Re: Getting a pure text widget?
- To: mathgroup at smc.vnet.net
- Subject: [mg61390] Re: Getting a pure text widget?
- From: "Steven T. Hatton" <hattons at globalsymmetry.com>
- Date: Tue, 18 Oct 2005 02:44:31 -0400 (EDT)
- References: <email@example.com>
- Sender: owner-wri-mathgroup at wolfram.com
David Park wrote: > Steven, > > I'm certainly glad I don't have Linux! A friend with Linux couldn't use a > style sheet I designed because it used Helvetica font in some of the > titles and apparently Linux doesn't handle Helvetica. I finally changed > the style sheet to accommodate Linux people. Your problems sound even > worse. If the notebook interface works so poorly that you are seriously > thinking of writing your own GUI, then I think you have real cause for > complaint. Maybe you should think of a Mac, or a Microsoft system, at > least for Mathematica. It certainly will be much less expensive than > writing your own GUI. There is nothing wrong with the OS. It handles many other programs without a problem. The Mathematica FrontEnd implementation on Linux is 1995 vintage. > I think that your difficulty with the notebook concept is mostly that you > don't have an implementation in which it works properly. That rather > destroys the chances of getting the most out of it. It's not difficult to > use - when it is working. I don't find it very well structured. > In any case, let's hope that Mathematica and Linux get better in step. > > In Mathematica notebooks you can put much of the source code in special > initialization sections that the reader might not look at, unless they are > curious. You can also put source code in a package. You can hide the input > code for graphics by putting it in a closed cell, which occupies only a > thin space with a bracket at the right. The user can evaluate the cell and > get the graphics without seeing the code. But the code has to be > somewhere, either in a package or special Sections or in closed cells, and > of course in the kernel. Otherwise there will not be an interactive > system. And sometimes you will want the reader to see the code. It depends > on what you're communicating and who you're communicating to. I think it > would be possible to hide all of the source code. Sometimes I like to do > derivations where I use Print statements to annotate the steps. One could > close the input cell and just have the steps with the annotation. Those are all presentational matters, and not directly related to the structure of the notebook, other than the fact that the features operate on parts of the notebook. > There are things in Mathematica notebooks that I would like to see > improved. I've posted on this before. None of the supplied style sheets > are entirely to my taste. Some of them are downright silly. But since we > can design our own style sheets (if we have a system that works!) this is > not a serious point. Except that it would be nice to work with a style > sheet that everybody has. Some of the things I would like to see are: > 1) The font size for text should be more commensurate with the font used > in input and output cells. In scientific papers the fonts are > approximately the same size. > 2) I would like to see group open/close icons on all Section type headers > - but NOT on anything else. The open/close icons are intuitive and > everybody figures out how to use them. Double-clicking the rhs bracket > sometimes eludes even super geniuses. > 3) I would like to see one extra level of Section header. I would like to see support of XML document specifications, and stylesheet which function orthogonally to the content. Ideally, I would like to be able to change stylesheets, and have vectors change representation from bold face, to over arrow, for example. A vector should be tagged as a vector and the form of presentation should be determined by the rendering engine. > If the reader has the option of typing any expository text he wants, if he > can evaluate any command that is in Mathematica, the notebook or a loaded > package, if he can make any kind of graphics he wants, then how can a > palette provide better options? Palettes generally do not give freedom but > restriction. The most useful ones are ones that simply paste very common > constructions that are difficult to type. It is very difficult to design > general purpose palettes that will have wide use. The more sophisticated > and interactive the palette is, the more difficult it is to learn. Also if > the output is displayed in the palette or widget, then it is also usually > replaced when other choices are made and there is not a permanent record. An example of how a pallet can be useful is finding the name for a particular symbol, and also the keyboard shortcut for entering the symbol. -- "Philosophy is written in this grand book, The Universe. ... But the book cannot be understood unless one first learns to comprehend the language... in which it is written. It is written in the language of mathematics, ...; without which wanders about in a dark labyrinth." The Lion of Gaul