Converting  Mathematics  to  Web-Readable  Form

What is the best way to post mathematical documents on the web? Well, that depends on who your intended audience is, what software you expect that audience to have, how long the document is, how high a quality of typesetting you want the document to have, and how quickly or easily you want the audience to be able to read it.

• To preserve the high quality typesetting provided by TeX, you may want to post your document as A DOWNLOADABLE FILE. For this purpose I usually recommend PDF files, though PS files come in as a close second choice. (PDF files have the advantage that everyone has, or can easily obtain, the required reader program. You can generate PDF files using the Windows Miktex program dvipdfm, or using the Unix programs dvips and ps2pdf, discussed below.) The method for posting downloadable files of any filetype is explained in another on-line documentation page, "Posting Files of Various Filetypes on the Web". (I do NOT recommend posting TEX or DVI programs, because they are not really standardized. Their display depends on style and font files and other files that are part of the TeX installation, which may vary from one computer to another.)

• On the other hand, if you're willing to give up some of the high quality of typesetting, you can post math on a web page that is INSTANTLY READABLE (i.e., no waiting or downloading needed). You can use one of these methods (or a mixture of these methods):

  • Use GIF files (i.e., pictures) -- in effect, you take "snapshots" of your document. These take a little time, but not much, to load into the browser of the person viewing your web page. By a variety of methods (discussed below), you can produce lots of little pictures of individual symbols or individual formulas; or you can produce a few big pictures, one for each page of your document. A minor disadvantage of this method is that the picture files will not be resized when the person viewing the web page changes his or her preferred font sizes. An advantage is that GIF files will work on all graphical browsers, and I think that JPG files probably will too. (For most efficient use of storage space, It is recommended that you use GIF files for diagrams and pictures involving 256 colors or less; use JPG files for photographs and other pictures involving more than 256 colors.) ... Other picture formats, such as PNG, will work with some browsers but not all.

  • Use the symbol font and other special HTML tricks. For instance, on most browsers,

    <font face=symbol>&#242;</font>

    displays an integral sign. This approach has the advantage that it is relatively to edit and it makes the formulas integrate smoothly with the text. It has the disadvantage that different browsers use different tricks of this sort; there is not a uniform method for all browsers. Indeed, the uniformity has diminished in recent years; Netscape 6 and Mozilla do not conform to the same methods followed by Netscape 4 and Internet Explorer. (At this time, I do not recommend using MathML (mathematical markup language) in your web pages. Althoguh it is now a part of the "official" HTML web language, as endorsed by the World Wide Web Consortium, it is not supported adequately by major browsers. Very few members of your intended audience are likely to have the unusual software to read MathML documents.)

How to Use ps2pdf

ps2pdf  filename.ps  filename.pdf

How to Use dvi2gif

1. Move your dvi file into a directory of its own, if you want to avoid mixing together too many different unrelated files in one directory.
2. If your dvi file is named "myfile.dvi", type "dvi2gif myfile" -- don't bother typing the ".dvi" after the "myfile".
3. Wait a few moments while dvi2gif does its stuff. It will create a collection of gif files, one for each page of your tex output. These will be named "myfile-p1.gif", "myfile-p2.gif", etc. It will also create a very basic web page, called "myfile.html." When dvi2gif is done, the computer will prompt you for your next command. Type "ls" or "ls -t" to see what files it has created.
4. The dvi2gif program assumes you wrote your tex files to print on ordinary 8.5 inch by 11 inch paper. If you intended to use a larger page, you're out of luck. If you intended to use a smaller page, you can crop your gif files using any appropriate graphics editor (e.g., Lview). If you don't know how to do that, don't worry about it; it just means that your gif files have larger margins than necessary.
5. The file "myfile.html" is a very minimal web page. It contains nothing more than links to the gif files. This html page is very crude; it is not really intended for you to post publicly. However, it will save you a little work. Before you delete this minimal web page, you can use cut-and-paste to copy its links into some more elaborate web page of your own. Edit the links if your web page is not in the same directory as the gif files.

By the way, dvi2gif was written by Chip Turner, one of our students; it is a script which knits together some complicated instructions worked out earlier by Prof. Ellingham.

How to Use Windows in place of dvi2gif

One drawback to this method is that it is somewhat tedious; it takes many steps. Another drawback is that it only works for one screenful at a time; in some cases you may want to create a larger image that covers 2 or 3 screenfuls. That can be done by pasting together several smaller images, if you know how to use your graphics editor, but again it is rather tedious.

How to Use latex2html

cos x dx = sin x + C

However, you don't have to do them by hand, if your favorite flavor of Tex is Latex (as opposed to Plain Tex, AMS-Tex, or some other variant). You can use the latex2html program. Here is how to use it:

1. If your latex file is named "myfile.tex", type "latex2html myfile.tex".
2. Wait a long time -- or a very long time, if your document is long. You'll see some activity on the screen, while different programs talk to each other and create files for you. It's not done until the computer gives you back your usual prompt -- a percent sign, or whatever is your usual prompt. The program sometimes pauses for several seconds with "GS>" on the screen, but that's not your prompt; you still have to keep waiting when you get that.
3. When the program is done, you'll find that it has created a new subdirectory called "myfile" within the directory that you started in. In that new directory, you'll find a large number of gif files, plus three html files and a bunch of other miscellaneous files. You can delete all those miscellaneous files -- they're just log files, byproducts of the program's work.
4. The main html file is "myfile.html". Try running Netscape on that file, to see how it looks. The latex2html program does fairly well with simple tex, but it won't always work well with complicated macros or sophisticated spacing. You may have to fix those things by hand -- i.e., by editing the html file. Evidently, this technique is only for people who know at least a little html and a little latex.
5. The two other html files are "node1.html" and "node2.html." Those contain information about the latex2html program. It would be courteous to leave those in, but if you really want to you can delete those files, and edit the "myfile.html" file so that it no longer contains links to "node1.html" and "node2.html".

Though the html file described above may be difficult to construct, it should display properly on essentially any version of any graphical browser -- Netscape, Explorer, Mosaic, etc.

The latex2html program does have one serious flaw: it does not convert paragraph breaks in the LaTeX document to paragraph breaks in the HTML document. You can add the paragraph breaks manually in the HTML document, or if you prefer you can just insert the code

We have a recent version of latex2html. Also, we have online documentation about latex2html. You can go to any text terminal on our system, and type "man latex2html", to get a brief text explanation. For a longer manual, go to one of our X-Windows terminals, type

netscape &

file:/usr/local/stow/latex2html/manual/manual.html

ftp://username@atlas.math.vanderbilt.edu/usr/local/stow/latex2html/manual/manual.html

Text-encoded mathematical symbols

For instance, subscripts and superscripts can be displayed with the "sub" and "sup" tags. If you want to display x-sub-3 and x-cubed in an HTML page, you could display them with these commands:

Unusual characters (ASCII/Unicode)

You must use the ascii encoding if you wish any of the characters & < > to appear on your web page. Just typing the characters you see on the keyboard won't work; those keystrokes will be interpreted as special codes by most web browsers. (If this explanation confuses you, just view the source file for this page.)

A non-breaking space is a space that looks just like an ordinary space between two words -- but it will keep those two words glued together, so that a line break cannot occur between them.

Many accented and other foreign-alphabet symbols can be encoded by number; most of them can also be encoded by name. I have put charts of those on a separate page. Many mathematical and other symbols can be encoded by number, but I am not aware of any name abbreviations available for the same characters:

Using these built-in characters instead of small gifs saves some downloading time -- i.e., your intended audience will not have to wait as long for the material to load into his or her viewing screen. Moreover, these built-in characters may make it easier for you to edit a web page: You can do everything in one file, using just one text editor; you don't have to keep track of dozens of separate little files. But be cautioned that these characters are not necessarily exhibited properly on all browsers.

Symbol font

The preceding technique only shows the "low ascii" characters of the "symbol" font. The "symbol" font also includes "high ascii" characters, which correspond to control-characters such as control-I or control-shift-I. Those are hard to edit, using ordinary text editors. However, we can encode those control-characters with ascii too. The result makes a slightly longer and more cumbersome ascii code, but it still occupies far fewer bits than does a graphical image.

  A VU Math web page, updated 27 May 2002 by the department's webmaster. Your comments are welcome.