Richard Fateman <fateman at cs.berkeley.edu>
  • Date: Mon, 4 Jul 2011 06:44:46 -0400 (EDT)
  • References: <201107020902.FAA10576@smc.vnet.net> <iup8au$kse$1@smc.vnet.net>

It is quite simply the case that any system which requires a reader or 
writer of a document to pay WRI an annual license will be rejected by 
the vast majority of people, as well as any journal published by a 
reputable academic/ scientific non-profit organization such as IEEE, 
AMA, ACM.

It appears that Mathplayer (etc) don't do the job for what you are 
talking about.

There are several competing products in this active-math document 
editing framework, including web-page based products which allow (they 
claim) easy call-outs to programs and programming languages without 
going through Mathematica, Mathlink, etc.

Mathematica is hardly the first language of choice for most of the 
world, and in spite of what you may think of its graphical, interactive, 
animation (whatever) is hardly unique.

I would add (though clearly others have expressed disagreement on this 
point) that embedding Mathematica semantics in numerical and some 
symbolic computations would be a grave disservice to actual scientific 
calculations.  Consider the situation when the most important 
characteristics of a formula are those locations where it is singular, 
but Mathematica leaves them out because they are not part of the 
"generic" solution.  Or the numerical computation which should be done 
to higher precision, but Mathematica displays the answer as 0. 
Neglecting to mention that it really could be anything at all, just that 
it has apparently lost all significance.

I am generally in favor of the notion that engineering or other 
scientific/mathematical journal articles or documentation should include 
material that can be cut out and pasted elsewhere, including in a 
computer algebra system.  Or having them be "clickable".

One way of doing this is via what is sometimes called "multivalent" 
documents.  See
http://www.eecs.berkeley.edu/Pubs/TechRpts/1998/CSD-98-999.pdf

If you look at this (circa 1998) paper, you will see that it shows, in 
principle, how one can take an existing document that includes math 
expressions, and via OCR and some help, present an additional layer to 
the reader that provides active computation or extraction into TeX.

Constructing a paper initially in digital form removes the need for OCR, 
but the multiple-layer notion seems like a good one.  Whether you 
consider that Mathematica provides all you need now or not, do you think 
that a document prepared with version 8.0 will still "work" in a later 
version?  Will you be able to effortlessly attach a layer of 
annotations, corrections, 3D-smell-o-vision, etc. in the future? Will 
you have layers of historical versions of the past?  Will you be able to 
send it to a colleague who does not have a license?

RJF



  • Prev by Date: Can I use NonlinearModelFit to model some data with a piece wise linear function??
  • Next by Date: Re: Insoluble marbles-in-urn problem?
  • Previous by thread: Re: How to write a "proper" math document
  • Next by thread: Re: How to write a "proper" math document