gnuplot-info — For questions and discussion about gnuplot

[Gnuplot-info] Re: Q: Fraction in seconds

[Hans-Bernhard B. <br...@ph...>]

On Thu, 27 May 2004, Reinhold Zwingel wrote:

> I'm using patchlevel 3.7.1.

Please update.  You're several years behind...

> I've a data file with the time format 
> HH:MM:SS.ssss
> When I read the time, the seconds are read as Integer 0 to 60 (only SS).
> How can I read the fraction .ssss ?

You can't, not even with version 4.0, but that's OK, because you probably
don't really have to.  It's practically impossible to need both time spans
of more than one hour, and sub-second resolution, in the same plot ---
that would require more then 3500 pixels of usable horizontal resolution
on the output format, which practically never is the case.


-- 
Hans-Bernhard Broeker (br...@ph...)
Even if all the snow were burnt, ashes would remain.

Re: [Gnuplot-info] Tic labels overwrite axes

[Hans-Bernhard B. <br...@ph...>]

On Fri, 28 May 2004, Periklis Akritidis wrote:

> On Thu, May 27, 2004 at 11:49:02PM +0200, Hans-Bernhard Broeker wrote:
> > On Thu, 27 May 2004, Periklis Akritidis wrote:
> > 
> > > Hello,
> > > 
> > > Even a simple 3d plot such as splot sin(x) + cos(y) results in numbers
> > > overwriting each other (at the points where the axes meat each other)
> > > as well as the axes. 
> > 
> > What terminal driver?  Which version of gnuplot?  Which font did you use,
> > at which size?
> 
> Postscript driver and gnuplot 4.0, I tried various fonts, but I need them
> to have a certain size so that they are readable, which I guess makes things
> worse (I can prevent them from overwriting each other by changing the xtics
> increment, the problem with overwriting the axes, x and y, remains).

A concrete example would have been better, but I've since found one
myself...

If you compare the behaviour of PostScript with that of the X11 terminal,
you'll find that the problem is almost completely absent in X11 (except at
the corners, where tick labels from neighboring axes do tend to overlap,
but I suspect there's really not much that can be done about that).

I.e. it's quite certainly the mismatch between PostScript's actual font
metrics and gnuplot's guesses of them that's causing this.

> I understand that it may be non-trivial to to it automagically, what I
> was hoping for was something like the xoff and yoff options to xlabel.

Alas such an option doesn't exist.  But here's an alternative that *might*
help: "set tics out".  This will put the tickmarks to the outside of the
graph box, thus pushing the tick labels further away from the borders. You
can even use the ticscale setting to decide how far way they'll be.

> So if such an option does not exist, I would like to submit a wishlist
> item:) What syntax would be appropriate for such an option?

Well, fell free to register this as a Feature Request in the
SourceForge.net project pages.  The syntax should probably be xoff and
yoff options to "set xtics" and friends.

-- 
Hans-Bernhard Broeker (br...@ph...)
Even if all the snow were burnt, ashes would remain.

Re: [Gnuplot-info] Tic labels overwrite axes

[Periklis A. <ak...@ic...>]

On Thu, May 27, 2004 at 11:49:02PM +0200, Hans-Bernhard Broeker wrote:
> On Thu, 27 May 2004, Periklis Akritidis wrote:
> 
> > Hello,
> > 
> > Even a simple 3d plot such as splot sin(x) + cos(y) results in numbers
> > overwriting each other (at the points where the axes meat each other)
> > as well as the axes. 
> 
> What terminal driver?  Which version of gnuplot?  Which font did you use,
> at which size?

Postscript driver and gnuplot 4.0, I tried various fonts, but I need them
to have a certain size so that they are readable, which I guess makes things
worse (I can prevent them from overwriting each other by changing the xtics
increment, the problem with overwriting the axes, x and y, remains).

> Note that for the majority of PostScript fonts this task is harder than
> you may be aware of: With proportional fonts, gnuplot has no means to know
> how wide a given string is actually going to be on output.  Even with
> fixed-width fonts, it has to guess that width from the font size you gave
> it, because it doesn't have access to font metrics.

I understand that it may be non-trivial to to it automagically, what I
was hoping for was something like the xoff and yoff options to xlabel.
Of course ideally you would not have to adjust such things, but with such
options I could manually adjust the offsets so that they do not overwrite
the axes in 3d plots even if gnuplot did not guess correctly the font
height and width. I haven't found such an option but I guess that if
it has been implemented for xlabel, it should not be more difficult
to implement it for the axes numbers too.  Actually, xlabel sometimes
has similar problems, but they are always fixable manually using xoff
and yoff.

So if such an option does not exist, I would like to submit a wishlist
item:) What syntax would be appropriate for such an option?

Thanks,
Pericles

-- 
Periklis Akritidis

Re: [Gnuplot-info] Tic labels overwrite axes

[Hans-Bernhard B. <br...@ph...>]

On Thu, 27 May 2004, Periklis Akritidis wrote:

> Hello,
> 
> Even a simple 3d plot such as splot sin(x) + cos(y) results in numbers
> overwriting each other (at the points where the axes meat each other)
> as well as the axes. 

What terminal driver?  Which version of gnuplot?  Which font did you use,
at which size?

Note that for the majority of PostScript fonts this task is harder than
you may be aware of: With proportional fonts, gnuplot has no means to know
how wide a given string is actually going to be on output.  Even with
fixed-width fonts, it has to guess that width from the font size you gave
it, because it doesn't have access to font metrics.

-- 
Hans-Bernhard Broeker (br...@ph...)
Even if all the snow were burnt, ashes would remain.

[Gnuplot-info] Tic labels overwrite axes

[Periklis A. <ak...@ic...>]

Hello,

Even a simple 3d plot such as splot sin(x) + cos(y) results in numbers
overwriting each other (at the points where the axes meat each other)
as well as the axes. This seems to be the result
of the rotated view. Is there any way to avoid this (at least for the
postscript terminal) without residing to tricks such as adding spaces
and newlines in the format (set format y "       %g")?

PS. Please Cc: me as I am not subscribed to the list.

-- 
Periklis Akritidis

[Gnuplot-info] subscribe info-gnuplot

[Zwingel, R. <Rei...@BS...>]

Re: [Gnuplot-info] how to cite gnuplot ?

[Hans-Bernhard B. <br...@ph...>]

On Wed, 26 May 2004, Gilles Hunault wrote:

> 
>  I would like how I should cite gnuplot in an
>  article since there are many authors. The use
>  of gnuplot for our article does not rely on a
>  specific version of gnuplot since it uses
>  basic (x,y) plotting.

While we don't have any particular policy about this, I guess the
correct citation would be

author: Thomas Williams, Colin Kelley et al.
Title: gnuplot"
reference:  http://www.gnuplot.info

-- 
Hans-Bernhard Broeker (br...@ph...)
Even if all the snow were burnt, ashes would remain.

[Gnuplot-info] how to cite gnuplot ?

[Gilles H. <gh...@in...>]

 I would like how I should cite gnuplot in an
 article since there are many authors. The use
 of gnuplot for our article does not rely on a
 specific version of gnuplot since it uses
 basic (x,y) plotting.

 Thank you for your help and for gnuplot.

<<Au plaisir de vous relire.>>
--
(.=.)
Gilles J. HUNAULT, gil...@un...
"Science needs culture and feelings"
http://www.info.univ-angers.fr/pub/gh/

[Gnuplot-info] FASD project: Online survey launched

[<ben...@id...>]

Dear Open Source developer

I am doing a research project on "Fun and Software Development" in which I kindly invite you to participate.
You will find the online survey under http://fasd.ethz.ch/qsf/. The questionnaire consists of 53 questions and you will need about 15 minutes to complete it.

With the FASD project (Fun and Software Development) we want to define the motivational significance of fun when software developers decide to engage in Open Source projects. What is special about our research project is that a similar survey is planned with software developers in commercial firms. This procedure allows the immediate comparison between the involved individuals and the conditions of production of these two development models. Thus we hope to obtain substantial new insights to the phenomenon of Open Source Development.


With many thanks for your participation,
Benno Luthiger


PS:
The results of the survey will be published under http://www.isu.unizh.ch/fuehrung/blprojects/FASD/.
We have set up the mailing list fa...@we... for this study. Please see http://fasd.ethz.ch/qsf/mailinglist_en.html for registration to this mailing list.

_______________________________________________________________________

Benno Luthiger
Swiss Federal Institute of Technology Zurich
8092 Zurich

Mail: benno.luthiger(at)id.ethz.ch
_______________________________________________________________________

[Gnuplot-info] Plotting discrete data with pm3d?

[Jerome L. <jer...@nr...>]

Hello,

	I'm having special sets of 2D data, where the y coordinate is
continuous but the x coordinate is discrete (the x's are labels for
spectral peaks and the y's represent the experimental variable). I want
to plot my results using 'pm3d map'.

	The problem is that pm3d uses the average of neighboring points to
attribute a colour to the region between these points; the result is
that the intensities of my 'peaks' are all mixed together.

	Is there a way to tell pm3d not to show the average value for a given
axis?  

Thanks,

Jerome Levesque

-- 
Jerome Levesque <jer...@nr...>

Re: [Gnuplot-info] 16 bit problem?

[Hans-Bernhard B. <br...@ph...>]

On Sat, 22 May 2004, Tore Johansson wrote:

> I get only 65536 number of samples plotted. Is this a limitation in
> gnuplot or a bug??

Hard to tell --- you don't tell what version of gnuplot, what platform you
experience this on, nor what terminal driver you were using.

-- 
Hans-Bernhard Broeker (br...@ph...)
Even if all the snow were burnt, ashes would remain.

[Gnuplot-info] 16 bit problem?

[Tore J. <gn...@bu...>]

Hi,

I am trying to plot a series of samples stored in a datafile.
It works fine if I do as follow:
plot 'datafile'

But if I do:
plot 'datafile' with lines 1

I get only 65536 number of samples plotted. Is this a limitation in gnuplot or a
bug??

BR
/Tore Johansson


----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.

Re: [Gnuplot-info] (no subject)

[Hans-Bernhard B. <br...@ph...>]

On Fri, 21 May 2004 sa...@em... wrote:

> Can you please tell me what the maximum file size restrictions are for
> visualization of a data file using a splot command. 

"The sky's the limit", as they say.

In more scientific terms: there's no fixed limit on file size imposed by
gnuplot.  The only effective limit is given by the amount of RAM you can
make available to gnuplot.  Depending on the plotting mode you show, each
datapoint will consume a fixed amount of storage.  As long as all of that
fits into your main memory, you'll be fine.

-- 
Hans-Bernhard Broeker (br...@ph...)
Even if all the snow were burnt, ashes would remain.

[Gnuplot-info] (no subject)

[<sa...@em...>]

Hello

Can you please tell me what the maximum file size restrictions are for
visualization of a data file using a splot command. I need to know that for my
thesis defense.

I would really appreciate if you could tell give me the relevant information

Thank you
salil

Re: [Gnuplot-info] parts chart

[Hans-Bernhard B. <br...@ph...>]

[Note: mailing list changed to the new one --- old one causes delays
because all redirected posts from dartmouth have to be approved
individually...]

On Wed, 19 May 2004, Dan Jacobson wrote:

> H> 	set size square
> H> 	set ticslevel 0
> 
> Odd that you squash by default. A rather one sided decision :-)

Not really.  Just remember we're in the plotting business, not in 3D CAD
or solid modelling.  Preservation of aspect ratios comes second to getting
good plots that use as much of the available real estate on screen and
on paper as possible.

-- 
Hans-Bernhard Broeker (br...@ph...)
Even if all the snow were burnt, ashes would remain.

Re: [Gnuplot-info] save a plot

[Hans-Bernhard B. <br...@ph...>]

On Wed, 19 May 2004, Francy wrote:

> How can I save a graphic of 'plot datafile'?

You don't exactly just "save" it.  You create it again, this time to
a file-based terminal driver you choose.  In the simplest of all cases:

	set term push  # remember what it was
	set term png ; set out 'myplot.png'  # turn to PNG creation
	replot
	set out ; set term pop  # return to screen terminal

-- 
Hans-Bernhard Broeker (br...@ph...)
Even if all the snow were burnt, ashes would remain.

[Gnuplot-info] save a plot

[Francy <fra...@ti...>]

How can I save a graphic of 'plot datafile'?
Thanks.
Francesca

Re: [Gnuplot-info] parts chart

[Dan J. <ji...@ji...>]

H> [Please restrict yourself to posting to *one* mailing list.  

I was guessing that it was another project that bounces bugs from
non-subscribers, so I tried full power mode. Glad to know that isn't
the case.

H> best be done as a demo script.
good idea

>> Splotting a sphere looks squashed.

H> 	set size square
H> 	set ticslevel 0

Odd that you squash by default. A rather one sided decision :-)

H> In spite of the docs saying 'set size square' doesn't affect 3D plots, it
H> does.

glad I didn't RTFM.
Ok, instant sphere on Debian
#show _evenly spaced_ points on a sphere
dxfsphere 5|dxf2vrml|
perl -nlwe "BEGIN{print q{unset key;set size square;set ticslevel 0;splot '-'}};
print if (/point/../Indexed/)&&/\d/"|
gnuplot -persist #not from emacs' compile mode though still

Re: [Gnuplot-info] gnuplot question

[<Die...@t-...>]

Hi Dave,
I do not entirely understand your issue. You have a C program generating da=
ta=20
(i.e. threed.dat) having a subroutine writing a plotfile for gnuplot (i.e.=
=20
threed.plt) and then do a
system ("gnuplot threed.plt");
and this does not work?
Or is your problem related to the three-D plotting?
Please give more information to help!
Take care


Dieter

Am Montag, 17. Mai 2004 21:47 schrieb Dave Anderson:
> Hi,
> I am trying to use gnuplot to output 3-dimensional images from a C progra=
m,
***

=2D-=20
=2D----------------------------------------------------------

                               |
                                \
                 /\_/\           |
                | ~x~ |/-----\   /
                 \   /-       \_/
  ^^__   _        /  _  ____   /
 <=B0=B0__ \- \_/     |  |/    |  |
  ||  ||         _| _|    _| _|

if you really want to see the pictures above - use some font
with constant spacing like courier! :-)
=2D----------------------------------------------------------

RE: [Gnuplot-info] Gathering Errors in Batch mode

[Hans-Bernhard B. <br...@ph...>]

On Tue, 18 May 2004, Lauren Clarke wrote:

> Interesting idea... If no one sees any immediate roadblocks to this, I might
> just try it. 
> 
> I'm not all that familiar with the innards of gnuplot, but have utilized
> cygwin before. Anyone have any hunches which environment (mingw, or cygwin)
> would be the most likely to yield results?

Either will, because they're essentially the same thing.  You can use the
Win32+X11 binary provided on our download sites, I think.


-- 
Hans-Bernhard Broeker (br...@ph...)
Even if all the snow were burnt, ashes would remain.

RE: [Gnuplot-info] Gathering Errors in Batch mode

[Lauren C. <la...@co...>]

Interesting idea... If no one sees any immediate roadblocks to this, I might
just try it. 

I'm not all that familiar with the innards of gnuplot, but have utilized
cygwin before. Anyone have any hunches which environment (mingw, or cygwin)
would be the most likely to yield results?

Thanks Dave.
-lc

> -----Original Message-----
> From: Dave Denholm [mailto:dde...@es...]
> Sent: Tuesday, May 18, 2004 7:55 AM
> To: Lauren Clarke
> Cc: gnu...@li...
> Subject: Re: [Gnuplot-info] Gathering Errors in Batch mode
> 
> "Lauren Clarke" <la...@co...> writes:
> 
> > Hi All,
> >
> > Apologies if this is an FAQ, but I'm in the process of integrating
> version 4
> > into our gnuplot-enabled web applications and have the same question I
> had
> > back in 2000 (on ver 3.7).
> >
> > When running on MS Windows, is there any mechanism whereby the errors
> that
> > would be displayed in interactive mode (with LOAD), can be sent to a
> file
> > when running in batch mode? Or, perhaps more simply, what is the proper
> way
> > to get error messages in batch mode?
> >
> 
> If you're running batch mode, and you don't need the windows terminal
> type, have you considered (trying to) build a mingw or cygwin version
> of gnuplot. This may behave more like you'd expect.
> 
> 
> dd
> --
> Dave Denholm              <dde...@es...>
> http://www.esmertec.com

Re: [Gnuplot-info] Gathering Errors in Batch mode

[Dave D. <dde...@es...>]

"Lauren Clarke" <la...@co...> writes:

> Hi All, 
>
> Apologies if this is an FAQ, but I'm in the process of integrating version 4
> into our gnuplot-enabled web applications and have the same question I had
> back in 2000 (on ver 3.7).
>
> When running on MS Windows, is there any mechanism whereby the errors that
> would be displayed in interactive mode (with LOAD), can be sent to a file
> when running in batch mode? Or, perhaps more simply, what is the proper way
> to get error messages in batch mode?
>

If you're running batch mode, and you don't need the windows terminal
type, have you considered (trying to) build a mingw or cygwin version
of gnuplot. This may behave more like you'd expect.


dd
-- 
Dave Denholm              <dde...@es...>       http://www.esmertec.com

Re: [Gnuplot-info] parts chart

[Hans-Bernhard B. <br...@ph...>]

On Mon, 17 May 2004, Dan Jacobson wrote:

[Please restrict yourself to posting to *one* mailing list.  
Cross-posting is bad enough in Usenet, we really don't need it in mailing
lists.]

> Maybe as quick way to find how to adjust different parts of a graph,
> there should be distributed in the docs some .png files with arrows
> pointing to each part of a graph, with names of that part.

That could be *very* hard to do, given the large number of output format
the documentation is available in.  Supplying it as a generatable 
plot like the 'test' page, might work better.  Actually, it'd probably
best be done as a demo script.

> E.g. it took me an hour to find out the name of the part that was
> bothering me was "key", I needed to unset "key", and not "title",
> "label" etc.

"help glossary" would have gotten you there a lot faster ;-|

> Splotting a sphere looks squashed.  I don't know how you can convey
> whatever one needs to do to unsquash it on the diagram I call for
> above, unless it is axis related...

That would be impossible to show in a single page, anyway.  The solution
is

	set size square
	set ticslevel 0

In spite of the docs saying 'set size square' doesn't affect 3D plots, it
does.

-- 
Hans-Bernhard Broeker (br...@ph...)
Even if all the snow were burnt, ashes would remain.

Re: [Gnuplot-info] gnuplot question

[Hans-Bernhard B. <br...@ph...>]

On Mon, 17 May 2004, Dave Anderson wrote:

> Hi,
> I am trying to use gnuplot to output 3-dimensional images from a C program, 
> but am having little success in doing so.  Any help would be greatly 
> appreciated,

It's hard, if not impossible, to help without knowing where exactly
your problem is.  Tell us what you tried, and how it fails.  Knowing
the platform you're on might help, too.


-- 
Hans-Bernhard Broeker (br...@ph...)
Even if all the snow were burnt, ashes would remain.

[Gnuplot-info] parts chart

[Dan J. <ji...@ji...>]

Maybe as quick way to find how to adjust different parts of a graph,
there should be distributed in the docs some .png files with arrows
pointing to each part of a graph, with names of that part.

E.g. it took me an hour to find out the name of the part that was
bothering me was "key", I needed to unset "key", and not "title",
"label" etc.

Splotting a sphere looks squashed.  I don't know how you can convey
whatever one needs to do to unsquash it on the diagram I call for
above, unless it is axis related...