Changing Firefox' cache directory ================================= Enno Middelberg, 26Apr2006 Under Linux, Firefox's cache directory defaults to ~/.mozilla/..., where it can quickly fill up one's disk allowance. To change it, enter "about:config" in the address bar and look for "browser.cache.disk.parent_directory". If it exists, change it, if not, create it by right-clicking into the "about:config: window and creating a new string with the appropriate value. Piecewise function definitions in Gnuplot ========================================= Enno Middelberg, 1Apr2006 (I'm NOT kidding here!) To define a function in Gnuplot which is only valid for a certain range, use the ternary operator, "?:". It works as follows: condition? operation1: operation2 If the condition is met, execute operation1, otherwise, execute operation2. For example: f(x)= 1986<=x && x<1998 ? m*x+b : 1/0 This function is defined only if 1986<=x<1998, otherwise, it is undefined. See also help ternary from the gnuplot prompt Using Thunderbird as email client from within Opera =================================================== Enno Middelberg, 6Oct2005 One can use Thunderbird as an email client when clicking on "mailto" links in Opera. In Opera, go to Tools -> Preferences -> Advanced -> Programs, and select "Use specific e-mail client". Below, enter $PATH/thunderbird -compose mailto:"%t" Configuring Opera to block advertising ====================================== Enno Middelberg, 5Oct2005 -Add the following line to the "[Adv User Prefs]" section of your opera6.ini: URL Filter File=/u/mid063/.opera/filter.ini -make a text file called "filter.ini" with the following content, and put it into the same directory as the opera6.ini. After that, Opera will block any content matching the patterns below. ; filter.ini ; This file is part of the Opera browser. [prefs] prioritize excludelist=1 [include] * [exclude] http://ad.* http://ads.* http://adserv* http://imgserv.ad* http://*.ads.* http://*/ads/* http://*banner* http://*Banner* http://count* http://par.ad.* http://*linkexchange* http://*.doubleclick.net/* http://*.fastclick.net/* http://*.burstnet.com/* http://*.hitbox.net/* http://*.ivwbox.de/* http://*.tradedoubler.* http://view.atdmt.com/* http://pop.mircx.com/* http://ffxcam.fairfax.com.au/* In KDE (and probably Gnome as well), you can extend popup menus to easily extend the list by modifying the default menu configuration. This will add a menu item to the popup menus when you right-click on unlinked and linked images which are ads and which you want to block: You will normally use the system default menu setup, but you can get a copy and modify it. In Opera, select Tools -> Preferences -> Advanced -> Toolbars, and make a copy of the "Opera Standard" Menu setup. You will then have a file called "standard_menu (1).ini" in the "menu" subdirectory of the directory where you found the opera6.ini file, ususally ~/.opera. Add the following line to the "[Image Link Popup Menu]" and the "[Image Popup Menu]" sections: Item, "Add to block list" = Copy image address & Execute program, "$PATH/operaadblock.sh", "%c" The file "operaadblock.sh" contains ---------------------8<------------------- #!/bin/sh block=`kdialog --title OperaAdBlock --inputbox Block? $*` if [ $? -eq 0 ]; then echo "$block" >> ~/.opera/filter.ini fi ---------------------8<------------------- Right-clicking on an ad and selecting the "Add to block list" item then opens a dialog with the image URL. You can modify it to make it more general and then click ok to add the pattern to the filter.ini Using Thunderbird as email client, but Opera as browser ======================================================= Enno Middelberg, 5Oct2005 I am using Thunderbird 1.5beta for email and news, but Opera for internet browsing. This is how to open links in emails in new tabs in Opera: -add the following lines to Thunderbird's prefs.js (substitute $PATH approprietly): user_pref("network.protocol-handler.app.ftp", "$PATH/opera-thunderbird.sh"); user_pref("network.protocol-handler.app.http", "$PATH/opera-thunderbird.sh"); user_pref("network.protocol-handler.app.https", "$PATH/opers-thunderbird.sh"); -make an executable shell script, "opera-thunderbird.sh", containing the lines #!/bin/bash exec opera -newpage "$@" That's it. Thunderbird seems to ignore command line switches, thus the workaround with the shell script is required. Renaming multiple files ======================= Enno Middelberg, 28Aug2005 If you have to rename multiple files, use the "rename" command. For example, to Examples: rename -r 's/\.JPG/\.jpg/' *.JPG - convert *JPG file extensions to lower case rename -r 's/ //' *.JPG - strip blanks from *JPG files rename -r 'y/A-Z/a-z/' * - convert all files to lower case Consider using the -i (interactive) switch to make sure you don't mess it up. Once you're sure it works, you can hit g for go and the rest will run without questions. Reading files directly from tape with dd ======================================== Enno Middelberg, 22Mar2005 When reading data from tape, most applications determine the block size (the size of a chunk of data written at once) themselves. dd can run into problems when the block size on tape is larger than it expects, resulting in a "dd: reading `/dev/rmt/5ln': Not enough space" error. Use dd's "bs" parameter to increase the block size, until it is large enough to be auto-detected by dd. The block size on Suns is 512 bytes, and on HPs is 1024, and a typical block size is 126 (don't ask me why), so a reasonable starting point would be larger than 126*512=64512 bytes. Try for example dd if=/dev/rmt/5ln of=file_to_write.fits bs=128k That should do it (in this case, for the DDS4 tape). Xemacs says "Minibuffer already active: abort it with '^]', enable new one with 'n'" ==================================================================================== Enno Middelberg, 9Mar2005 Press C-] to bring Xemacs back to its normal behaviour Compiling Miriad tasks ====================== Enno Middelberg, 11Jan2005 -Copy the Fortran code (task.for and task.h) from $MIR to your directory -Edit the code -compilte it with fortran -o output task.for `mirlibs` -make sure the new task comes first in your search path $PATH Installing Gnuplot 4.0.0 ======================== Enno Middelberg, 17Nov2004 -get gnuplot-4.0.0.tar.gz and unpack -run ./configure --prefix=/data/PLASTE_2/mid063/software/gnuplot -run make -run make install -enjoy Burning DVDs on PLASTE ====================== Enno Middelberg, 29Oct2004 My PC Plaste now has a DVD burner (Pioneer DVR-107D). Here's how to get this darn thing to work: A derivate of cdrecord, cdrecord-prodvd, is used to burn a DVD. The program is free for private and educational use, to make it work requires the environment variable CDR_SECURITY to include an expiring key. Please visit http://ftp.berlios.de/pub/cdrecord/ProDVD/ for details. cdrecord-prodvd addresses the DVD burner as dev=ATA:1,1,0, otherwise, it's pretty similar to burning a CD. I have written a shell script, /u/mid063/bin/burndvd.sh, which takes an ISO image as argument and burns this image onto a DVD. If no argument is specified, it will generate an ISO image from the current working directory and subsequently burn it. To blank a DVD-RW, use /data/PLASTE_2/mid063/software/bin/cdrecord-prodvd dev=ATA:1,1,0 blank=fast and to burn an ISO, use /data/PLASTE_2/mid063/software/bin/cdrecord-prodvd dev=ATA:1,1,0 filename.iso Fonts in PDF documents look ugly when using the latex -> dvi -> ps -> pdf conversion chain ========================================================================================== Enno Middelberg, 28Sep2004 -try using dvips -Ppdf filename -This issue is well explained here: http://carol.wins.uva.nl/~portegie/tex2pdf.html Programming loops in tcsh ========================= Enno Middelberg, 27Sep2004 If you want to repeat an operation for each file in a directory, say, convert all JPGs to PNGs, use this: foreach fn ( *.jpg ) # read in the filelist /usr/bin/convert $fn png:$fn.png # convert it to png using ImageMagick's convert end # close the loop For more info, try a Google search on "CSH one liner", you should get plenty of results. Here is one of them: http://www.luv.asn.au/overheads/csh-one-liners.html How to install and configure Tim Huege's "ascore" ================================================= Enno Middelberg, 3Sep04 Tim Huege has written a Perl script which searches the daily mail from astro-ph for keywords. Each keyword can be highlighted in a color specified by the user and adds a score to an article's score sum. It then sorts by the score and generates HTML files from the articles. As an example, take a look at file://DATA/PLASTE_1/mid063/astro-ph/20040519.html All you need is the ascore script and a configuration file. Copy the script from /DATA/PLASTE_2/mid063/software/bin to a directory listed in your $PATH, and for a start, copy the configuration file .ascorerc from my homedir /u/mid063/.ascorerc to yours. Now edit the configuration file.: astro-ph mails need to be automatically moved to a certain "folder" where ascore can find them. From the file system's view, this folder is a file. Specify this file in .ascorerc: mailfile=/path/to/file Specify where the html files should be stored: htmlpath=/path/to/dir/ (don't forget the / at the end) Then edit the score types. It's pretty self-explanatory: Middelberg=s:100,c:orange,i:atnf.jpg This means, my articles get a score of 100, my name will be in orange and there will be an icon added to the article. The other sections of the configuration file are similar. The icons need to exist in the same directory as the HTML files. Execute ascore and it will generate one HTML file for each day's astro-ph email. In the default mode, existing HTML files will not be re-generated each time you call ascore, but you can force that using the -f command line option. Hint: ascore makes a pattern search for the specified expressions. For example, if you specify "AGN=s:100", this will also match "Magnetic". To avoid this behavious, add a space before the expression. Ppgplot says "_ppgplot.typeerror: object is not an array" ========================================================= Enno Middelberg, 21Jul04 There is a trap with ppgplot and the Numeric/numarray packages. Ppgplot can be compiled with only one of these packages, and you then have to stick to that. When compiling, Ppgplot checks whether Numeric or numarray is installed (it reports what it has found, first line after executing "python setup.py install"). If you prefer Numeric, do the follwing: -Edit "setup.py". Comment-OUT the line: from numarray.numarrayext import NumarrayExtension -AND ALSO comment-IN the line (a few lines below): raise ImportError -Rebuild and install (thanks to http://www.astro.ruhr-uni-bochum.de/laboca/download/boa_master_doc/3_2_5ppgplot.html) Using Windows applications from a Unix desktop ============================================== A Windows emulator can be run on Solaris machines in the following way. -log in to venice -launch /usr/local/wincenter/winm This will bring up a Windows login screen. Log in with your NEXUS account and password. To access your homedir, start the file manager, drop down the "Tools" menu and select "map network drive". In the menu, enter "\\lynx\IDENT", where IDENT is, well, basically, your ident. This will bring up a file browser showinh you files in your homedir which you now can access with all the colourful toys Windows offers. If you have ssh-agent running, you can use the following command to go directly to the Windows login screen: ssh -l IDENT venice -s /usr/local/wincenter/winm Setting up ssh ============== -A guide on how to generate a passphrase and the related key can be found at http://acd.ucar.edu/~fredrick/mpark/ssh/rsa-unix.html -modify your .xinitrc file to include eval `ssh-agent ` ssh-add ~/.ssh/id_rsa ~/.ssh/identity You will need to type in the passphrase once upon login and can then log in from machine to machine without typing your password each time. -Generate the file "config" in your .ssh directory and add the line ForwardX11 yes to enable automatic forwarding of X windows. You will be able to use things like Xemacs and AIPS windows without having to set the environment variable "DISPLAY" and without allowing the remote host to connect to your machine with "xhost +MACHINENAME" Exported disks from other machines ================================== by Enno Middelberg, 16Jun2004 Local hard disks from other machines can be found at /DATA/MACHINE-NAME_X where MACHINE-NAME is (guess what!) the machine's name written in capitals, and X is the number of the disk. The automounter will take care of the disks being mounted to your machine and they will be unmounted after an idle time of approximately 15 mins. The first disk of any machine (and sometimes also the second), named MACHINE-NAME_1, is a scratch disk, and its contents will be removed every Monday morning. See the file SCRATCH_DISK.txt in these areas. Tape backups ============ by Enno Middelberg, 16Jun2004 Tapes in the "ATNF computers" room are attached to attila, so ssh to attila and chdir to the directory that you want to backup, say ~username for your homedir or /DATA/MY_MACHINE_3/MYDIR if it is a local drive. Assuming you have inserted a tape into the drive (I use the DAT on /dev/rmt/5ln), do a Mounting and checking the tape mt -f /dev/rmt/5ln rewind to rewind the tape. The device name is written onto the drive. Use mt -f /dev/rmt/5ln status to check the status. The line "file number" tells you how many files have been written or advanced. Making the backup To start the backup, use, eg, tar cvbf 126 /dev/rmt/5ln . > ~/logs/backup.`date +%d.%m.%Y` The "b" option, together with the "126" will set the size N of the tape "blocks" to 126*512 bytes instead of the default of 20*512 bytes. More bytes per block means larger blocks and less gaps between them, so the writing speed is increased a little (see http://www.gnu.org/software/tar/manual/html_node/tar_142.html for details). The blocking factor has to be used when the backup is read, so easiest is to use always the same blocking factor and to write it onto the tape cover. All output from tar, including the list of written files, is piped into the file ~/logs/backup.`date +%d.%m.%Y` for later reference, where `date +%d.%m.%Y` is replaced with the current date in dd.mm.yyyy format. Making another backup on the same tape To write another backup to the same tape without overwriting the existing one, insert the tape into the drive and do a mt -f /dev/rmt/5ln eom Here, "eom" means "end of media", and the tape will be forwarded to the position behind the last written file. You can check how many files you skipped on the tape using mt -f /dev/rmt/5ln status again. Then use the tar command above to write your backup. Note: I now use the command tar cvbfX 126 /dev/rmt/2ln excludes * > ~/logs/dlt-backup.`date +%d.%m.%Y` for backup (on DLT though). The "X" parameter will cause the Sun tar command to look up filenames to exclude from the archive in a file called "excludes" in the directory where tar was called. Use this e.g. to exclude browser cache and temporary directories. **************************************************************************** Old stuff follows, some MPIfR-specific **************************************************************************** Put this in your .Xmodmap to enable german Umlauts and the Euro sign ==================================================================== by Walter Alef, 26Mar2003 -----------------------------------------8<-------------------------------- keycode 64 = Meta_L keycode 0x6D = Multi_key ! ! Peter Müller, MPIfR Bonn 16. Mai 1999 ! ! This file makes the following changes: ! ! ä on ! ö on ! ü on ! ß on ! µ on ! ¤ (EURO) on ! The "Alt Gr" key generates Mode_switch and Multi_key keysym u = u U udiaeresis Udiaeresis keysym o = o O odiaeresis Odiaeresis keysym a = a A adiaeresis Adiaeresis keysym s = s S ssharp degree keysym m = m M mu section keysym e = e E EuroSign EuroSign keysym Alt_R = Mode_switch Multi_key clear Shift clear Lock clear Control clear Mod1 clear Mod2 clear Mod3 clear Mod4 clear Mod5 add Shift = Shift_L Shift_R add Lock = Caps_Lock add Control = Control_L Multi_key add Mod1 = Meta_L add Mod2 = Num_Lock add Mod3 = Mode_switch add Mod4 = Super_L Super_R add Mod5 = Scroll_Lock Hyper_R -----------------------------------------8<-------------------------------- Enabling Umlauts ================ by Enno Middelberg, 26Mar2003 To enable Umlauts by pressing the Windows key, vowel and '"', put this in your .Xmodmap: !! !! Xmodmap: Some Settings for a standard PC key board !! !! Note: the scan codes of the keys (keycode) used here !! are specific for PC key boards. !! !! by werner 1993, 1997, 1998 !! enables use of German Umlauts using the left Windows key keycode 115 = Multi_key How to avoid window splitting on Xemacs startup =============================================== If you start Xemacs, it might come up with the main window split into halves, the lower one containing error messages a la (key-mapping/warning) Put this in your .Xmodmap file to remove it: !! avoid window splitting on Xemacs startup !! due to an error message remove Mod1 = Alt_R Reading Redmond's .doc format under Linux (or: How to install Open Office) ========================================= by R. W. Porcas, 31Mar2003 1) Login to pc192 2) execute /opt/OpenOffice.org/program/soffice install as indicated 3) execute /opt/OpenOffice.org/program/soffice xxxx.doc 4) reject unwanted windows which pop up 5) opt for writing to a .ps file 6) ghostview or print the .ps file Writing cds on PC099 ==================== by Enno Middelberg, 02Apr03 If you simply want to burn some data on a cd using PC099, try this: (">" denotes a shell prompt) -Log in to PC099 > ssh pc099 -If you have a re-writable cd, blank it with cdrecord, otherwise, skip this step > cdrecord-1.6.1 -dev=0,3,0 -blank=fast -Get a script called "burncd.sh" from my home directory; make sure it is in your path and executable > cp /homes/emiddelb/bin/burncd.sh $HOME/bin -Put all files you want to burn in one directory > cp -R $MYSTUFF $FILES_TO_BURN -chdir to that directory and run burncd.sh > cd $FILES_TO_BURN > burncd.sh -you should get some screens full of messages and after some time the writer will spit out a cd with your data. Be aware, it might still be warm :-) -for more instructions on how to burn cds from a shell, see http://forum.mlug.org.au/bbshowpost.php?bbsubject_id=165 http://www.linuxdocs.org/sln/cdcommands/ (this didn't work last time) How to use your Windows-mounted notebook or your home PC as LINUX/UNIX terminal =============================================================================== Bong-Won Sohn, 30Apr2003 Normally one confiures her/his home PC(notebook) as a dualboot(Win/Linux) machine in order to use it as a Linux/Unix terminal. If you need your PC only as a terminal not (as a stand-alone Linux machine) (e.g. in order to log-on one of MPIfR machine or AUIB machine), there is an alternative way. There is a free platform-independent X-terminal emulation program, so-called 'VNC server/client'. Visit http://www.uk.research.att.com/vnc/ for more infomation. Current SUSE Linux versions (8.1 & 8.2 maybe older vesrions too) install VNCserver automatically on your Linux machine. Type 'vncserver' to check, if VNCserver is installed in your Linux/Unix machine. Or ask F. Pascher to install it. If VNCserver is already installed. Now you need to install 'vncclient' in your PC/Notebook. Download from http://www.uk.research.att.com/vnc/ the client program. Consult http://www.uk.research.att.com/vnc/. Don't forget: 1. In order to quit VNCserver, type on your vncserver machine, 'vncserver -kill : #' ; # is the display number which you can see when you begin vncserver. 2. VNCserver is also helpful if you need to use old 8bit color programs like MIDAS or MAPA. You can start VNCserver which 8bit color, then you will have no problem to use these old program which otherwise blame 24bit graphic card. Mathematica complains about missing fonts ========================================= Enno Middelberg, 27May2003 (thanks to T. Huege) During startup, Mathematica may complain about missing fonts that are needed to display formulae. This can be solved by adding the following line to your ~/.Xdefaults file: XMathematica*fontList: fixed After editing the file, you have to restart the X server or to type xrdb -load .Xdefaults for the hange to take effect. Installing a useful spamfilter under Linux ========================================== Enno middelberg, 5Jun2003 If you're sick of spam mails, try the following approach. It works fine for me: I used to receive some 10-20 spam mails per day, this has now dropped to about 1-2 per week that pass the filters. Copy the following files to your homedir: /homes/emiddelb/.spamblock /homes/emiddelb/.spamblock.local /homes/emiddelb/.forward /homes/emiddelb/.procmailrc Edit the files according to your needs: .spamblock ---------- This file contains the filter rules. You can comment out those rules that catch mails that you actually want to receive. Note, that some of the rules are not absolute, ie only if a given mail matches several rules it will be discarded. As an example, if you are an Asian fellow, you may want to exclude the rule that is looking for Asian charactersets. .spamblock.local ---------------- Insert those domains, for which your rules should not apply. That means, even if a mail matches a number of filters, it will nevertheless pass them if you enter the mailer's domain here. .forward -------- This file activates procmail, the program that actually moves the spam into a specified folder. I don't have a clue of how it works, but you need to specify your username here. .procmailrc ----------- This file specifies where the spam will be stored: MAILDIR=$HOME/nsmail - your mail directory LOGFILE=$MAILDIR/.log - the name of the logfile INCLUDERC=$HOME/.spamblock - the folder that takes the spam (this is actually a regular file, it is only treated as a mail "folder") Converting from ssh to ssh2 =========================== Enno Middelberg, 23Sep2003 If you already use ssh to hop from machine to machine, you can switch to ssh2 using the following procedure: cd ~/.ssh ssh-convert -k identity ssh-convert -a authorized_keys Sorting and cleaning up of BibTeX files ======================================= Enno Middelberg, 29Sep2003 (X)Emacs provides an algorithm to sort out only those references from a bibTeX file that are actually being used in the .tex file. Download bib-cite.el from http://savannah.gnu.org/cgi-bin/viewcvs/auctex/auctex/bib-cite.el?rev=HEAD&content-type=text/vnd.viewcvs-markup or contact me for a copy and add the following lines to your .emacs file (I guess bib-cite.el must be in $PATH): ;; Load and enable bib-cite.el package (require 'bib-cite) (autoload 'turn-on-bib-cite "bib-cite") (add-hook 'LaTeX-mode-hook 'turn-on-bib-cite) (add-hook 'bibtex-mode-hook 'turn-on-bib-cite) (setq bib-cite-use-reftex-view-crossref t) You then can use M-x bib-make-bibliography to create a file with all refs needed. If you want to sort out double entries from your bibtex file, get JBibtexManager from http://csb.stanford.edu/nbatada/JBibtexManager/ and start it with, e.g., java -jar ~/local2/bin/JBibtexManager_3.2.2.jar From there on, it's self-explanatory. Big brackets and split in Latex =============================== Enno Middelberg, 6Oct2003 Big brackets will not work if a linebreak is introduced inbetween in a split environment: \begin{equation} \begin{split} a &= \left( b + c \\ & + e + f \right)\\ \end{split} \end{equation} Use \bigl(, \bigr) and relatives instead. Installing ppgplot (pgplot for python) ====================================== Enno Middelberg, 8Oct2003 Check that you have the libcpgplot.a in the pgplot directory. If not, re-install with (see $PGPLOT/install-unix.txt): make make cpg Download ppgplot-1.1.tar.gz, unpack and edit setup.py: -add path to libcpgplot.a: library_dirs.append("/homes/emiddelb/pgplot") -add path to cpgplot.h, eg: includes.append("/homes/emiddelb/pgplot") -add path to python module Numeric, eg: includes.append("/usr/lib/python2.2/site-packages/Numeric") This should be done in the section where setup.py checks for os name="posix": (...) if os.name == "posix": libraries.append("X11") library_dirs.append("/usr/X11R6/lib/") library_dirs.append("/homes/emiddelb/pgplot") includes.append("/homes/emiddelb/pgplot") includes.append("/usr/lib/python2.2/site-packages/Numeric") libraries.append("m") libraries.append("g2c") (...) Then run python setup.py build You'll find the ppgplot.sp in a subdir called "build/lib.linux-i686-2.2" (or similar). Copy it to some meaningful place and be sure to set environment variables accordingly. I use a script called python_with_ppgplot.sh and then run python (or my python progs) from the same shell. #!/bin/bash export PYTHONPATH=$PYTHONPATH:/homes/emiddelb/local2/python export PGPLOT_DIR=/homes/vlbsoft/libl export PGPLOT_FONT=/homes/vlbsoft/libl/grfont.dat export LD_LIBRARY_PATH=/homes/vlbsoft/libl Need to know's about XMGrace! ============================= Uwe Bach, 28Oct2003 1. At the XMGrace homepage one can find a useful user guide http://plasma-gate.weizmann.ac.il/Grace/doc/UsersGuide.html and a tutorial about the first steps to create nice graphs http://plasma-gate.weizmann.ac.il/Grace/doc/Tutorial.html 2. Placing the legend box: Specifying the placement of the graph by entering the coordinates can be painful, especially the fine tuning. To alleviate this problem, a graphical method is also available, although not readily apparent. After a legend appears, it may be dragged to a new location. To do this, press Ctrl-L with your mouse on the main canvas. You should see the arrow cursor turn into a hand. If this doesn't work, double click on the main canvas (to get its attention) and then press Ctrl-L. Click on the legend and drag it. To cancel the legend drag mode (as with all other modes), click on mouse button 3. 3. Adding and placing text or objects (arrows, lines, ...) in the graph: Just click on the menu 'Windows' and select 'drawing objects'. This opens a dialog where you can choose from various objects like text, line, box ... After inserting your object you can by drag and drop to a new location. To do this, press Ctrl-M with your mouse on the main canvas. You should see the arrow cursor turn into a hand. Click on the object and drag it. To cancel the object drag mode (as with all other modes), click on mouse button 3. 4. List of hot-keys: Ctrl A: Autoscale the current graph Ctrl D: Delete an object Ctrl L: Move current graph legend Ctrl M: Move an object Ctrl T: Place timestamp Ctrl U: Refresh hotlinks Ctrl V: Set the viewport with mouse Ctrl Z: Zoom Ctrl Alt L: Draw a line Ctrl Alt B: Draw a box Ctrl Alt E: Draw an ellipse Ctrl Alt T: Write a text string 5. Using different fonts and/or symbols in labels: Grace permits quite complex typesetting on a per string basis. Any string displayed (titles, legends, tick marks,...) may contain special control codes to display subscripts, change fonts within the string etc. For example to label an axis with the apparent velocity $\beta_{\rm app}$ (latex style) you can type "\x b\f{}\s app", where '\x' means change font to Symbol, 'b' is then the greek character beta, '\f{}' return to "normal" font and '\s' means subscript. A complete list of all typesetting control codes can be found here http://plasma-gate.weizmann.ac.il/Grace/doc/UsersGuide.html#ss7.1 6. You can change the default settings of xmgrace (fonts, linewidths, etc) by copying the default file "Default.agr" from, e.g., /usr/X11R6/lib/xmgrace/templates/ to $HOME/.grace/templates and editing it according to your needs. Including multi-page Postscript in Latex documents ================================================== Enno Middelberg, 13Feb2004 I had a multi-page Postscript file that I wanted to include in another dosument using Latex. Here is how it worked for me. 1. Split up the document into single pages using psselect. This can be done for each page using a bash loop: for x in 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21; do psselect -p$x myfile.ps page$x.ps; done The resulting 21 single Postscript files are named page1.ps .... page21.ps 2. Convert each page to epsi using ps2epsi: for x in 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21; do ps2epsi page$x.ps; done Include each epsi file in Latex using \includegraphics. You might get a "Too many unprocessed floats" error. I solved that using the \clearpage command after every fifth page. Removing files older than a certain time ======================================== by Enno Middelberg, 4Mar2004 If you need to remove all files in a directory which are older than a certain time, e.g., 30 days, use the following command: find /path/to/dir -mtime +30 -exec rm '{}' \; Extract from the find man page: TESTS Numeric arguments can be specified as +n for greater than n, -n for less than n, n for exactly n. -exec command ; Execute command; true if 0 status is returned. All following arguments to find are taken to be arguments to the command until an argument consisting of `;' is encountered. The string `{}' is replaced by the current file name being processed everywhere it occurs in the arguments to the command, not just in arguments where it is alone, as in some versions of find. Both of these constructions might need to be escaped (with a `\') or quoted to protect them from expansion by the shell. The command is executed in the starting directory.