Revised on May 10, 2005 to include information on gnumeric documentation. This section also describes documentation using the info reader. A key file related to the use of info has a date of Jul 28, 2000, so Until the next revision, that is probably the previous date of this section.
Documentation of gnumeric 1.0.4 in html format can be found here. This program is available on workstations in some offices, but the documentation wasn't installed. A copy has been found, but its usability has not been tested. Report problems to bumby@math.rutgers.edu.
(the older entries that follow have not been checked on current systems) A frequent question is, "Where is the documentation?" In Unix, the standard answer is to use the man command to read the official documentation, but if you try that with some basic commands, like ls, you are told that the manpage is not maintained, since the official documentation is in info. If you use emacs, you are familiar with info, but you may not realize that it can also be used (and may even be better behaved in some ways) from any command line (such as an xterm). You should be able to follow your unsatisfying man ls with info ls and get the recommended documentation. If you are the GUI type of person who prefers to read manpages in xman, bring your xterm window to the front and move your hands to the keyboard. You can also use emacs, where you enter info using the 'ctrl-h i ' key sequence.
If this doesn't work, you may need to add an INFOPATH variable to your environment. On the new machines, a useful setting is "/usr/local/info:/usr/local/gnu/info:/usr/local/teTeX/info". Most of the documentation is in usr/local/gnu/info, but a useful start of the index file has been added to /usr/local/info. If you start info without any additional arguments, it will build an index from the files named dir in the directories in its path. A large menu in this directory node is useful for browsing the documentation to see what is available. The files needed to build such a menu have be constructed. Any problems should be reported to me.
Footnote (April 19, 2002). Under Solaris 8, e.g., on math, the command-line man program reads documentation in roff and sgml formats from a single branch of the directory tree. This may be most current description of programs supported by Sun (although gnu programs continue to be documented in info). Unfortunately, the xman program is not able to read sgml, so a MANPATH environmental variable is needed to find other locations of the documentation that it can handle. Even if this variable is correctly set, the information it finds may not be current.
This section was prepared in November, 2001. Comments about PostScript level were added on April 22, 2002.
The main printers for general use are lp29 in Hill 323 and lp59 in Hill 531. Both of these printers also have separate queues for special billing methods obtained by appending i (for instruction), a (for administrative), g (for grant), e (for external), or p (for personal) to the name of the printer. For a current list of printers (and print queues) available the network, you can type "printers" at a shell prompt.
These printers are level 2 PostScript printers (for more information about differnet levels of the PostScript language, see the relevant part of the Adobe web site). Some programs, including Acroread 4, can produce level 3 PostScript, but our printers will reject such jobs. In the case of Acroread, the .acrorc file in your home directory has a line beginning *PSLevel: that can be used to set your preferred level. For simple text, level 1 suffices and is often immune from problems that arise from expecting equipment to live up to its advertised capabilities. A frequent problem is that a document will garble fonts after the first page. Disabling the option to "Download fonts once" will often correct this. Postscript level 1 also forces fonts to be downloaded on a page by page basis, so this also avoids this problem.
Here is a link to a longer description of department printers.
This section was prepared sometime in the twentieth century, but is accurate as of April 19, 2002.
If you want to use emacs as an external editor for a mail or news program, but don't want a fresh version of emacs to be started for each use, you should modify your usual emacs to allow it to operate as a server and have the other program call "emacsclient" as its external editor. Details are in the "Emacs Server" node of emacs.info.
This section was composed on May 16, 2002.
Unix has all the file management utilities you need available from a shell prompt. At the other extreme, window environments (X or otherwise) have file managers that allow "drag and drop" operations for moving or copying files. If you always have an emacs window open on desktop, you have a utility that combines the best features of these two systems. It is called dired, and is documented in the info files for emacs. It is usually bound to the key sequence "ctrl-x d" (quotes only used to delimit the sequence; they are not part of the key sequence). You will be prompted for the directory that you want to show in a buffer. Unless you custom it otherwise, you get the output of 'ls -la' in an emacs buffer. This shows all files, including those that are usually hidden, with the mode, owner, size and name of the file. If your path is set to use the gnu version of ls (in most cases, this is not the current default) or if you have customized the command, you will also see the group that the file belongs to. Non-destructive commands are usually immediate, but deleting files requires marking the file (with d) and expunging it (with x). Typically, many files are marked, and then they are all removed with a single x command.
Additional operations in dired buffers are enabled by dired-x. If you want this, you need to add instructions to your .emacs file to load support. This is documented in an info file, but that file is outside of the emacs info tree. It will be listed in the main info directory, which should appear when you first enter info. (At present, this documentation is not installed on math. This oversight has been reported.)
Updated December 19, 2002 to include suggestions on erasing files. Previously updated June 07, 2002 to include section on plug-ins.
You are supposed to set options in Netscape using the menu bar: Edit->Preferences. You should try this first. The lines that need to be changed are at the bottom of some secondary items on the list: be sure to click on the triangle to expand the list to find the items in "Identity", "Mail Servers", "Newsgroup Servers" under "Mail and Newsgroups"; and "Cache" under "Advanced". Unfortunately, some very important things appear not to be touched by this system. For example, when the location of my home directory was moved, I found that I lost access to my "Bookmarks", and I could find no way to tell the system where it was. The easiest way to correct this problem is to find the file ~/.netscape/preferences.js and edit it manually. You can safely ignore the warning, "This is a generated file! Do not edit." as long as you limit your action to changing the name of the server in the absolute pathname to your home directory. The use of absolute paths here with no provision for changing them is annoying, but it is not a serious problem.
Someone has reported that just editing preferences.js didn't work since those changes were overwritten. It appears that this information is cached at the system level, not just within Netscape, so you should log out after making these changes. Exiting Netscape does not appear to be sufficient, although that should certainly be done before trying to edit these files by hand.
(December 19, 2002) Another approach to Netscape preferences is to simply delete troublesome configuration files in the (dot)netscape directory in your home directory. They will be rebuilt. You will lose any customization that you made previously, and you will need to accept the license again, but it may be a net time-saver. With Netscape 4.7x (the standard on most of our machines), this is the preferred way to deal with a bloated cache when you need to recover some disk space. When I tried the utility to clear the cache from the preferences item in the Edit menu, it removed the preferences file but did not clear the cache.
Updated to include the MIME type for Maple worksheets on December 19, 2002. Further comments dealing with more recent versions of Netscape or Mozilla added April 21, 2004. Still more about mozilla added September 9, 2004.
(April 21, 2004: These comments apply to Netscape 4.7x; later versions will be discussed below). In the "Navigator/Applications" section of the preferences menu, you will need to implement programs for some mime types that we use all the time, but are not common in the rest of the Netscape universe. Here are my settings:
| Description | Mime Type | Suffix | Application |
|---|---|---|---|
| teX DVI data | application/x-dvi | dvi | xdvi %s |
| Portable Document Format | application/pdf | acroread -tempFile %s | |
| Postscript Document | application/postscript | ps,ai,eps | ghostview %s |
| Maple Worksheet | application/x-maple | mws | /usr/local/bin/xmaple %s |
In all cases, the "Application" button needs to be selected before entering the name of the program in the box on that line. You may need to use the full path with these programs, but a properly set PATH should allow the short form to work. You do need the "%s" to assure that the program is given the temporary file that Netscape downloads when you select a link. The option for acroread is set to avoid adding these useless names to the menu of recent files. (December 19, 2002) The ability to launch Maple is new. Since it depends on the web server defining this MIME type, it may not work generally, but this type has been defined on our server so this will work for files on the Rutgers math site. Students will probably find it more convenient to add this MIME type with the Save to Disk action.
In the case of acroread, that name will launch the most recent version, but older versions can be used, if necessary, by modifying the program name. In particular, on math and fermat, acroexch is the name of release 3 of the Acrobat Exchange. This is more than just a reader: it allows modifications to be introduced into pdf files. There is also release 4 of the reader with the name acroread4.
(Added April 21, 2004.) Some workstations have later browsers. For example, the computer on my desk, running RedHat Linux 7.3, has Mozilla 0.9.9. The relevant menu is called "Helper Applications" under "Navigator". There is no longer a complete list of default behavior for all known MIME types. You need to use the "New Type" button to add types that are not already known (and properly handled). I have added types for application/x-maple and application/x-maple-xm , since these are the types defined by our server for Maple worksheet files.
On math and fermat, Netscape 4.75 has plug-ins installed for Flash and Acroread5. In the case of Acroread, you can use the "Navigator/Applications" section of the preferences menu (described above to switch between the plug-in and a separate program. Each has its own advantages. The plug-in allows easier navigation between html and pdf files and use of forms; the separate Acroread window gives access to a more complete menu bar including reports on Document Properties.
If you use Netscape for news, there are some things that you need to be careful about. There are two separate news servers at Rutgers, so if you have been using one and try to use the other, the message numbers may be inconsistent, so you the new system will not know which messages you have seen. Netscape does not seem to offer a simple way to customize its news handling. You may be presented with all available groups when you want only to have graphical access to one particular group. The use of a single .newsrc file can allow Netscape to make return to your usual newsreader difficult. Finally, Netscape uses a special cache, called the "xover-cache" for news. You are not offered an option to limit its size in the Preferences dialog, and it can grow to an unreasonable size. You may need to bring your own broom to this directory if you allow Netscape to create it.
(April 21, 2004) This section is probably obsolete: I have not found news to work properly for several years. Various ways to read news were available, but there were authentication problems that made it impossible to post a reply. This limitation, combined with unpleasant side effects of lurking, as described above, led me to abandon news.
(September 9, 2004) Mozilla-based browsers allow you to select the helper application for a MIME type, but not to pass options to that application. In most cases, the default behavior seems reasonable: a temporary file is downloaded and the application started on that file while environmental variables are set to allow the URI of the web page to be known to the application (which is necessary if the file is a pdf file containing local hyperlinks).
April 21, 2004: The current way to check your quota is to use Webtools, although "quota -v" still works.
June 26, 2002: A low quota generates reminders to review material that you are saving in you home directory. If you really need to save that much, you shuld ask for a larger quota. This may require moving your directory. Then you will need to teach Netscape how to find important files.
My workstation now gives a meaningful response when "quota -v" is typed in a console or xterm window. Terminal windows on other machines also give the same response. In addition to your usage in the normal home of my filesystem, I see an entry for /math/u2 that appears to provide some redundancy to improve system reliability.
The following was needed in 2001. Although it is not currently needed, it is kept available in case the need for a similar trick arises.
If you keep saving files, you will soon exhaust your quota and bad things will start to happen. The usual way to check quota is to say "quota -v" at a shell prompt. This only works if the shell is on a machine that can access the quota information. If you usually work on elsewhere (e.g., on your own workstation), this gives no useful information. The simplest way to examine your quota is to force the quota program to execute on one of the servers by changing your request to "rsh lagrange quota -v". This works for me, so I have added a line to my .login file that says:
alias q rsh lagrange quota -v
This allows me to check my quota using the simple command "q" from any shell. Feel free to use this, replacing "q" with your choice of name, and "lagrange" with the name of any server that gives reliable information.
This section describes the original installation of teTeX 1.0, with modifications to support a local texmf directory. It was modified when support for revTeX4 was installed on Feb. 2, 2002. A further modification was on July 25, 2003 to give status of TeX on various machines. The most recent modification was a description of pdftex, added on October 10, 2003. The link to the thesis style file was added on Feb. 24, 2006
With the acquisition of new servers and workstations running Solaris 7, our old TeX installation will be replaced by teTeX 1.0. If you have been using plain TeX, you should remove all commands in your .login and other startup files that set environmental variables that have TEX in their name. The new system is able to find everything it needs without these variables and any setting will force the programs to look in the wrong place. There is a lot of documentation in the form of manpages, info files and other formats. The first two are found by the usual systems that present these forms of help. The texdoc command can find documentation in other formats.
New versions of TeX rely on configuration files to set parameters for all programs in the "TeX and friends" suite. Most of the difficulties appear in printing or building PostScript files for conversion to the PDF format for posting, controlled by the settings in the config.ps file. Without options, the dvips command should print to your $PRINTER. If you want to construct a PostScript file for the Adobe Distiller, the "-P pdf" option should give correct settings. New workstations and servers should have correct settings (originally, the TeX installation in the machine named "math" was a test installation that was not upgraded, so other machines gave better results, but it has had a stable version of TeX for some time). Individual linux machines with RedHat 7.3 or newer also have teTeX 1.0. You can modify settings with a .dvipsrc file in your home directory, or with files in your personal texmf subdirectory tree (users on the Rutgers math network can examine ~bumby/texmf/ for hints), or even in the current directory when you are running the program. While powerful and flexible, this also allows to set things up so that the same instruction will behave differently when small changes are made in the runtime environment. To assure uniformity, a local texmf tree has been created on /math/s2 and individual teTeX installations configured to use it. Difficulties with system settings should be reported so that they can be corrected in this directory. The many steps to get PDF output starting with TeX can be confusing. In particular, if dvips is run without the "-Ppdf" option, the final result is not pleasant to view because Type 3 bitmapped fonts are used rather than Type 1 PostScript fonts. In most cases, it is better to use pdftex to go from TeX source to pdf in a single step instead of using separate programs to get dvi and ps files along the way. The same viewer (typically acroread) can be used for the working copies produced while composing and for the finished product. There is documentation that you can read using "texdoc pdftexman" (there is a separate section of this page describing texdoc) that will be of some help if you want to use the extensions of TeX to produce full-featured pdf files. (The pdfTeX documentation can also be found using this direct link.) In particular, inclusion of figures is done differently since it is not possible to include "Encapsulated PostScript", which is the most convenient format in the tradition method of producing pdf files form PS files. The TeX distribution includes a script called "epstopdf" in its bin directory that will convert figures in eps format into a one page pdf file that can be included by pdftex. LaTeX makes it easy to switch between different conventions for including graphics, but it is not difficult to construct definitions for including graphics in plain pdftex.
There is an ruthesis.cls file for production of a Rutgers thesis in LaTeX. This is used by declaring this class in the header of a LaTeX file. The settings for loading files in tetex will take of the rest in any machine on the math system. A copy of the file is available for downloading to other systems.
Checked for accuracy, with minor updates added, on April 19, 2002, and again on July 25, 2003.
Newer versions of emacs have more useful menu bars at the top of the screen. The new servers and workstations are using version 20.3.3. (Version 20.7.2 is currently installed on Solaris 8 machines. RedHat linux 7.3 includes version 21.2.1 with support for an Xtoolkit and Xaw3d scrollbars.) These instructions are intended for emacs20 and newer and may not work under version 19.
When you are editing a TeX file, there will be a "TeX" entry on the menu bar that provides access to utilities to run TeX on the current file and view or print the corresponding dvi file. Running TeX (or LaTeX) should work with no further action. The system (usually) already knows (and reports on the mode line) whether your file is plain TeX or LaTeX, and will run either "tex" or "latex" accordingly. Our system is set up to find the correct program from this description. If you have problems, you may have something in your .emacs file left over from a previous system. If you find things with TeX in their name, you should probably disable them. A semicolon at the start of a line causes it to be treated as a comment, so adding some semicolons is the easiest way to have portions of your .emacs file ignored on future readings.
Viewing and printing are a different story. Although the menu items exist they may not point to the right programs. I use "xdvi" to view dvi files and "dvips" to print. As presently configured, dvips (with no further options) pipes its output to lpr, so it will be printed on your default printer. To tell emacs to use these programs, you should customize your emacs. The help menu begins with a submenu that gets you into this subsystem. If you start at the top level, drill down through Emacs, Wp, TeX, View to get to menu you want. You will then be able to change the DviPrintCommand and DviViewCommand. When you have entered the values, go back to the top of the page and click on "set" to have the value available in the current session and "save" to make it available in future sessions (this is done by modifying your .emacs file, which you could do manually, but it is better to let the customizer make all such modifications). There is also space for an Alternate Print Command. I have this set to "dvips -Ppdf", which generates a PostScript file (instead of printing) with those features enabled that give best results when distilled to pdf.
If you run TeX outside of emacs and want to use the "e" response to errors to start emacs (or emacsclient) to edit the file, see the "Editor invocation" info page in Web2c (look in the index for web2c.info since the route from the top node is not obvious).
This section describes the implementation of Texdoc in teTeX 1.0, as installed in February 2000. Additional information on the use of this system was added on June 2, 2005.
If you want to find out more about changes introduced in AMS-LaTeX version 2.0, you can scan the documentation using the following commands
These files are in different directories, but texdoc will do the searching for you. It will also recognize the type of documentation available and start an appropriate viewer.
(added on June 2, 2005) The texdoc “program” is actually a script. There is no separate configuration file – the different programs used to do the work are assigned to environmental variables in the text of the script. You may prefer (or, possibly, need) different programs. You can either define these variables in the shell from which you call texdoc or modify the script itself. If you are unwilling (or unable) to modify the script on your workstation, you can make a copy in your home directory and modify that. It is only slightly more difficult to run such a script than one installed in a directory in your path.
(added on June 2, 2005) There is a related program called “texdoctk” that is part of newer tetex distributions. This uses Tk widgets to provide a more flexible user interface. Unfortunately, it is unlikely to work completely. It relies on separate packages that were once common, but are not always installed in newer systems. A new TeX installation is planned for Summer 2005 that will include a working texdoctk.
Spring 2002: nfermat has become fermat and nlagrange is being tested as a replacement for lagrange. There is also some progress in upgrading Solaris 7 machines to Solaris 8. A separate page will record information about the configuration of individual servers.
As of November, 2000, replacement of servers is progressing slowly, but a machine named nfermat is available to allow the department to preview the new system. Connect to it using the 'x' or 'rlogin' command. If you need a newer version of software than is available on our old network look for it on nfermat. If it isn't there, be sure to let us know that you want it available.
Another new machine is compute. Various packages like Macaulay2, Maple7, Matlab, gp (pari) are installed there. Upgrades of compilers are also likely to be installed on this machine first.
June 2002: Acroread5 has been added to some machines, and the ability to use it as a Netscape Plug-in. It is also available on eden and rci. There may be local variation in which version is run under the name "acroread" and whether the default Netscape behavior is the program or the plug-in.
If you use acroread as a plug-in, you will be able to navigate links in the pdf file with your browser. Unless a BaseURL has been set in the pdf file, this will not work if you start a separate acroread from your browser (since acroread will be reading a downloaded copy of the file from a temporary directory). If you are interested in information about the file, you may need a separate instance of the reader in order to access all options in the acroread menus.
Spring 2002: Acroread was upgraded to release 4 on some machines. This is better behaved, by not giving an error message every time it starts, as well as being more capable that release 3. Release 3 of the Acrobat Exchange continues to be available on servers for working with pdf files.
If you are using an NCD terminal and attempt to read a pdf file on math or another machine running Acroread4, you will be presented with a license agreement that appears not to allow any way to accept it. Eric Luhrs reports that entering TAB, then ENTER from the keyboard will allow you to accept the license. If this doesn't work, find someone who will allow you to login through a workstation since only the NCD terminals appear to have this problem. After accepting the license, you will be able to freeely use Acroread4, Acroread 3 and the Acrobat Exchange (also version 3).
This section was started on June 24, 2002 with the following statement. The "Course Materials" part of the department web site has just been made easier to use. A new group called "webpub" was created that includes everyone who is active in creating course web pages. Directories in this area are now owned by this group and have the setgid flag set so that all new files will be owned by this group. Files that are set up to be group writable will stay writable by this group.
Added September 03, 2002. The above paragraph is intentionally vague about to find the "Course Materials" part of the department web site. This description suffices for reading the files, and only those authorized to write to that area need to know how to find the directory on the department network. The department computer committee (R. Bumby, P. Feehan, S. Hartke, C. Sims at the time this was written) can advise members of "webpub" on use of ths area.
New information added January 21, 2004. For security reasons, the disk containing the web site is configured to be mounted read-only over the department network. This means that extra effort must be made to create web pages in this area.
This section introduced June 30, 2002.
SSH has been installed on math, and important Rutgers machines like rci and eden have been configured to accept ssh connections with X11 forwarding. It is intended to replace rlogin and rsh, adding secure encrypted communications between the hosts and allowing X11 connections to be forwarded over the secure channel. On your first ssh connection to a particular host, a public key pair will be set up, with one of the pair stored on the host and the other one in your home directory. This pair will be used to authenticate future connections.
The manpage seems to imply that X11 forwarding is automatic, but it must be set on the command line of in a config file. Once this is done, your DISPLAY variable will be set to appear to be on the remote machine, but output will be encrypted and forwarded to your terminal.
All other methods of connecting to remote machines can be expected to be replaced by ssh.
To find out more, read the ssh manpage on math.
This section introduced November 27, 2002, with minor revision on April 16, 2003.
By popular demand, the Rutgers Faculty of Arts and Sciences has implemented a system that allows grades for exams and other components of courses to be posted on a secure site that allows students to retrieve their individual grades. Although the program is still being improved, it is already easy to use and effective for its purpose.
This site deals only with current courses. The documentation of the site says, "The grades are available immediately after entry and will remain in the system until the beginning of the following major semester (i.e. Fall or Spring). On the day prior to the beginning of a new semester, all course, assignment/exam and grade information will be cleared from the system." With this limitation, it is extremely flexible. Although a course is entered using the "School (2 digits), Subject (3 digits), Course (3 digits), Section (6 alpha-numeric digits), Description (up to 50 chars)" format, this is not checked against the official schedule. In particular, the Section field could be used to identify a single lecture section by giving the meeting time of the course (I have done this). You can also create a fictitious course for testing.
Grades may be entered individually or by pasting data from a spreadsheet, emacs buffer, or anything else that can be copied to your system clipboard. I have successfully copied grades for more than 70 students from an emacs buffer containing data exported from a spreadsheet at a different site. The data should consist of lines with two or three tab-separated fields: student-ID; grade; comment. The full student ID number written as 9 digits is used, and the comment is optional. The Add Grades mode is primarily used to add grades to a newly created assignment, but it also allows an additional list of grades to be added to an existing assignment. However, you will normally use Edit Grades to update single entries from a complete list of grades.
(added April 16, 2003) Newer versions of the program allow bulk entry , but the ability to paste from an emacs buffer has been adequate for my needs. If you post a link to this site, it is helpful to add TARGET="gradebook" to the anchor tag. This will give a new browser window that the user can close by selecting Exit program from the gradebook menu (if you select that from your only browser window, you exit your browser). Since this isn't implemented in the server, it is a matter of good citizenship to make our references to this site more user-friendly.
Students retrieve information by connecting to the same secure site. A student is shown all information associated with his or her student ID.
When you connect to the site, you must change a listbox to "faculty" before logging in, since there is only one portal to this site and the options are set for student use. There is adequate help on the site, so you can learn more by visiting the site.
This section introduced December 20, 2002. Modified on September 9, 2004 (at end of section).
Most of the world now uses attachments to email messages that look mysterious to the many members of this department who consider emacs RMAIL synonymous with electronic mail. In RMAIL you see many equal length lines of printable characters instead of a graphic image, music file, or Maple worksheet that other mail programs generate from these symbols. It takes a little more work, but you can join this multimedia world without leaving your favorite mail program.
One program is called mimencode. It is installed on math and probably other servers. With options it will handle both encoding and decoding. There does not appear to be any documentation, but you can get a synopsis by entering "mimencode --help" at the shell prompt. This synopsis will tell you that the program is called mmencode, but it no longer responds to that name. To use this program you need the "-u" option for decoding and the "-o" option with a filename to say where to put the output. The last argument is the name of the input file. The input should be stripped to only the encoded lines. Usually, all but the last will be of the same length. With this preparation, the program worked on math and gave a usable version of the attachment.
A more user-friendly program is munpack, that has a matching mpack program to do the coding. There are manpages for these programs that you should consult for details. This program reads the entire mail message, including headers, and reads the destination from the header. This has been tested on math and it works as advertised.
If you use Netscape for mail, consult the Netscape section of this file for information about setting preferences for handling different types of attachments.
Pine may need more care. If you use Netscape for anything, it will think that it owns the (dot)mailcap (for assigning programs to types declared as part of the attachment) and (dot)mime.types (for assigning types to file name extensions when attaching a file), so anything that you add manually can get corrupted if you edit your Netscape preferences. I have created separate pine.mailcap and pine.mime.types files and used the pine setup to tell pine to use these files instead of the default files. This has allowed me to use "antiword" for instant access to Microsoft Word attachments (this program is not on the system, but it is small and I can provide you with what you need to put it in your personal bin directory). Any other file that has been properly identified by the sender can be opened in this way. Unfortunately, non-standard types (like Maple worksheets) are likely to be classified as APPLICATION/OCTET-STREAM, which is not specific enough to be useful.
(September 9, 2004) Since most mail programs now handle attachments automatically, separate programs for dealing with base64 encoding no longer seem to have much priority. Programs mentioned above may not be installed on the machines you use regularly. However, current versions of uuencode and uudecode include the ability to work with base64 encoding and uudecode can recognize the start of the encoded segment, allowing it to work on the message with attachment without any prior editing.
This section introduced July 11, 2003. Additional material added September 9, 2004.
Many users now access the math network from a PC running linux. Most software is run directly on the PC, but the X-Window system allows the output of programs run elsewhere on the network to be directed to the display of your PC. To create a consistent environment and to allow you to access your files anywhere on the network, your home directory from math is used as home directory on all machines. Some important software is not comfortable with this arrangement. This section will describe the difficulty and offer suggestions for recovering from problems.
The gnome desktop environment is standard on our linux PCs. This provides a window manager for X11 and other features like a "panel" and "start menu" familiar to users of another well known windowing system. A similar system called KDE is also available. Both of these save your preferences in your home directory. In principle, this means that any machine that you use will have the same appearance. Unfortunately, different machines may have different releases of the operating system with different levels of gnome or KDE. Earlier versions are sure to be confused by settings designed for later ones and could corrupt the directory in which these settings are saved.
If you are happy with the standard settings, you have a simple remedy if you find the system behaving strangely. You can remove the ~/.gnome (or ~/.kde) directory. It will be rebuilt using system defaults the next time you use gnome (or KDE).
(September 9, 2004) The Dell machines introduced in the summer of 2004 use a version of Debian linux that includes window managers with fewer features but better performance that those described above. The standard window manger on these systems is Fluxbox. Another window manger that is available is IceWM. Both can be configured using files in a configuration subdirectory of your home directory. To find out more about configuring these window managers, start with the manpage. It is possible to switch window managers from the main menu, so you can experiment. If you want to make a permanent change, send mail to help@math.
This section introduced July 25, 2003.
The Adobe Portable Document Format has become the standard for many uses. If you use a PC with a recent flavor of Microsoft Windows, or a Mac, you can buy the recently released Acrobat 6 from Adobe for your machine and create files in the PDF 1.5 format. The department currently has Acrobat 5 (which produces the 1.4 format) on some administrative machines.
For *nix, the most recent version of Acrobat is Acrobat3, although version 5 of the free Acrobat reader is available and has been installed throughout our network. We have several licenses for Acrobat3. Its main components are "distill" and "acroexch". On "math", you can use "distill" to convert PostScript files to PDF, and "acroexch" (an abbreviation for "Acrobat Exchange") to modify properties of existing PDF files. Both programs are well documented.
Ghostscript also allows conversion of Postscript to PDF. The interface uses a script called "ps2pdf". Prior to version 6.5 of ghostscript, it was not possible to include Type 1 outline fonts in PDF files produced in this way, so the results were not pleasant to view on your computer monitor (although they gave reasonable results when printed). Since older versions of ghostscript are on our servers, examples of this output can sometimes be found on web pages. RedHat 7.3 includes ghostscript 6.52, so it should produce reasonable results if the PostScript file contains all fonts in Type 1 format. This program can produce files compatible with either PDF 1.2, 1.3 or 1.4.
TeX produces files in DVI format using font metrics, but no other information about fonts. This format contains information for finding fonts when you want to view or print the file. The program "dvips" is usual way to produce a PostScript file. It relies on a configuration file to locate the desired version of each font. Unfortunately, the creator of teTeX had the belief that the fonts created by metafont print better on most printers, so they are used unless a special request for Type 1 fonts is made. As noted in another section of this file, you make this request by saying "dvips -Ppdf".
An alternative to "distill"or "ps2pdf" is "pstill". This is a free program but it is not currently installed on our network. I believe that some members of the department have used it, but I have not tested it. Another approach uses "dvipdfm", which converts a dvi file directly to PDF without creating a PostScript file first. It allows instructions to be put in your TeX file for requesting features of the PDF1.2 format. Since Fall 2001, course web pages linked to my home page, have included material created with dvipdfm.
The future belongs to pdfTeX, but major distributions contain only early versions. It is not currently supported on the math network although it may be available in some linux distributions. This program builds PDF files directly from the TeX source file using extra macros (instead of specials) to describe key pdf objects.
The NSF FastLane has made PDF its standard format. It claims to be able to work with dvi files that can be converted to PDF on their system. Except for using "distill" or a recent version of "ps2pdf" after "dvips -Ppdf", the methods for producing PDF files discussed here do not lead to acceptable files, so the use of dvi files is encouraged. Formats more recent than PDF 1.2 are not necessary, but having the correct font embedded and allowing concatenation of files from separate sources are essential.
This section introduced April 21, 2004.
Maple allows figures or entire worksheets to be exported in various formats.
Exporting figures is not completely satisfactory since there appears to be no way to change the settings of the figure to be exported. A more robust approach is to redirect output using plotoptions. I have done this, but don't have easy access to details at this point, so I will fill in this information later.
Whole worksheets may be prepared for viewing outside of Maple by printing, either to paper or to a PostScript file. Two other formats are also available as exports of the worksheet: LaTeX and html. All export options can be found under "File" on the menu bar. Both of these formats convert the outline structure of the worksheet into a something suitable for the new format.
The LaTeX version will be arranged into section and subsections with the usual LaTeX conventions for headings. To process a file exported to LaTeX, you need access to several style files that are part of the Maple distribution. They can be found in the etc subdirectory of the Maple directory. The current setup works best if these files should be copied to the directory where you saved the file created when you exported the worksheet. You can then make changes to match your LaTeX style. These files for Maple 8 and Maple 9 are almost identical: only the copyright dates were changed. More information on the use of these files can be found in a file called "latex.txt" in the main Maple directory. Some tuning may be necessary. The commands loading the Maple style files may need editing if you want options other than the standard ones. I also found that long lines of Maple output were not broken, even if they were much larger than the page, so some output may need hand editing.
Exporting to html uses frames: one for the worksheet content and one for an index generated from the outline. The output is rendered as gifs which are stored in an images subdirectory. Maple can also produce the output as MathML, which will be more desirable in the future. However browsers on department computers are not set up to display this format.
A revision to describe the availability of photos added June 29, 2006. This section revised January 18, 2006 to describe html format of rosters. Originally introduced September 9, 2004.
The university allows authorized instructors (and most of us are authorized before the semester begins) to download rosters. Unfortunately, the instructions emphasize the use of programs other than those used by members of this department, and the description of the format of the rosters, while good enough for most people, is not precise enough for us. Previously, the files described as Excel worksheets were really another format that Excel is able to import automatically. The correct name of the format is SYLK. When you clicked on the link to download the file, you saw the SYLK file, and it looked like fairly healthy ASCII with some markup tags. The gnumeric spreadsheet could import this format: older versions required that you identify the file as SYLK, but newer versions recognized this format automatically. Unfortunately, when the roster tool was modified at the end of 2005, a different format was introduced.
In January 2006, I could get rosters by asking for a file download, but no option to choose the name of the downloaded file was provided. Once I had the file, I could open it in gnumeric, using the Open entry in the File menu. Find the file that you want and then open the list on the File format line, Select the entry that reads HTML. The roster will be created in a new worksheet. To be useful, you will could use Save As from the File menu to save the worksheet in gnumeric format. Alternatively, you can copy the page to an existing gnumeric workbook. Then you can start formatting it. Mostly, all you need to do is adjust the width of the columns to allow you to see the contents of the cells.
Previously, the student ID numbers were formatted as numbers and leading zeros were not shown. I haven't checked whether this is still the case. If you meet this problem, it can be corrected by using the Custom option of the Number page of the Format cells notebook. A format of nine zeros will show all nine digits including leading zeros. There does not appear to be a way of creating a format that looks like the usual way that these numbers are written, although you should be able to force this by converting this value to a string and using some of the string functions.
Another feature has been added to the roster site: photos of your students are now available, packed 6 to a line in pdf format. I found it while downloading rosters for my summer course. Three of the thirteen pictures were missing, but it provided a good start on recognizing my students.
This section introduced October 29, 2004. Minor modification introduced January 18, 2006. However, information about availability of Office suites may be stale since it was not checked when the modification was added.
The popularity of Microsoft Office programs, e.g., Word and Excel, leads to the expectation that everyone can read files in these formats easily. Users of our *nix systems can work with these files, but it is not always easy.
One general method is to use Star Office. Version 5.2 of these programs is installed on "compute". The program name is "soffice". The first time you use this program, it will build a profile in your home directory. Once this is set up, the program will start directly.
A similar program is Open Office. This is based on Star Office, but has a less restrictive license. Version 1.0 is installed on "matrix". The program name is "ooffice".
Other programs that may be found on linux workstations are:
(January 2006) If you are using Windows but not Microsoft Office, the free WordPad program does a good job of reading a typical Word file, and rendering it as plain text. From there, you can add any formatting information in TeX that you want.
Introduced January 22, 2007. Mathematica is available on the math network by connecting (via ssh with X11Forwarding on) to "matrix". The current version is 5.2. Someone who had used the program in the past reported that it was no longer recognizing anything in the Notebook window. The system also did not respond when asked to play the tutorial (that window, selected from the main window appeared but didn't respond to the button to play the demonstration). After some study where different users on the same machine got different results, it was discovered that it is necessary to have NumLock OFF when using the graphical interface of Mathematica (the text interface always works). It is not necessary to change this setting before starting the program, the NumLock key toggles the behavior of the program as well as the behavior of the keyboard.
This section was created on January 30, 2007.
The W3C -- the group in charge of setting standards for the Web -- currently recommends XHTML as a replacement for the HTML that has been used as a markup language for web pages. There are some simple changes to replace strict a HTML 4.01 page by an equivalent XHTML 1.0 page. Unfortunately, Microsoft Internet Explorer refuses to recognize the type used to describe such documents (see a discussion in answers.com for more information).One fix is described in a letter to a W3C mailing list, but this must be done by everyone who wants to read these documents in Internet Explorer, and must be done by editing the Windows Registry. Also, this fix only allows display as HTML, so the whole reason for using XHTML is lost.
An important motivation for mathematicians to move to XHTML is to facilitate the inclusion of MathML in web pages. Maple is able to export its worksheets in MathML, but I have not yet run a successful experiment to export a file and view it unless the equations were exported as images in gif format.