Hello,

RPM packages of ploticus 1.41 (plus older versions) are available at:

	 http://www.megaloman.com/~hany/RPM/ploticus.html

It's the same link as mentioned in http://www.sgpr.net/doc/Download.html
page but only version 1.39 is reported there while version 1.39, 1.40 and
1.41 are actualy available there.

Thanks to Stephen for this great application.

Sincerely

Peter Hanecak

--
=================================================================
= Peter Hanecak             = GnuPG public key:                 =
= hany@...          = http://terminus.sk/\              =
= http://terminus.sk/~hany/ =   ~hany/gnupg-hany-public-key.txt =
=================================================================

Steve,

Yes, that's sounds right. Thank you very much, I appreciate it.

Regards,

Michael

--- In ploticus@y..., "Stephen C. Grubb" <scg@j...> wrote:
> Michael,
>
> OK, now I see what you are asking.
>
> You are using #proc curvefit to generate the curves.
> #proc curvefit fits a smooth curve to the points.
>
> You probably should use #proc lineplot if you want
> your data points shown exactly.  See the attached
> script.
>
>
> Stephen C. Grubb    scg@j...
> Scientific Software Engineer, The Jackson Laboratory
> 600 Main Street  Bar Harbor, Maine 04609 USA
> -------------------------------------------------------

27 Jun 2001

Hey Stephen!

Just wanted to drop you a line to say thanks for the software.
Ploticus was a very quick, easy
and robust solution for our web page.

I am using ploticus through Perl CGI scripting to create GIF images of
stellar spectra, allowing users
to get a "quick glance" at a spectra before downloading it's large (and
cryptic) pixel file.   Ploticus was so well suited for this task that
our data formats never changed to accommodate Ploticus.  With a few
tweaks to the script file, I had GIF's being generated on the fly and
delivered to web browsers within a day of development.

You can find this demonstration at our site
http://stellar.phys.appstate.edu/   (through Online Spectra) where we
are releasing data from our spectroscopic survey of the nearby stars (
within ~120 l.y.) through NASA's Nstars project.  For more info, please
visit the web page.

If you would like more info on how this Perl CGI creates and serves the
GIF images with Ploticus, you can e-mail me at mcfaddnm@...

Thanks again!!!

-Michael "Flip" McFadden

PS.  Just today, our ascii "pixel list" files went through some personal
growth, and gained comment lines within the data lines.  I was so happy
to find the "commentchar" attribute to "getdata".  Again, a very robust
piece of software that conforms to MY data, not the other way around.
Thanks!

On Wed, 27 Jun 2001, Jason Burnett wrote:

> ok, that's working for me, but what's the formula for inches vs. screen
> resolution.  i.e. if I want a 320x200 image, how many inches is that?

The conversion factor is 100 pixels per inch, so that would be 3.2" x 2"

  -Steve

Stephen C. Grubb    scg@...
Scientific Software Engineer, The Jackson Laboratory
600 Main Street  Bar Harbor, Maine 04609 USA

Michael,

OK, now I see what you are asking.

You are using #proc curvefit to generate the curves.
#proc curvefit fits a smooth curve to the points.

You probably should use #proc lineplot if you want
your data points shown exactly.  See the attached
script.


Stephen C. Grubb    scg@...
Scientific Software Engineer, The Jackson Laboratory
600 Main Street  Bar Harbor, Maine 04609 USA
-------------------------------------------------------

Hi Steve,

Sorry about that, forgot to include more detail. The problem is that
the data contains plots at 22:13, 23:17 and 23:24, but these times
are not plotted on the graph.

BTW, I tried to download the graph you created but an error occurred
each time.

Regards,

Michael


--- In ploticus@y..., "Stephen C. Grubb" <scg@j...> wrote:
> Michael,
>
> I ran your example and it seems to be working correctly
> to me.  You are supplying data for 3 consecutive days,
> and drawing 3 curves using 3 data fields.
>
> The curve is drawn from the first point to the last.
>
> Attached is the gif I generated.
>
> Perhaps you could describe what you want, in more detail.
>
>   -Steve
>
>
> Stephen C. Grubb    scg@j...
> Scientific Software Engineer, The Jackson Laboratory
> 600 Main Street  Bar Harbor, Maine 04609 USA

Michael,

I ran your example and it seems to be working correctly
to me.  You are supplying data for 3 consecutive days,
and drawing 3 curves using 3 data fields.

The curve is drawn from the first point to the last.

Attached is the gif I generated.

Perhaps you could describe what you want, in more detail.

   -Steve


Stephen C. Grubb    scg@...
Scientific Software Engineer, The Jackson Laboratory
600 Main Street  Bar Harbor, Maine 04609 USA

On Wed, 27 Jun 2001, Jason Burnett wrote:

> Is there a way to force ploticus to spit out png images in a fixed
> size, no matter what the data?
>
> It'd make my life a lot easier if I could do that.


Jason,

   Suppose you want a fixed size of 4" wide by 3" high.
   You should be able to get this always, using pl ... -crop 0,0,4,3

   This will crop the result using absolute locations 0,0 (lower left)
   and 4,3 upper right.

   Big images may require -pagesize as well e.g. -pagesize 4,3

  -Steve



Stephen C. Grubb    scg@...
Scientific Software Engineer, The Jackson Laboratory
600 Main Street  Bar Harbor, Maine 04609 USA

A bug was reported where proc getdata was
being used to read fairly long data records,
and 'filter' was being used to process them.
The user was getting instability and garbage
result.  This occurred in 1.41 and earlier versions.

This was due to a buffer size problem in getdata.c

The fix is available at

http://www.sgpr.net/download/getdata.c

(drop the file into ploticus src directory
and do a 'make')

Thanks to Jason Burnett for tracking down this bug.

   -Steve

Hi,

I am having problems plotting the data file below (see transfer.dat
below). The x-axis contains dates (ie. 20/06/2001) and the y-axis
times (ie. 01:30). All the data is in time format (ie. 14:23). The
problem is that pl skews or sort of squashes or forces the plot
within a range that is not specified in my configuration script (see
my configuration script below). What am I doing wrong here? What I
require is to graph the times as they appear in the data file, except
the hyphens. Note, this also occurs when there are no missing times
in the data file.

The file transfer.dat contains the following data. Note, that a
hyphen indicates that there was no transfer for the specific file
type on that day... Thank you for any assistance.

Regards,

Michael

// transfer.dat
20/06/2001 08:05 08:14 08:35 -
21/06/2001 22:13 -     23:17 23:24
22/06/2001 09:12 09:28 09:40 09:59

The configuration file follows...

// My configuration file
#proc page
#if @DEVICE in png,gif
   scale: 0.9
   backgroundcolor: redorange
#endif

#proc getdata
file: filetrans_info_prototype.dat
fieldnames: date FileType1 FileType2 FileType3

#proc areadef
    title: Unison File Transmissions
    titledetails: size=16  align=L color=brightblue
    rectangle: 1 1 8 4
    xscaletype: date dd/mm/yyyy
    xrange: 21/05/2001 26/06/2001
    yscaletype: time hh:mm
    yrange: 00:00 24:00
    frame: color=black width=2.0

// do months using proc xaxis
#proc xaxis
   location: min-0.2
   stubs: increment 1 month
   stubformat: MMM
   stubdetails: adjust=0,+0.1 color=claret
   ticlen: 0.05
   axisline: none
   stubslide: 1(s)

// do days using proc xaxis
#proc xaxis
   stubs: none
   stubs: increment 1
   //stubrange: 010701
   stubrange: 21/05/2001 26/06/2001
   gridblocks: drabgreen drabgreen
   stubformat: dd
   stubdetails: size=6
   ticlen: 0.05
   label: Date in days
   labeldistance: 0.6
   labeldetails: color=oceanblue

#proc yaxis
   stubs: inc 1 hour
   //grid: color=gray(0.8)
   stubformat: hh:mm
   label: Time in hours
   labeldistance: 0.6
   labeldetails: color=oceanblue

#proc curvefit
xfield: date
yfield: FileType1
linedetails: color=yellow width=0.3
legendlabel: FileType1

#proc curvefit
xfield: date
yfield: FileType2
linedetails: color=red width=0.3
legendlabel: FileType2

#proc curvefit
xfield: date
yfield: FileType3
linedetails: color=brightblue width=0.3
legendlabel: FileType3

#proc legend
location: max+0.4 max
seglen: 0.2

Please be aware of a doc problem I just noticed.

In a number of places in ploticus,
double at-signs (@@) are used.  The are
used in 'select' attributes,
proc getdata 'filter', 'proc print', and other places.

In the docs, these mistakenly show up as single at-signs(@).
They should be double (@@).  The problem is mostly in example
code listings.  The problem occured because I now
produce these man pages using a scripting system.

Pages affected include the Faq, proc getdata, proc bars,
proc print, and condex.

I will put the corrected pages up as soon as I can.. sorry
for the confusion.

   -Steve

Apparently the sgpr.net web site hosting service
is having problems.. www.sgpr.net has been unavailable
all day yesterday (6/25) and now again today (6/26).
FTP has also been unavailable.

Sorry for the inconvenience.

I see that their stock share price is into the pennies..
maybe they are going under as we speak.  :>

   -Steve

Hi, Steve

First I'd like to thank you for developing such a powerful graphing package
that generates a lot of beautiful plots.
I've been using version 1.40 since late March and I just downloaded version
1.41 release 18 earlier this evening.  I am running Solaris 8 on a Sun
Ultra 5.  When I compiled the src,  everything went well except for the
following warning:

gcc -DSOLARIS2 -DPLOTICUS -DDIST -I/usr/openwin/include -c sleep.c -o sleep.o
sleep.c: In function `Eusleep':
sleep.c:27: warning: passing arg 1 of `nanosleep' from incompatible pointer type

I don't think the above warning would impact the binary in any way.  I just
include it as a FYI.

I use 'plpng' to plot a data file (testfile) into png graph.  In the second
graph where the yaxis crosses at 0, instead of '0', the stub became
'-5.55112e-17' which is very close to '0' of course, but I would love to
see a '0' printed there.  Is there anything I am doing wrong?  The command
I am using is:

	 plpng -png testfile.p -o testfile.png


(On your web site, I saw you sometimes said 'pl -png', so should I use
that or 'plpng -png'?)
I am enclosing 'testfile' and 'testfile.p'.  I would appreciate it very much
if you can help point me in the right direction.

As a sidenote, is it possible to overlay a ploticus graph on top of an
existing 'gif' or 'png' file?  One possibility would be something like:

	 #proc embed
	 file:  <existing png file>

	 #proc getdata
	 file:  <file to be plotted>

	 #proc areadef
	 .
	 .

Last but not least, thank you for a Wonderful product!

	 - Wing Moy (wingmoy@...)

Jan-01-2001 1455477 0 0 0
Jan-02-2001 1460550 5073 0.35 0.35
Jan-03-2001 1461274 724 0.05 0.40
Jan-04-2001 1465388 4114 0.28 0.68
Jan-05-2001 1470082 4694 0.32 1.00
Jan-06-2001 1473823 3741 0.25 1.25
Jan-07-2001 1479151 5328 0.36 1.62
Jan-08-2001 1484688 5537 0.37 1.99
Jan-09-2001 1484346 -342 -0.02 1.97
Jan-10-2001 1487781 3435 0.23 2.20
Jan-11-2001 1491209 3428 0.23 2.43
Jan-12-2001 1494108 2899 0.19 2.62
Jan-13-2001 1496320 2212 0.15 2.77
Jan-14-2001 1501575 5255 0.35 3.12
Jan-15-2001 1506461 4886 0.33 3.45
Jan-16-2001 1506210 -251 -0.02 3.43
Jan-17-2001 1510556 4346 0.29 3.72
Jan-18-2001 1514168 3612 0.24 3.96
Jan-19-2001 1517229 3061 0.20 4.16
Jan-20-2001 1520260 3031 0.20 4.36
Jan-21-2001 1526150 5890 0.39 4.75
Jan-22-2001 1531455 5305 0.35 5.10
Jan-23-2001 1528379 -3076 -0.20 4.90
Jan-24-2001 1532662 4283 0.28 5.18
Jan-25-2001 1536142 3480 0.23 5.40
Jan-26-2001 1539755 3613 0.24 5.64
Jan-27-2001 1543295 3540 0.23 5.87
Jan-28-2001 1549053 5758 0.37 6.24
Jan-29-2001 1554083 5030 0.32 6.57
Jan-30-2001 1554532 449 0.03 6.59
Jan-31-2001 1558581 4049 0.26 6.85
Feb-01-2001 1562116 3535 0.23 7.08
Feb-02-2001 1565813 3697 0.24 7.32
Feb-03-2001 1569045 3232 0.21 7.52
Feb-04-2001 1574787 5742 0.37 7.89
Feb-05-2001 1580505 5718 0.36 8.25
Feb-06-2001 1584072 3567 0.23 8.48
Feb-07-2001 1583536 -536 -0.03 8.45
Feb-08-2001 1586655 3119 0.20 8.64
Feb-09-2001 1589462 2807 0.18 8.82
Feb-10-2001 1592298 2836 0.18 9.00
Feb-11-2001 1597732 5434 0.34 9.34
Feb-12-2001 1603081 5349 0.33 9.67
Feb-13-2001 1603193 112 0.01 9.68
Feb-14-2001 1606562 3369 0.21 9.89
Feb-15-2001 1610064 3502 0.22 10.11
Feb-16-2001 1613236 3172 0.20 10.31
Feb-17-2001 1616664 3428 0.21 10.52
Feb-18-2001 1622442 5778 0.36 10.88
Feb-19-2001 1627591 5149 0.32 11.19
Feb-20-2001 1626819 -772 -0.05 11.15
Feb-21-2001 1632237 5418 0.33 11.48
Feb-22-2001 1637012 4775 0.29 11.77
Feb-23-2001 1640967 3955 0.24 12.01
Feb-24-2001 1644685 3718 0.23 12.24
Feb-25-2001 1651058 6373 0.39 12.63
Feb-26-2001 1656576 5518 0.33 12.96
Feb-27-2001 1661387 4811 0.29 13.25
Feb-28-2001 1662026 639 0.04 13.29
Mar-01-2001 1666282 4256 0.26 13.55
Mar-02-2001 1670689 4407 0.26 13.81
Mar-03-2001 1674323 3634 0.22 14.03
Mar-04-2001 1680157 5834 0.35 14.38
Mar-05-2001 1685670 5513 0.33 14.70
Mar-06-2001 1690773 5103 0.30 15.01
Mar-07-2001 1695032 4259 0.25 15.26
Mar-08-2001 1693418 -1614 -0.10 15.16
Mar-09-2001 1696783 3365 0.20 15.36
Mar-10-2001 1700296 3513 0.21 15.57
Mar-11-2001 1705964 5668 0.33 15.90
Mar-12-2001 1710966 5002 0.29 16.20
Mar-13-2001 1715518 4552 0.27 16.46
Mar-14-2001 1720594 5076 0.30 16.76
Mar-15-2001 1719203 -1391 -0.08 16.68
Mar-16-2001 1722773 3570 0.21 16.89
Mar-17-2001 1726755 3982 0.23 17.12
Mar-18-2001 1733362 6607 0.38 17.50
Mar-19-2001 1739204 5842 0.34 17.84
Mar-20-2001 1744009 4805 0.28 18.11
Mar-21-2001 1750076 6067 0.35 18.46
Mar-22-2001 1755135 5059 0.29 18.75
Mar-23-2001 1759205 4070 0.23 18.98
Mar-24-2001 1762575 3370 0.19 19.17
Mar-25-2001 1768342 5767 0.33 19.50
Mar-26-2001 1773852 5510 0.31 19.81
Mar-27-2001 1778404 4552 0.26 20.07
Mar-28-2001 1783260 4856 0.27 20.34
Mar-29-2001 1787748 4488 0.25 20.59
Mar-30-2001 1791527 3779 0.21 20.80
Mar-31-2001 1794790 3263 0.18 20.99
Apr-01-2001 1800592 5802 0.32 21.31
Apr-02-2001 1806448 5856 0.33 21.63
Apr-03-2001 1812716 6268 0.35 21.98
Apr-04-2001 1798631 -14085 -0.78 21.20
Apr-05-2001 1804512 5881 0.33 21.53
Apr-06-2001 1809026 4514 0.25 21.78
Apr-07-2001 1813237 4211 0.23 22.01
Apr-08-2001 1818908 5671 0.31 22.33
Apr-09-2001 1823139 4231 0.23 22.56
Apr-10-2001 1822627 -512 -0.03 22.53
Apr-11-2001 1826542 3915 0.21 22.75
Apr-12-2001 1829541 2999 0.16 22.91
Apr-13-2001 1832531 2990 0.16 23.07
Apr-14-2001 1832426 -105 -0.01 23.07
Apr-15-2001 1836961 4535 0.25 23.32
Apr-16-2001 1841354 4393 0.24 23.56
Apr-17-2001 1845307 3953 0.21 23.77
Apr-18-2001 1843358 -1949 -0.11 23.66
Apr-19-2001 1846389 3031 0.16 23.83
Apr-20-2001 1848394 2005 0.11 23.94
Apr-21-2001 1851166 2772 0.15 24.09
Apr-22-2001 1854997 3831 0.21 24.29
Apr-23-2001 1858633 3636 0.20 24.49
Apr-24-2001 1857838 -795 -0.04 24.45
Apr-25-2001 1861144 3306 0.18 24.63
Apr-26-2001 1863740 2596 0.14 24.77
Apr-27-2001 1866581 2841 0.15 24.92
Apr-28-2001 1869390 2809 0.15 25.07
Apr-29-2001 1873712 4322 0.23 25.30
Apr-30-2001 1877698 3986 0.21 25.51
May-01-2001 1882009 4311 0.23 25.74
May-02-2001 1880555 -1454 -0.08 25.66
May-03-2001 1884717 4162 0.22 25.89
May-04-2001 1888120 3403 0.18 26.07
May-05-2001 1891084 2964 0.16 26.22
May-06-2001 1895311 4227 0.22 26.45
May-07-2001 1899280 3969 0.21 26.66
May-08-2001 1897926 -1354 -0.07 26.58
May-09-2001 1900692 2766 0.15 26.73
May-10-2001 1903417 2725 0.14 26.87
May-11-2001 1905729 2312 0.12 27.00
May-12-2001 1907957 2228 0.12 27.11
May-13-2001 1911571 3614 0.19 27.30
May-14-2001 1915283 3712 0.19 27.50
May-15-2001 1914648 -635 -0.03 27.46
May-16-2001 1918550 3902 0.20 27.67
May-16-2001 1918550 0 0.00 27.67
May-17-2001 1922144 3594 0.19 27.85
May-18-2001 1925363 3219 0.17 28.02
May-19-2001 1924393 -970 -0.05 27.97
May-20-2001 1927924 3531 0.18 28.15
May-21-2001 1931438 3514 0.18 28.34
May-22-2001 1934704 3266 0.17 28.51
May-23-2001 1932767 -1937 -0.10 28.41
May-24-2001 1937672 4905 0.25 28.66
May-25-2001 1941753 4081 0.21 28.87
May-26-2001 1945580 3827 0.20 29.07
May-27-2001 1950763 5183 0.27 29.33
May-28-2001 1954845 4082 0.21 29.54
May-29-2001 1959126 4281 0.22 29.76
May-30-2001 1958318 -808 -0.04 29.72
May-31-2001 1962089 3771 0.19 29.91
Jun-01-2001 1965744 3655 0.19 30.10
Jun-02-2001 1968904 3160 0.16 30.26
Jun-03-2001 1973405 4501 0.23 30.49
Jun-04-2001 1977711 4306 0.22 30.71
Jun-05-2001 1982577 4866 0.25 30.95
Jun-06-2001 1986916 4339 0.22 31.17
Jun-07-2001 1990614 3698 0.19 31.36
Jun-08-2001 1988502 -2112 -0.11 31.25
Jun-09-2001 1992090 3588 0.18 31.43
Jun-10-2001 1996258 4168 0.21 31.64
Jun-11-2001 2000060 3802 0.19 31.83
Jun-12-2001 2004465 4405 0.22 32.05
Jun-13-2001 2003181 -1284 -0.06 31.99
Jun-14-2001 2006290 3109 0.16 32.14
Jun-15-2001 2009035 2745 0.14 32.28
Jun-16-2001 2011993 2958 0.15 32.43
Jun-17-2001 2013335 1342 0.07 32.49

#proc getdata
file: testfile

#proc areadef
//#clone: A
rectangle: 1 3 5 5
xscaletype: date mmm-dd-yyyy
xrange: Jan-01-2001 Dec-31-2001
yrange: -16000 10000
title:

#proc xaxis
ticincrement: 1 month
grid: color=skyblue
label: daily variation
axisline: none
tics: none
labeldetails: adjust=0,0.2 size=8

#proc yaxis
stubs: inc 2000
grid: color=skyblue

#proc bars
barwidth: 0.001
locfield: 1
lenfield: 3
thinbarline: color=green
select: @@3 > 0
crossover: 0.0
#saveas: B

#proc bars
#clone B
select: @@3 < 0
thinbarline: color=red

#proc line
notation: scaled
points: Jan-01-2001 0 Dec-31-2001 0

#proc areadef
//#clone: A
rectangle: 1 0.6 5 2.6
xscaletype: date mmm-dd-yyyy
xrange: Jan-01-2001 Dec-31-2001
yrange: -1.0 1.0
title:

#proc xaxis
ticincrement: 1 month
grid: color=skyblue
label: daily variation (%)
axisline: none
tics: none
labeldetails: adjust=0,0.2 size=8

#proc yaxis
stubs: inc 0.2
//stubrange:  0.2
grid: color=skyblue

#proc bars
barwidth: 0.001
locfield: 1
lenfield: 4
thinbarline: color=green
select: @@4 >= 0
crossover: 0
#saveas: C

#proc bars
#clone C
select: @@4 < 0
thinbarline: color=red

#proc line
notation: scaled
points: Jan-01-2001 0 Dec-31-2001 0

> There are some limitations with the command line.
> The command line has a length limitation of the arguments passed.

The command line args are placed into ploticus variables, which have a
documented max content length of 120 or so.  So that's the bottleneck.
Variables were intended for scalar parameters rather than big lists.

Sorry you had trouble getting your 'large' parameters into the script; I
hope to make improvements in this area soon.

> P.S. we managed to find a way to create real-time user specific custom
> web-based charts without using any temp files.

Explain?

  -Steve

Stephen C. Grubb    scg@...
Scientific Software Engineer, The Jackson Laboratory
600 Main Street  Bar Harbor, Maine 04609 USA

There are some limitations with the command line.
The command line has a length limitation of the arguments passed.

we were trying to pass on a list of numbers as a "-varg" argument (were talking
over 200 characters) to use in a sql statement and pl went ahead and cut off our
list at a certain number of characters.
we then attempted to grab the list using the #command from the getdata section
inside PL. no help there.
our only solution was to have the #command  run a shell script and a temporary
sql table contain our list  (which outputs the final dataset )

so in the next version, can you extend the maximum accepted memory size of the
command line arguments to at least a bajillion. Thanks


Jeff & Darryl @ BGI
P.S. we managed to find a way to create real-time user specific custom web-based
charts without using any temp files.

> As of today, ploticus is covered under GPL.

Awesome! That is very cool. My only reservations before about using ploticus
was it being non-GPL. (I used it anyways... Gnuplot can't compare for plotting
actual data.)

I also have something sort of neat you might want to put up on the examples
page:

http://mierle.com/keir/icqstats/

That is the generated statistics and graphs. The sources can be found at

http://mierle.com/keir/icqstat.tar.gz

I would suggest checking out the README for more information. (And bring
a sense of humor!)

Thanks in advance, Keir
--
Keir Mierle
keir@...
705.645.1711
UIN.7673934

Dear Stephen,

This e-mail is to express my enthousiasm about ploticus. As a matter of fact
I came to your webpage quite exidentally, I was "lured" there by your getgui
program but that didn't get much attention (I'm sorry) after I saw ploticus.

I don't know how long and with howmany people you've been working on this
but this is quite something and I'm surely going to use this frequently.

I've started my company something like half a year ago and decided that I
would be running Linux for a large part on an idological basis (If I had
done payed work in the extra time it cost me to setup everything under Linux
I would have erned enough money to buy the most expensive office suite ever
made). And it's paying off. The high "scriptability" is making my life a lot
easier and I can now produce impressive documents with hardly any effort
(well other than the actual contents of the document ofcourse).

What was missing was a good plot generation program. I don't want my
documents to look like "Oh, he's using gnuplot" so I had the intention to
write my own plot generation software in "off-time" to get out of the
gnuplot stage. This development is now arrested due to your great effort
called ploticus.

Moreover your software fits very nicely in the scriptability tradition as
you prove with the the mouse phenome project.

I am considering a "credits" section in my report layout, I haven't made a
decision about it yet, if so you'll be in it.

Again a very big thank you and keep up the good work.

Kind regards,


Jacco van Muiswinkel




--

-------------------->8---------------------->8----------
| Jacco van Muiswinkel          | Karpaten 62          |
| Van Muiswinkel consultancy    | 3424 DE Utrecht      |
| consultancy@...  | tel 030 - 267 10 17  |
| www.vanmuiswinkel.nl          | fax 030 - 267 10 18  |
-------------------->8---------------------->8----------

On Mon, 11 Jun 2001, [iso-8859-1] Peter T Garner wrote:

> What I need to know is, do you have any plans to make
> Ploticus a daemon ?  I have done some timing figures
> on my PC and the system("pl ..." ) takes approximately
> 0.5 seconds ( but it's only a Pentium 150/48Mb). I
> would think that most of this overhead is perl doing a
> system call and loading Ploticus. I reckoned that if
> Ploticus was a daemon, most of the overhead would be
> removed - this is pure theory of course :-)
>
> Even if it's not part of the Grand Design, it works
> really well- thanks for a great piece of software!


Thanks Peter, This was suggested awhile back by Tim Churches
(tchur@...) .. here is that dialogue.


Tim> One advantage Grace has is the ability
> to read input from a named pipe, as well as write to
> stdout, which means it can sit in
> memory waiting for commands.

I had considered this but not yet done it..  it would require improvement
of errorhandling (you can't just exit() when running in "server" mode) and
memory leaks could become an issue.. in general running in this mode
requires long term robustness that isn't as big a deal when pl exits after
every plot.  But, as you say, eventually.

Tim> Keeping ploticus in memory as a long-running
> process is not really an issue
> given the small size of the binary. Better to
> concentrate on new graphical features and
> speed optimisation when called as a one-shot process,
> and just give up and die in the
> face of an error. On our new linux server the binary is
> clearly cached permanently in
> memory so the start-up latency is really small anyway,
> particularly compared to how
> long it takes to pull the necessary data out of MySQL.


I do hope to implement this at some point.

Thanks for your interest in ploticus.

   -Steve


Stephen C. Grubb    scg@...
Scientific Software Engineer, The Jackson Laboratory
600 Main Street  Bar Harbor, Maine 04609 USA

Hello Steve !

I'm working towards the end of a proof-of-concept
project incorporating Ploticus 1.40. What happens is
that I have set up a perl script to act as an email
"bot". It takes a product number in the subject line
as a parameter, and then uses perl DBI to get results
from a MySQL database. The results are then input to
Ploticus which then produces a graph that is returned
as an email reply. Ploticus works really well in this
- I'm very impressed!

What I need to know is, do you have any plans to make
Ploticus a daemon ?  I have done some timing figures
on my PC and the system("pl ..." ) takes approximately
0.5 seconds ( but it's only a Pentium 150/48Mb). I
would think that most of this overhead is perl doing a
system call and loading Ploticus. I reckoned that if
Ploticus was a daemon, most of the overhead would be
removed - this is pure theory of course :-)

Even if it's not part of the Grand Design, it works
really well- thanks for a great piece of software!

Best wishes,

Peter Garner



____________________________________________________________
Do You Yahoo!?
Get your free @yahoo.co.uk address at http://mail.yahoo.co.uk
or your free @yahoo.ie address at http://mail.yahoo.ie

This is incorporated in 1.41

   -Steve

Stephen C. Grubb    scg@...
Scientific Software Engineer, The Jackson Laboratory
600 Main Street  Bar Harbor, Maine 04609 USA

Steve,

there was a bug in Ploticus 1.40 with ticks not being drawn at bars when
truncating was not switched on.
Please find a patch attached below.

Regards,
Michael


*** bars.c.pl140-2001-01-19 Mon Jun  4 19:16:46 2001
--- bars.c Mon Jun  4 19:53:01 2001
***************
*** 520,543 ****
  	 /* otherwise, render bar as a rectangle */
  	 else Ecblock( xleft, Ea( Y, y0 ), xright, Ea( Y, y ), color, doo );

!  /* do tics if requested */  /* don't do if trunc && outside area - scg
11/21/00 */
  	 if( leftticfld >= 0 ) {
  		 ytic = fda( i, leftticfld, axis );
! 	 if( !Econv_error() && trunc && Ef_inr( axis, ytic ) ) {
  			 Emov( (xleft+halfw), Ea(Y,ytic) );
  			 Elin( (xleft+halfw)-ticlen, Ea(Y,ytic) );
  			 }
  		 }
  	 if( rightticfld >= 0 ) {
  		 ytic = fda( i, rightticfld, axis );
! 	 if( !Econv_error() && trunc && Ef_inr( axis, ytic ) ) {
  			 Emov( (xleft+halfw), Ea(Y,ytic) );
  			 Elin( (xleft+halfw)+ticlen, Ea(Y,ytic) );
  			 }
  		 }
  	 if( midticfld >= 0 ) {
  		 ytic = fda( i, midticfld, axis );
! 	 if( !Econv_error() && trunc && Ef_inr( axis, ytic ) ) {
  			 Emov( (xleft+halfw)-(ticlen/2.0), Ea(Y,ytic) );
  			 Elin( (xleft+halfw)+(ticlen/2.0), Ea(Y,ytic) );
  			 }
--- 520,543 ----
  	 /* otherwise, render bar as a rectangle */
  	 else Ecblock( xleft, Ea( Y, y0 ), xright, Ea( Y, y ), color, doo );

!  /* do tics if requested */  /* don't do if trunc && outside area - scg
11/21/00 */ /* fixed it - mr 2001-06-04 */
  	 if( leftticfld >= 0 ) {
  		 ytic = fda( i, leftticfld, axis );
! 	 if( !Econv_error() && ( !trunc || Ef_inr( axis, ytic ) ) ) {
  			 Emov( (xleft+halfw), Ea(Y,ytic) );
  			 Elin( (xleft+halfw)-ticlen, Ea(Y,ytic) );
  			 }
  		 }
  	 if( rightticfld >= 0 ) {
  		 ytic = fda( i, rightticfld, axis );
! 	 if( !Econv_error() && ( !trunc || Ef_inr( axis, ytic ) ) ) {
  			 Emov( (xleft+halfw), Ea(Y,ytic) );
  			 Elin( (xleft+halfw)+ticlen, Ea(Y,ytic) );
  			 }
  		 }
  	 if( midticfld >= 0 ) {
  		 ytic = fda( i, midticfld, axis );
! 	 if( !Econv_error() && ( !trunc || Ef_inr( axis, ytic ) ) ) {
  			 Emov( (xleft+halfw)-(ticlen/2.0), Ea(Y,ytic) );
  			 Elin( (xleft+halfw)+(ticlen/2.0), Ea(Y,ytic) );
  			 }

Ken,
If the data for series 1 and series 2 are separate fields in
the same dataset, you can use (eg if series are in fields 3 and 5):

#proc areadef
   yautorange: datafields=3,5
   ..etc..

   -Steve

Stephen C. Grubb    scg@...
Scientific Software Engineer, The Jackson Laboratory
600 Main Street  Bar Harbor, Maine 04609 USA

Steve:

I'm absolutely loving Ploticus... it is pretty easy to pick up and quite
versitile, and it puts out nice looking graphs.

I am having one problem, however.  I am plotting two different series
on a line graph, the X axis is a date and time, and my Y axis is a
number.  The two series run over the same Y and X axis.

My problem occurs when trying to specify the Y axis range.  I can never be
sure which series will have the highest Y value.  My question is this: can
I tell the Y axis to make it's max value the max value of both series 1
and 2?

For instance, series 1 has a max of 10 and series 2 has a max of 2.  I
would want the axis to range to 10.  However, if it were reversed, I would
still want the series to range to 10, but it would get it from the second
series and not the first.

--Ken Speich

--
Ken Speich
Manager, Network Surveillance Tools
kspeich@... 		 301-598-0500 x2637

"How can you shoot the devil in the back? What if you miss?"

> Why not release Ploticus under the GPL?
>
> I understand your want to stop potential forking, but your current
> license seems rather opposed to the open source philosophy.


As of today, ploticus is covered under GPL.

Thanks for the suggestion.

  -Steve


Stephen C. Grubb    scg@...
Scientific Software Engineer, The Jackson Laboratory
600 Main Street  Bar Harbor, Maine 04609 USA

Why not release Ploticus under the GPL?

I understand your want to stop potential forking, but your current
license seems rather opposed to the open source philosophy.

Thanks in advance, Keir

--
Keir Mierle
keir@...
705.645.1711
UIN.7673934

Jason, the bug you reported should be fixed
in ploticus 1.41, released today.
   -Steve


On 23 May 2001, Jason Kwong wrote:

> I expect the tics along the x-axis to be in red and the stubs to be in green. 
However, when I plot this, both the tics and the stubs are in green.  Weirder
yet, if I turn on the grid for the x-axis, the tics and stubs both show up in
red!  It's a small problem, but it'll be great if it can be fixed.
>

Stephen C. Grubb    scg@...
Scientific Software Engineer, The Jackson Laboratory
600 Main Street  Bar Harbor, Maine 04609 USA

First off, I would like to say that Ploticus is an amazing plotting application.
I was fortunate enough to have found it early on in my search for a plotting
tool.  I love how much control I have over the appearance of a graph.  We'll be
using it in our current project (plotting stock data).  I have, however, ran
across one small problem.  I'm using v1.40 and I see a problem with the colour
of tics and stubs.  Here's a little sample script:

---------------

#proc getdata
    data:
       0 0
       1 1
       2 2

#proc areadef
    rectangle 1 1 5 5
    xscaletype: linear
    yscaletype: linear
    xautorange: datafield=1
    yautorange: datafield=2

#proc xaxis
    tics: color=red
    stubs: incremental
    stubdetails: color=green

---------------

I expect the tics along the x-axis to be in red and the stubs to be in green. 
However, when I plot this, both the tics and the stubs are in green.  Weirder
yet, if I turn on the grid for the x-axis, the tics and stubs both show up in
red!  It's a small problem, but it'll be great if it can be fixed.

Once again, a great plotting tool!  Thank you very much!

Jason Kwong
Toronto, Canada

On Fri, 1 Jun 2001, Jason Burnett wrote:
> I have a data file that looks like:
>
> hosta 2
> hosta 3
> hosta 1
> hosta 6
>
> and I want to do a bar graph where:
>
> hosta = 3
>
> in other words, average all the values in the file for each host (will be
> several), and graph that value.
>
> I've looked through the docs for proc processdata, but I don't see it...

One way to do this would be to use #proc rangebar with statsonly: yes
and meanmode: yes

This will set a variable called RANGEBARMEAN, which you can then plot.
The code would look something like this:

   #proc getdata
    file:  <your hosta file>

   #proc rangebar
    meanmode: yes
    statsonly: yes

   #proc getdata
    data:  @RANGEBARMEAN

   #proc bars
    ..etc..


> as a secondary wish, if I have 10 hosts in there, I'd like to only plot
> the 5 with the highest averages.

I can't think of a straightforward way to compute a collection of
averages in pl, then find the "top 5".  You could set up a loop and use
#proc processdata action: breaks with the above technique to iteratively
compute all averages and plot each.  But I don't know how you might save
the averages so that the top 5 would be selected.

This processing may need to be performed by another program before
invoking pl.


   -Steve

Stephen C. Grubb    scg@...
Scientific Software Engineer, The Jackson Laboratory
600 Main Street  Bar Harbor, Maine 04609 USA

howdy

I have a data file that looks like:

hosta 2
hosta 3
hosta 1
hosta 6

and I want to do a bar graph where:

hosta = 3

in other words, average all the values in the file for each host (will be
several), and graph that value.

I've looked through the docs for proc processdata, but I don't see it...

as a secondary wish, if I have 10 hosts in there, I'd like to only plot
the 5 with the highest averages.

any idea?

--
Jason Burnett
Wham Engineering and Software, Inc.

Welcome to this group.

You can post questions/comments/responses directly to this group.

I will forward all questions/comments sent to me (steve@...)
to this group.

   -Steve

Stephen C. Grubb    scg@...
Scientific Software Engineer, The Jackson Laboratory
600 Main Street  Bar Harbor, Maine 04609 USA