Re: Re: Book recommendation for
- To: mathgroup at smc.vnet.net
- Subject: [mg9539] Re: [mg9498] Re: Book recommendation for
- From: seanross at worldnet.att.net
- Date: Thu, 13 Nov 1997 01:40:04 -0500
- Sender: owner-wri-mathgroup at wolfram.com
Sergio Rojas wrote: > > Ersek_Ted%PAX1A at mr.nawcad.navy.mil wrote: > > : |Even though one may find > : |a lot of references about Mathematica, few of them are really useful, > : |and the end result is that the program is still poorly documented, | > : Yes, it bugs me that there is virtually no documentation for the menu > : command in Version 3.0. > : But the Mathematica Book (3rd edition) is 1400 pages, and documents the > : Kernal features very well. > > I still do not get what is it that makes the thick Mathematica 3.0 > book a well documented reference. In my understanding, a well > organized documentation for something will allow the user to find what > is needed without need of going to a THOROUGHLY review of the > documentation. Moreover, it usually happen that while browsing a well > written documentation set, looking for a particular subject, one finds > information about others things that are not of need for that moment, > but will be so later on. This, in my view, is not the case of the > thick Mathematica book, and it can be illustrated in the process of > creating an acceptable 2-D plot ready to be included in a document. > > Let's say there is a file containing 3 columns of data and the idea > is to plot , to make it simple, columns 2 and 3 versus column one > separately. A QUICK look at the book only shows plot of functions, so > one goes to the index. In the way one take notes of pages 133, 492, and > 1042 for AxesLabel and some other things that may be needed to complete > the goal. However, one finds NOTHING close to "plot of data". But, a > somewhat "special" intuition let the initiated suspect that ListPlot > is what is needed, so the user goes to page 157, and there it is !!! > section 1.9.9 "Plotting List of Data". The initiated find that the > example on page at the beginning of page 159 is more or less fine > (except for the labels), and goes to section 1.11.3 in order to read > the data file. After reading section 1.11.3, no way to put the data > in the form given on page 159, so the initiated go to section 2.11.7, > but to find out that still it is NOT SHOW how to read a file with 3 > columns and set up the data in a way allowing the plotting of columns 2 > and 3 versus column one. The end result is that the great thick book is > put to a side, and some other books need to be looked at to find what > is needed. > > Other things to look at are: How many plots, in the great book, > shows the plotting of error bars? How many of them shows how to label > the plots using long or sort labels? How many ... > > Sergio I think you are missing the difference between a good reference, a good textbook and a good tutorial. The mathematica book is, in my opinion, a good reference, with a few mediocre tutorials. It is a lousy textbook. The things you describe, like how to plot specific kinds of data or certain formats are the function of a textbook or tutorial, not a reference. As an example, suppose a certain book was billed as a good C++ reference. Would you expect it to tell you how to write a program to read in certain formats of data and format them on the screen with error bars? Certainly not, for a C++ reference, you would expect to find a detailed description of every command available to you in the language. You would need to turn to a textbook on C++ graphics to learn how to make a pretty graph. Mathematica is no different-it is a programming language, not a program. A final note: I called the mathematica book a good reference, not a great one. To be a great one, they would need to include a list of ALL available options for ALL built-in objects. There are too many objects that only have listings of the default options, not a listing of all options. --