iraf.net

>From rajat@samba.tifrbng.res.in Thu Aug 14 12:32:41 1997
From: Rajat Varma <rajat@ncbs.tifrbng.res.in>
To: Mike Fitzpatrick <fitz@noao.edu>
Subject: some trivial problemsDear Mike,
writing to you after a long time! I have just acquired a HP
deskjet printer, and I have been just configuring everything so that I
can print images and graphics from iraf. I have basically two problems.a) When printing multiple graphs on the same graph, how can I asign
different colors to different lines on the graph.b) When printing an image from ximtool, which is in size smaller than
512x512, how do I print just the image and not the surrrounding black
region, it wastes my toner.rest is all fine.
rajat>From fitz Thu Aug 14 12:50:37 1997
To: rajat@ncbs.tifrbng.res.in
Subject: Re: some trivial problemsHi Rajat,
Good to hear from you again.> a) When printing multiple graphs on the same graph, how can I asign
> different colors to different lines on the graph.
This depends on the task and your iraf version, e.g. the GRAPH
task can assign colors but some other tasks cannot. Even though you
can get colors on the screen this information is lost when making hardcopy
because of the limitations of the SGI kernel. If you were to use the
PostScript kernel in STSDAS (if you have it installed) you could retain
the color information when printing, note this requires some graphcap
changes to be enabled.> b) When printing an image from ximtool, which is in size smaller than
> 512x512, how do I print just the image and not the surrrounding black
> region, it wastes my toner.
Probably you just want to switch the fore/background colors, e.g.
defining the following resources in your .Xdefaults file should work: XImtool*imagewin.color0: white
XImtool*imagewin.color1: blackThese may also be specified on the command line using % ximtool -xrm "*imagewin.color0:white" -xrm "*imagewin.color1: black"Cheers,
-Mike

>From rajat@samba.tifrbng.res.in Thu Aug 14 13:30:42 1997
From: Rajat Varma <rajat@samba.tifrbng.res.in>
To: Mike Fitzpatrick <fitz@noao.edu>
Subject: Re: some trivial problemsDear Mike,

> because of the limitations of the SGI kernel. If you were to use the
> PostScript kernel in STSDAS (if you have it installed) you could retain
> the color information when printing, note this requires some graphcap

Is it possible to install just the Postscript kernel in STSDAS
and not the whole package, if so please let me know what are the graphcap
changes to be made.Thanks a millionrajat>From fitz Thu Aug 14 15:42:27 1997
To: rajat@samba.tifrbng.res.in
Subject: Re: some trivial problemsRajat,
If you let me know what kind of machine you have I'll see if I
can find a precompiled binary, but otherwise the PS kernel isn't packaged
so it can be built/installed separately. The graphcap entries we have
are[graphcap entries deleted]You would used the device 'g-psiland' or 'g-psiport' to send the output to
the default printer. If all you're going to install is the binary then
you'll also need to change the :kf field for psi_def so the path to the
binary is correct. Let me know if you have problems or questions.Cheers,
-Mike

>From rajat@ncbs.tifrbng.res.in Fri Aug 15 11:41:22 1997
From: Rajat Varma <rajat@ncbs.tifrbng.res.in>
To: Mike Fitzpatrick <fitz@noao.edu>
Subject: real problemsDear Mike,
I could download STDAS and Tables from the eso mirror site, I
downloaded the rpm files as I am running iraf on redhat. After I
installed as instructed, I am just not able to run iraf I get the
following junk.[rajat@image ~]$ cl
INTERNAL ERROR: parser gagged
cl ()
cl ()
Error while reading login.cl file - may need to rebuild with mkiraf
Fatal startup error. CL dies.
[rajat@image ~]$ mkiraf
Initialize uparm? (y|n): y
-- initializing uparm
Terminal types: xgterm,xterm,gterm,vt640,vt100,etc.
Enter terminal type: xgterm
A new LOGIN.CL file has been created in the current directory.
You may wish to review and edit this file to change the defaults.
[rajat@image ~]$ cl
INTERNAL ERROR: parser gagged
cl ()
cl ()
Error while reading login.cl file - may need to rebuild with mkiraf
Fatal startup error. CL dies.
[rajat@image ~]$ so as mentioned above I remade iraf and now all is kind of lost, do you
have any suggestions as to what could have gone wrong?rajat>From fitz Fri Aug 15 12:48:37 1997
To: rajat@ncbs.tifrbng.res.in
Subject: Re: real problemsHi Rajat,
A 'parser gagged' error when starting the CL after installing a
new package usually means there's some sort of syntax error in the
package declaration in the hlib$extern.pkg file, e.g. you didn't put
in an '=' sign, didn't escape the newline when editing the helpdb string,
etc. If you can't spot the error feel free to send me the file. The error
might also be in your loginuser.cl, but only if you've changed it recently.
Note also that the RPM files in the contrib directory were generated
by another user who wanted to create ELF binaries for IRAF and various pkgs.
We can't vouch for whether these work or not, if you have problems you can
get the source and linux binaries from the /contrib/stsdas135 directory
and just install those directly. Let me know if you still have problems.Cheers,
-Mike

>From rajat@ncbs.tifrbng.res.in Sat Aug 16 03:39:09 1997
To: Mike Fitzpatrick <fitz@noao.edu>
Subject: Re: real problemsDear Mike,
Thanks a lot that was easy, there was one spurious statement
which got appended after the keep statement, which was causing all the
trouble. Meanwhile, I am downloading the other the two files,
stsdas135.tar.gz and stsdas135-bin.linux.gz, do I need both or either,
which one is recommended, and do I need to get tables software also if I
have to use the stsdas? Thanks again,rajat>From fitz Sat Aug 16 13:43:59 1997
To: rajat@ncbs.tifrbng.res.in
Subject: Re: real problemsRajat,
TABLES is needed by STSDAS if you're building from source, and some
tasks also use it. If you just want the PS kernel you can ignore it for now.
Just untar the source file, then unpack the binaries in the bin.linux dir-
ectory and declare it in extern.pkg and you're done.Cheers,
-Mike

>From rajat@ncbs.tifrbng.res.in Sat Aug 16 16:09:24 1997
To: Mike Fitzpatrick <fitz@noao.edu>dear mike,
I have installled both tables and stsdas on my computer here, now
I have just one last problem and that is, the entry in the graphcap file,
making the required entries was trivial(cp,paste), but from the examples
in the psikern.temp file, which line has to be prepended to graphcap
before required entries. the following are the examples given# EXAMPLES: The below entries are examples of top-level entries for output
# to specific printers. There are four basic entries, two for the two
# orientations, Landscape or Portrait, and for each orientation, two different
# default line widths, one that uses the default, and another that uses a
# line width thinner than the default. The print queue where the output
# will be sent is defined by the QQ parameter. See the examples below for how
# the QQ parameter is used.
nem409n|nem|nemlw|nemlp|lw3|lp3|Nemesis (N409) :QQ=lp3:tc=psi_land_q:
nem409nl|neml|nemlwl|nemlpl|lw3l|lp3l|Nemesis :tc=nem409n:
nemlws|nemls|nemlps|lp3s|lw3s| :QQ=lp3:tc=psi_land_qs:
nem409np|nemp|lw3p|lp3p|Portrait :QQ=lp3:tc=psi_port_q:
nem409nps|nemps|nemps|Thin Portrait :QQ=lp3:tc=psi_port_qs:# EXAMPLES: The below examples describe a generic way of sending the output
# to queues if the commands for all the queues are the same, i.e. for UNIX it
# is the "lpr" command.
psi_port_qs|PostScript Kernel send output to queue QQ, thin lines, Portrait:\
W#.00008I#.5:tc=psi_port_q:
psi_land_qs|PostScript Kernel send output to queue QQ, thin lines, Landscape:\
W#.00008I#.5:tc=psi_land_q:
psi_port_q|PostScript Kernel send output to queue QQ, Portrait:\
D=psi_def,tmp$psk,!{ lpr -P$(QQ) $F ; rm $F ; }&:tc=psi_port:
psi_land_q|PostScript Kernel send output to queue QQ, Landscape:\
D=psi_def,tmp$psk,!{ lpr -P$(QQ) $F ; rm $F ; }&:tc=psi_land: I have defined my printer as lp and I use the lpr command to print
Thanksrajat>From fitz Sat Aug 16 16:22:04 1997
To: rajat@ncbs.tifrbng.res.inRajat,
Change the first set of examples to readlp| :QQ=lp:tc=psi_land_q:
lps| :QQ=lp:tc=psi_land_qs:
lpp|Portrait :QQ=lp:tc=psi_port_q:
lpps|Thin Portrait :QQ=lp:tc=psi_port_qs: And sending the plot to device 'lp' will use the PS kernel. The entries
I sent last time from our graphcap will use the default printer for
the system, either set should work.Cheers,
-Mike>From rajat@ncbs.tifrbng.res.in Sat Aug 16 22:20:45 1997
To: Mike Fitzpatrick <fitz@noao.edu>Dear Mike,
I made the required changes in the graphcap file, the entries you
sent earlier, I have prepended those and before that I have added the one
you have just sent me. I have declared stdplot=lp in login.cl, if I try
to print it gives the following error messageERROR: No entry in device capabilities file for device (psi_land_q)Obviously I have not made a correct change somewhere, can you let me know
what has gone wrong?thanks
rajat>From fitz Sat Aug 16 23:44:01 1997
To: rajat@ncbs.tifrbng.res.inRajat,
Graphcap entries are inked together by the :tc field, so an entry
likelp| :QQ=lp:tc=psi_land_q:psi_land_q|PostScript Kernel send output to queue QQ, Landscape:\
D=psi_def,tmp$psk,!{ lpr -P$(QQ) $F ; rm $F ; }&:tc=psi_land:is linked by the ":tc=psi_land_q", but note the second entry also has a
:tc field so you also need the 'psi_land' entry to complete the description.
psi_land should be in the examples page or else there's one in the stuff I
sent. If psi_land_q isn't being seen on your system be sure you're not
missing an escaped newline somewhere that may be terminating the field
early (e.g. a cut-n-paste might break up a long line as it did in the mail
you sent for these entries. Let me know if you still have problems.Cheers,
-Mike

>From rajat@ncbs.tifrbng.res.in Sun Aug 17 01:49:55 1997
To: Mike Fitzpatrick <fitz@noao.edu>
Subject: hiDear Mike,
I have finally given up, I am sorry to bug you so much, but I am sending
you my login.cl as well as graphcap file, please see what is wrong, now
the error message is the followingERROR: No GIO kernel specified for graphics device (lp)in printtool for redhat, I have given printer name as lp, but the command
I use is lpr. So accordingly please suggest what changes are to be made
in login.cl and graphcap. Please take your own time, I am in no hurry so
whenever you are free you can look in to this.thanks a million
rajat>From fitz Sun Aug 17 20:15:20 1997
To: rajat@ncbs.tifrbng.res.inRajat,
You need to change the :kf field (kernel file) to read :kf=stsdas$bin($arch)/x_psikern.e:\As it stands now it assumes that STSDAS is configured already for your
architecture, if not the 'bin' directory points to an empty bin.generic.
Remember that after you reset the graphcap to use your private file you'll
also need to type "gflush" so the changes get read. Let me know if you
still have problems.Cheers,
-Mike

>From rajat@ncbs.tifrbng.res.in Mon Aug 18 11:57:27 1997
To: Mike Fitzpatrick <fitz@noao.edu>
Subject: Re: hiDear Mike,
I made this change as it is in the graphcap file, am i supposed
to type linux in place of the word arch down below. if I made the change
as it is nothing happens, nor do any errors appear but it doesnt print
either, there are several kf fields in the graphcap file, do I have to
make this change everywhere or only at the beginning of the file, where I
have the psi_def? and in the login.cl I just have to assign stdplot as lp
or lpr, lp is what I have given the name in the printtool but I use lpr
to print, and if I do give lpr I do get a print out but it is in black
and white and not colour, using sgraph I see color on the screen, do I
have to use some other plotting command?
Sorry mike I am sounding very naive but thats what the situation
is and I am not able to figure out what is going wrong.Thanks
rajat>From rajat@ncbs.tifrbng.res.in Mon Aug 18 12:13:02 1997
To: Mike Fitzpatrick <fitz@noao.edu>
Subject: Re: hi

Dear Mike,
I have tried something which generated a postscript file in /tmp/
called pskxxx something and it is showing the proper colours, this is
what I typedst> sgraph s2.dat,s9.dat device=psi_land
st> gflush
/tmp/psk1121aSo now I can view the above file from ghostview but how can I direct it
to the printer directly ,somewhere lp -> psi_land information is not
being passed.Thanks
rajat>From fitz Tue Aug 19 00:06:24 1997
To: rajat@ncbs.tifrbng.res.inRajat,> am i supposed to type linux in place of the word arch down below.
The "bin($arch)" will resolve to bin.linux automatically in your
system so type it in directly as I did, otherwise use "stsdas$bin.linux".> there are several kf fields in the graphcap file, do I have to make this
> change everywhere or only at the beginning of the file, where I have the
> psi_def?
Sorry, you should change the :kf field under the psi_def entry
only, all the rest of the PSIKERN entries should link to this.> and in the login.cl I just have to assign stdplot as lp or lpr,
> lp is what I have given the name in the printtool but I use lpr
> to print, and if I do give lpr I do get a print out but it is in black
> and white and not colour, using sgraph I see color on the screen, do I
> have to use some other plotting command?
'lp' is the correct name but only so long as that's the *first*
entry in the graphcap file, i.e. the new one you just added and not the
old 'lp' entry already in there that goes to the standard SGI kernel and
B&W output. The graphcap you sent me had the new stuff at the very top
so this should have been picked up, were you sure to do a "gflush" after
the "reset graphcap" command so it was reinitialized? Otherwise the only
'lp' device recognized is the old one.
What do you mean by 'printool' (the OpenWin command)? Do you
print using "lpr -Plp" or some other command, this is what needs to be
in the graphca DD string.> I have tried something which generated a postscript file in /tmp/
> called pskxxx something and it is showing the proper colours, this is
> :
>
> So now I can view the above file from ghostview but how can I direct it
> to the printer directly ,somewhere lp -> psi_land information is not
> being passed.
What you did by accessing the psi_land device is use the low-level
entry for the kernel where the DD string says to move the generated PS
file to pskXXX.eps in the current working directory (i.e. /tmp). What
you're trying to get with the 'lp' device is an entry where the 'DD' string
uses an 'lpr' command to instead print the file, i.e.lp| tc=psi_land_q:psi_land_q|PostScript Kernel send output to queue QQ, Landscape:\
D=psi_def,tmp$psk,!{ lpr -Plp $F ; rm $F ; }&:tc=psi_land:I've already deleted the graphcap you sent, if you still have problems would
you mind resending it?Cheers,
-Mike

>From rajat@ncbs.tifrbng.res.in Tue Aug 19 12:30:33 1997
To: Mike Fitzpatrick <fitz@noao.edu>
Subject: hiDear Mike,
Thanks a lot, the only problem in all this was that I had not
reset the graphcap file,now that I have done it I am getting the desired
color printouts, now I have only one more querry, which is the file I
need to modify to change the various colours in the border of the xgterm
window that appears, as it stands the default color is greyish green, so
I want to get rid of that because it unnecessarily prints out that border
in the graphics as well, once I know which file to modify then I can play
around with it and do what I want to. Also is there a way I can play
around with the size of the graph box that is printed out, can I reduce it?
rest is fine, thanks once again.
rajat>From fitz Tue Aug 19 13:23:54 1997
To: rajat@ncbs.tifrbng.res.in
Subject: Re: hiRajat,
I'll append an FAQ that describes how the colors are determined,
probably what you want to do is change the 'fr' part of the 'glbcolor'
environment variable or else reset the color9 resource for XGterm to
be white.> Also is there a way I can play around with the size of the graph box that
> is printed out, can I reduce it?
Yes, but you'll need to change the graphcap so you link to entries
with a different viewport size. For example,lp| :YO#0.21:XO#0.2576:yr#1156r#1532:tc=psi_land_q:The :XO and :YO fields are the offset of the plot on page in NDC coordinates
(in the range 0.0 -> 1.0), the r and :yr fields are the resolution. For
details see the "help psikern opt=sys" page or you can ask specific questions
at help$stsci.edu since they maintain the software.Cheers,
-Mike[deleted FAQ 14.1]

>From rajat@ncbs.tifrbng.res.in Thu Aug 21 08:07:41 1997
From: Rajat Varma <rajat@ncbs.tifrbng.res.in>
To: Mike Fitzpatrick <fitz@noao.edu>
Subject: hiDear Mike,
I could figure out how to change colours in the window of xgterm
but there are a few problems.a) The change in colour doesnt reflect in the printouts I print in the
following manner.psikern 57.meta device=stdplotb) Once the graphcap has been changed to use the new PS kernel, I can no
longer print graphs asgkiextract 57.meta |stdplotso does it mean I have to always use psikern to print any graph?c) I was unable to change the environment variables, is doing the
following way correct?reset fr=0I finally changed the colour entries in the .Xdefaults file, but suppose
I want to play around more with colours, then how shall I do it?
Rest is fineThanks for all the help you have been extending!rajat>From fitz Thu Aug 21 11:22:24 1997
To: rajat@ncbs.tifrbng.res.in
Subject: Re: hiHi Rajat,> a) The change in colour doesnt reflect in the printouts I print in the
> following manner.
>
> psikern 57.meta device=stdplot
This depends on what the value of the 'stdplot' environment variable
is. If it's the 'lp' entry you added then it should send color PS to the
printer, if it's some other device it may use the old sgi kernel. Or do you
mean you get color, just not the what you expecting?> gkiextract 57.meta |stdplot
>
> so does it mean I have to always use psikern to print any graph? The STDPLOT kernel also doesn't preserve the color, you'll need to
call the PSIKERN directly as with cl> gkiextract 57.meta 1 | psikern dev=lp> c) I was unable to change the environment variables, is doing the
> following way correct?
>
> reset fr=0
You need to reset the value of the 'glbcolor' variable, e.g. cl> reset glbcolor = "pt=3,fr=0,al=3,tl=6,ax=5,tk=5"> I finally changed the colour entries in the .Xdefaults file, but suppose
> I want to play around more with colours, then how shall I do it?
I'm not sure what you mean. XGterm can only set 9 colors at a time
using the X resources. These can be any color you want, but you have to set
them as a resource and restart XGterm for them to take effect. On the
command line you could start xgterm using something like % xgterm -xrm "*color9:sienna"as a way of experimenting until you find a color you like, then just set
the resource in your .Xdefaults file and reload the database with % xrdb -load ~/.Xdefaults.Using 'glbcolor' you can select from the available xgterm colors for the
plot options, this change should take effect immediately. Does this answer
your question?Cheers,
-Mike

>From rajat@ncbs.tifrbng.res.in Thu Aug 21 14:12:26 1997
From: Rajat Varma <rajat@ncbs.tifrbng.res.in>
To: Mike Fitzpatrick <fitz@noao.edu>
Subject: Re: hi> > a) The change in colour doesnt reflect in the printouts I print in the
> > following manner.
> >
> > psikern 57.meta device=stdplot
> This depends on what the value of the 'stdplot' environment variable
> is. If it's the 'lp' entry you added then it should send color PS to the
> printer, if it's some other device it may use the old sgi kernel. Or do you
> mean you get color, just not the what you expecting?Dear Mike,
Well what I meant above was that if I change fg color from
darkslate grey to white, I see white on the screen but not on the
printout, I shall try again, the printout still gave the default colour
that was there before.Thanks
rajat>From fitz Fri Aug 22 11:40:39 1997
To: rajat@ncbs.tifrbng.res.in
Subject: Re: hiRajat,
The glbcolor just selects which of the 9 available colors are to
be used for various parts of the plot label. By changing the fr=0 you
say that you want the frame border to be done in color 0, which is whatever
*color0 resource you defined for the gterm widget. I just tried this again
and the frame color was correctly changed in the saved postscript file
generated by the psikern. Did you type gflush after resetting glbcolor?
Did you reset any of the Gterm widget color resources?> if I change fg color from darkslate grey to white ...
Note the darkslategrey isn't a foreground/background color, it's
the color9 resource value. If all you did was restart xgterm using the
-fg or -bg switch you're not changing the color9 resource. If fr=9 you'll
get whichever color9 is set to, if it's fr=0 you'll get color0 (background).
Let me know if you still have problems.Cheers,
-Mike

>From rajat@ncbs.tifrbng.res.in Fri Aug 22 23:45:27 1997
To: Mike Fitzpatrick <fitz@noao.edu>Dear Mike,
I think I have created a lot of confusion, I have followed
everything you have said and I have played around with everything as
well, but it has still not solved my problem, I have changed colours and
I know they get affected by psikern as well, but I will just tell you now
what i really want to do. I have attached a postscript file generated by
psikern in /tmp, if you look at the graph, the central area appears white
while there is a dark border around the graph in which the tick labels
appear in blue and the graph title appears in green, I want this dark
area to be white as well, because I waste a lot of ink while printing
otherwise and it looks horrible. I hope I have not bothered you too much
but I am not an expert in all this so I need to get help sometimes.Thanks a lot for being so patientrajat[PS file deleted]>From fitz Sat Aug 23 01:08:16 1997
To: rajat@ncbs.tifrbng.res.inHi Rajat,
Perhaps it is I who have confused you: I've looked at the plot
and from what I see the white are is the background (color0) color and
the dark frame border is what you get from the color9 resource value (it
still looks like darkslategray). Resetting glbcolor to get a white frame
really does work, but if you're starting xgterm with "-rv" or by resetting
the fg/bg colors it may be be getting confused. The fr=0 says the area
around the plot should be drawn in the color0 value (i.e. the background
value). Using "-rv" swaps the fg/bg (color0/color1) values so if you're
doing this perhaps what you need is fr=1.
Assuming xgterm uses a dark bg and a light fg color, and you don't
define any of the color resources you should get a light background by
setting fr=0 in the glbcolor variable, this will draw the frame in the same
color as the center of the plot. To see what color resources are defined
for XGterm (either personally or with a system app-defaults file) use
the command % xrdb -query | grep colorif you get any output at all some color is being set somewhere. What happens
if you do the following: % xrdb -query | xrdb -remove
% xgterm -e cl
....from the new xgterm window..
cl> reset glbcolor = "pt=3,fr=9,al=3,tl=6,ax=5,tk=5"
...then create a plot on disk, should have dark bg
cl> reset glbcolor = "pt=3,fr=0,al=3,tl=6,ax=5,tk=5"
cl> gflush
...then create a plot on disk, should have light bgThe blue axis labels are non-standard, do you always see this or did you
reset a color resource or glbcolor env variable to get it?

> I hope I have not bothered you too much ...
Not a problem...Cheers,
-Mike>From rajat@ncbs.tifrbng.res.in Sat Aug 23 02:07:01 1997
To: Mike Fitzpatrick <fitz@noao.edu>Thanks a lot Mike, finally it worked, I just followed step by step what
you asked me to and I am getting the desired output. but everytime I
start a iraf session I will have to do the resetting of glbcolor? or is
there a way of permanantly setting these colors? also do I have to always
do the following as well i.e. xrdb -query | xrdb remove, because I feel
this is the thing that has done the trick.Thanks again
rajat>From rajat@ncbs.tifrbng.res.in Sat Aug 23 03:54:44 1997
To: Mike Fitzpatrick <fitz@noao.edu>Well Mike,
I have tested out everything, you need not bother about anything
else, I have included a set statement for glbcolor in my login.cl that
takes care of everything, so thanks a lot once again.rajat>From fitz Sat Aug 23 13:11:56 1997
To: rajat@ncbs.tifrbng.res.inRajat,
As you've found you can declare glbcolor in your login.cl or
loginuser.cl file, just be sure it goes before the final 'keep' statement
at the end of the file. If you think it really may be a resource problem
them use
% xrdb -query to look for 'color' commands, specifically color9. You can also just try
it with and without removing the resources to see if it makes a difference,
if there is no color9 defined it probably doesn't matter. If there is then
look either in your .Xdefaults file or the system file which is usually
either in /usr/lib/X11/app-defaults/XGterm or in /usr/openwin/lib/app-defaults.Cheers,
-Mike

>From rajat@ncbs.tifrbng.res.in Sun Aug 24 05:30:31 1997
To: Mike Fitzpatrick <fitz@noao.edu>Dear Mike,
One last querry! How do I change the font size of axis title and
plot title in all the graphs I have been generating? Rest is all working
very finerajat>From fitz Sun Aug 24 23:02:26 1997
To: rajat@ncbs.tifrbng.res.inRajat,
For the PSIKERN you would need to modify the :cw and :ch fields
of the graphcap entry (probably in psi_def). But be careful since it appears
the spacing for the strings isn't adjusted properly is you make if too big.
In general though using ":.txset size=1.5" is one way to adjust the size
of other text, e.g. stuff drawn with a 'T' keystroke command. You might
also try resetting the font graphcap entries to something which includes a
size, I haven't tried this though. Does this answer your question?Cheers,
-Mike

>From rajat@ncbs.tifrbng.res.in Fri Sep 5 02:19:15 1997
From: Rajat Varma <rajat@ncbs.tifrbng.res.in>
To: Mike Fitzpatrick <fitz@noao.edu>Dear Mike,
Hi, Hope this mail finds you in great weekend spirits. I have one
yet another small querry. While using sgraph or graph, how does one
increase the thickness of line segments connecting points, I have to make
slides out of these, so I want them to look a little more prominent than
what they are now.thanks
rajat>From fitz Fri Sep 5 12:44:45 1997
To: rajat@ncbs.tifrbng.res.inHi Rajat,
The only way to change the line width, for those tasks, is to
change the W (pen width) graphcap entry for the hardcopy output. However,
this also changes the width of the plot labels and can sometimes make
it unreadable. Things to try would be something likelw1| W#2.6:tc=uapl1:The format of the parameter is 'origin'.'slope' so you might experiment
a bit to find a pair of values that works for you.
Another way to do this would be to plot using IGI, the LWEIGHT
command there lets you set the line width for just the data if you want.
Hope this helps.Cheers,
-Mike

>From rajat@ncbs.tifrbng.res.in Fri Sep 5 20:54:23 1997
To: Mike Fitzpatrick <fitz@noao.edu>
Subject: Re: hiDear Mike,
As usual I have a question. In the graphcap file there are
various PW entries, which one should I choose to make the appropriate
changes, or should I just append a line like the one you have sent me below?
I am still not familiar with using IGI, so there I have a problem. Thanks rajat>From fitz Fri Sep 5 21:16:49 1997
To: rajat@ncbs.tifrbng.res.inRajat,
You'll need to edit the graphcap entry for whichever printer
you're using. Most likely it won't have a W field you'll see so just
add it before the :tc as in my example. Be sure after editing your
graphcap file you type 'gflush' to reinitialize the system.Cheers,
-Mike