[OT] OSS for dissertations (Was re: [vox-tech] tell gnuplot to use a single X11 window)
On Tue, May 24, 2005 at 09:14:42PM -0700, Dylan Beaudette wrote: I like your take on using only open source software, I am trying to do the same thing for my thesis in the field of landscape modeling and soil science. So far my professor and a couple others have become interested in OSS as well. Heh, be sure to send them our[*] way! ;^) -bill! [*] our as in LUGOD's ___ vox-tech mailing list vox-tech@lists.lugod.org http://lists.lugod.org/mailman/listinfo/vox-tech
Re: [vox-tech] tell gnuplot to use a single X11 window
On Tue 24 May 05, 9:14 PM, Dylan Beaudette [EMAIL PROTECTED] said: On May 23, 2005, at 3:37 PM, Peter Jay Salzman wrote: [...snip...] It's loaded with options. You can change the point size and even convert your graphs to eps files, suitable for including into your latex dissertation. ;-) Hope you find it useful. It was a lot of fun to write. Pete PS- I'm not the best Perl hacker, but I can always get the job done. Thanks again Pete. not only was the demo fun to watch, your plotter script with be handy for all sorts of stuff around the lab! Sure. If you can think of improvements, lemme know! I am not sure if I will understand much of your dissertation, but I would be interested in hearing about when it is done as well. I like your take on using only open source software, I am trying to do the same thing for my thesis in the field of landscape modeling and soil science. So far my professor and a couple others have become interested in OSS as well. That's cool; definitely keep at it! I lucked out: the physics department at UCD is firmly and utterly behind Linux. Professors that rely on Windows stick out like a sore thumb Pete -- Every theory is killed sooner or later, but if the theory has good in it, that good is embodied and continued in the next theory. -- Albert Einstein GPG Fingerprint: B9F1 6CF3 47C4 7CD8 D33E 70A9 A3B9 1945 67EA 951D ___ vox-tech mailing list vox-tech@lists.lugod.org http://lists.lugod.org/mailman/listinfo/vox-tech
Re: [vox-tech] tell gnuplot to use a single X11 window
On Monday 23 May 2005 15:37, Peter Jay Salzman wrote: any demos of the output that you mentioned? Thanks, Yeah, sure thing. Download it here: http://www.dirac.org/p/dylan.tar.bz2 Neat, definitively. Was this a part of your dissertation? Are you ever going to post your dissertation to your web site? Your web site looks like it hasn't been updated since you went east. Its ok to be modest but don't over do it. One of my instructors told me he was told he could not be a theoretical physicist because he wasn't arrogant enough. Richard Harke ___ vox-tech mailing list vox-tech@lists.lugod.org http://lists.lugod.org/mailman/listinfo/vox-tech
Re: [vox-tech] tell gnuplot to use a single X11 window
On Tue 24 May 05, 3:10 AM, Richard Harke @earthlink.net said: On Monday 23 May 2005 15:37, Peter Jay Salzman wrote: any demos of the output that you mentioned? Thanks, Yeah, sure thing. Download it here: http://www.dirac.org/p/dylan.tar.bz2 Neat, definitively. AFAIK, what you saw is the first numerical demonstration of quantum mechanical wavefunction collapse, and the first known numerical solution to the Schrodinger Newton quantum gravity scheme. Was this a part of your dissertation? Yeah, if you change was to is. I'm still writing, but hope to be done in about 2 weeks. Part of the problem is that I'm firmly, utterly, and completely a *theoretical* physicist, but my dissertation's main thesis is the proposal of an experiment to disprove a theory of quantum gravity. Unfortunately, experimentalists haven't been doing mass-on-spring or Young's double slit experiments for 2 or 3 hundred years. Modern experiments are considerably complicated, and have a theoretical framework that is very challenging. I know the gist of modern day matter-wave experiments, but I need to become an expert very quickly (if anybody really understands the Talbot effect, the Lau effect or Talbot-Lau interferometry, please email me!) One thing I've learned in grad school is respect for experimentalists. The types of theory a *real* experimentalist does (not a grad student performing someone else's experiment) requires theory that's often less abstract but more complicated than what a theoretician does. As an example of this, pick up any book on electricity and magnetism. Very easy to expand the Coulomb potential if you know the theory of expanding 1/|x-x'| in terms of orthogonal functions like spherical harmonics. But try to use the theory to calculate something useful, like the potential at the tip of a semi-spherical probe or even the electric field off-axis of a solenoid, and you have a problem that would make most theoreticians pee in their pants. Also, it didn't help that my main workstation died last week (hence, my recent postings on vox-tech). Luckily, I back everything up like a crazed maniac on steroids. I lost nothing. Absolutely nothing. Except time. Are you ever going to post your dissertation to your web site? Your web site looks like it hasn't been updated since you went east. That's pretty much true. I had a full time professorship at a local college teaching physics and calculus. You can't get away with the shenanigans that some professors do at UC Davis at community college. Try it, and you'll have a 90% fail rate. They definitely need a lot of hand holding because half of them are woefully unprepared for classes in math and physics. The metric that community colleges use to indicate success is graduation rates. High graduation rate == college effectiveness, so it behooves them to push as many people through as possible as fast as possible. It looks good for the college and they get loads of federal/state $$$ to do that. That also means when you teach calc, you have a class that couldn't graph y = mx + b without a graphing calculator. BTW, it's not just community colleges --- many high schools are also guilty of this. The punchline is -- I didn't have much time to work on my dissertation in the past year. I really tried, but it was beyond difficult. The job was great experience, and is good resume material. But was terrible for getting my thesis done. Luckily my position was eliminated, and I declined summer adjuncting, so I've been going full steam on my thesis. Losing the job was the best thing that ever happened to me. I definitely plan on posting the dissertation. I'll post an announcement. You'll probably hear me yell huzzah all the way from NJ. ;) I think the dissertation is also worth 2 or 3 physics papers, and perhaps another 2 or 3 papers in some kind of numerical research journal. Maybe computational fluids. I've been thinking that some Dr. Dobbs articles might be appropriate too. A lot of neat stuff went into solving this. And the whole thing was done, from start to finish, using nothing but open source tools. In fact, in my dissertation, I spend a few paragraphs explaining why it would be detrimental from a *technical view* to use certain proprietary tools. I'll post those papers too, including latex source. ;) Its ok to be modest but don't over do it. One of my instructors told me he was told he could not be a theoretical physicist because he wasn't arrogant enough. I've heard that before. I think busy often gets confused with arrogant. We normally talk with a daemon running in our head that filters or rearranges what comes out of our mouth. So instead of saying you are an idiot and have your head up your ass, most of us would say I think you're looking at the problem the wrong way. Most successful researchers don't seem to want that daemon taking CPU cycles. It's like running portmap when there's no RPC services.
Re: [vox-tech] tell gnuplot to use a single X11 window
On Tuesday 24 May 2005 05:35, Peter Jay Salzman wrote: AFAIK, what you saw is the first numerical demonstration of quantum mechanical wavefunction collapse, and the first known numerical solution to the Schrodinger Newton quantum gravity scheme. Awesome! Was this a part of your dissertation? Yeah, if you change was to is. I'm still writing, but hope to be done in about 2 weeks. Part of the problem is that I'm firmly, utterly, and completely a *theoretical* physicist, but my dissertation's main thesis is the proposal of an experiment to disprove a theory of quantum gravity. I'll want to read it for sure. Unfortunately, experimentalists haven't been doing mass-on-spring or Young's double slit experiments for 2 or 3 hundred years. Modern experiments are considerably complicated, and have a theoretical framework that is very challenging. I know the gist of modern day matter-wave experiments, but I need to become an expert very quickly (if anybody really understands the Talbot effect, the Lau effect or Talbot-Lau interferometry, please email me!) One thing I've learned in grad school is respect for experimentalists. The types of theory a *real* experimentalist does (not a grad student performing someone else's experiment) requires theory that's often less abstract but more complicated than what a theoretician does. As an example of this, pick up any book on electricity and magnetism. Very easy to expand the Coulomb potential if you know the theory of expanding 1/|x-x'| in terms of orthogonal functions like spherical harmonics. But try to use the theory to calculate something useful, like the potential at the tip of a semi-spherical probe or even the electric field off-axis of a solenoid, and you have a problem that would make most theoreticians pee in their pants. Completely right. Good experiments are very hard to do. Also, it didn't help that my main workstation died last week (hence, my recent postings on vox-tech). Luckily, I back everything up like a crazed maniac on steroids. I lost nothing. Absolutely nothing. Except time. Are you ever going to post your dissertation to your web site? Your web site looks like it hasn't been updated since you went east. That's pretty much true. I had a full time professorship at a local college teaching physics and calculus. You can't get away with the shenanigans that some professors do at UC Davis at community college. Try it, and you'll have a 90% fail rate. They definitely need a lot of hand holding because half of them are woefully unprepared for classes in math and physics. The metric that community colleges use to indicate success is graduation rates. High graduation rate == college effectiveness, so it behooves them to push as many people through as possible as fast as possible. It looks good for the college and they get loads of federal/state $$$ to do that. That also means when you teach calc, you have a class that couldn't graph y = mx + b without a graphing calculator. BTW, it's not just community colleges --- many high schools are also guilty of this. The punchline is -- I didn't have much time to work on my dissertation in the past year. I really tried, but it was beyond difficult. The job was great experience, and is good resume material. But was terrible for getting my thesis done. Luckily my position was eliminated, and I declined summer adjuncting, so I've been going full steam on my thesis. Losing the job was the best thing that ever happened to me. I definitely plan on posting the dissertation. I'll post an announcement. You'll probably hear me yell huzzah all the way from NJ. ;) I think the dissertation is also worth 2 or 3 physics papers, and perhaps another 2 or 3 papers in some kind of numerical research journal. Maybe computational fluids. I've been thinking that some Dr. Dobbs articles might be appropriate too. A lot of neat stuff went into solving this. And the whole thing was done, from start to finish, using nothing but open source tools. In fact, in my dissertation, I spend a few paragraphs explaining why it would be detrimental from a *technical view* to use certain proprietary tools. I'll post those papers too, including latex source. ;) Its ok to be modest but don't over do it. One of my instructors told me he was told he could not be a theoretical physicist because he wasn't arrogant enough. I've heard that before. I think busy often gets confused with arrogant. We normally talk with a daemon running in our head that filters or rearranges what comes out of our mouth. So instead of saying you are an idiot and have your head up your ass, most of us would say I think you're looking at the problem the wrong way. Most successful researchers don't seem to want that daemon taking CPU cycles. It's like running portmap when there's no RPC services. They want to end whatever interaction they're engaged in quickly
Re: [vox-tech] tell gnuplot to use a single X11 window
On May 23, 2005, at 3:37 PM, Peter Jay Salzman wrote: [...snip...] It's loaded with options. You can change the point size and even convert your graphs to eps files, suitable for including into your latex dissertation. ;-) Hope you find it useful. It was a lot of fun to write. Pete PS- I'm not the best Perl hacker, but I can always get the job done. Thanks again Pete. not only was the demo fun to watch, your plotter script with be handy for all sorts of stuff around the lab! I am not sure if I will understand much of your dissertation, but I would be interested in hearing about when it is done as well. I like your take on using only open source software, I am trying to do the same thing for my thesis in the field of landscape modeling and soil science. So far my professor and a couple others have become interested in OSS as well. cheers, -- Dylan Beaudette Soils and Biogeochemistry Graduate Group University of California at Davis 530.754.7341 ___ vox-tech mailing list vox-tech@lists.lugod.org http://lists.lugod.org/mailman/listinfo/vox-tech
[vox-tech] tell gnuplot to use a single X11 window
Hi everyone,
is there any good way to get gnuplot to send its output to a single window,
such that any successive plots update an existing X11 window?
i have looked over the gnuplot man page, and can't seem to find what i am
looking for.
while it is a little ridiculous, i like to keep track of my progress while
writting with this little script:
while `true`; do wc -w main.tex | awk '{print $1}' word_count ;\
echo set ylab 'words'; set xlab 'minutes'; \
plot 'word_count' with lines notitle | gnuplot -persist;\
sleep 60; done
this produces a nice little graph updated once a minute. however, gnuplot on
linux produces a new window with every plot. on OSX it is possible to keep
all output in a single window* .
* there is a bit of a difference on OSX, as an application called AquaTerm is
being used to render the output, not X11.
any ideas on how to do this with regular X11?
thanks!
--
Dylan Beaudette
Soils and Biogeochemistry Graduate Group
University of California at Davis
530.754.7341
___
vox-tech mailing list
vox-tech@lists.lugod.org
http://lists.lugod.org/mailman/listinfo/vox-tech
Re: [vox-tech] tell gnuplot to use a single X11 window
On Mon 23 May 05, 2:55 PM, Dylan Beaudette [EMAIL PROTECTED] said:
Hi everyone,
is there any good way to get gnuplot to send its output to a single window,
such that any successive plots update an existing X11 window?
i have looked over the gnuplot man page, and can't seem to find what i am
looking for.
while it is a little ridiculous, i like to keep track of my progress while
writting with this little script:
while `true`; do wc -w main.tex | awk '{print $1}' word_count ;\
echo set ylab 'words'; set xlab 'minutes'; \
plot 'word_count' with lines notitle | gnuplot -persist;\
sleep 60; done
this produces a nice little graph updated once a minute. however, gnuplot on
linux produces a new window with every plot. on OSX it is possible to keep
all output in a single window* .
* there is a bit of a difference on OSX, as an application called AquaTerm is
being used to render the output, not X11.
any ideas on how to do this with regular X11?
thanks!
I had the same problem. My fix was to write a small perl script to wrap
around gnuplot.
The script writes a small gnuplot script and then invokes gnuplot with the
script.
When it comes time for a new plot, the script rewrites the gnuplot script,
kills all instances of gnuplot, and re-runs gnuplot with the new gnuplot
script.
It works surprisingly well.
Pete
PS- The data is output of a time evolution partial differential equation.
It creates a movie of the time evolution of a quantum wavefunction. Very
cool.
--
Every theory is killed sooner or later, but if the theory has good in it,
that good is embodied and continued in the next theory. -- Albert Einstein
GPG Fingerprint: B9F1 6CF3 47C4 7CD8 D33E 70A9 A3B9 1945 67EA 951D
___
vox-tech mailing list
vox-tech@lists.lugod.org
http://lists.lugod.org/mailman/listinfo/vox-tech
Re: [vox-tech] tell gnuplot to use a single X11 window
On Monday 23 May 2005 03:03 pm, Peter Jay Salzman wrote:
On Mon 23 May 05, 2:55 PM, Dylan Beaudette [EMAIL PROTECTED] said:
Hi everyone,
is there any good way to get gnuplot to send its output to a single
window, such that any successive plots update an existing X11 window?
i have looked over the gnuplot man page, and can't seem to find what i am
looking for.
while it is a little ridiculous, i like to keep track of my progress
while writting with this little script:
while `true`; do wc -w main.tex | awk '{print $1}' word_count ;\
echo set ylab 'words'; set xlab 'minutes'; \
plot 'word_count' with lines notitle | gnuplot -persist;\
sleep 60; done
this produces a nice little graph updated once a minute. however, gnuplot
on linux produces a new window with every plot. on OSX it is possible to
keep all output in a single window* .
* there is a bit of a difference on OSX, as an application called
AquaTerm is being used to render the output, not X11.
any ideas on how to do this with regular X11?
thanks!
I had the same problem. My fix was to write a small perl script to wrap
around gnuplot.
The script writes a small gnuplot script and then invokes gnuplot with the
script.
When it comes time for a new plot, the script rewrites the gnuplot script,
kills all instances of gnuplot, and re-runs gnuplot with the new gnuplot
script.
It works surprisingly well.
Pete
PS- The data is output of a time evolution partial differential equation.
It creates a movie of the time evolution of a quantum wavefunction. Very
cool.
Cool!
thanks for the tip Pete. I will give it a try tonight (this thing is due in a
couple of hours!) when i have some time.
any demos of the output that you mentioned?
Thanks,
--
Dylan Beaudette
Soils and Biogeochemistry Graduate Group
University of California at Davis
530.754.7341
___
vox-tech mailing list
vox-tech@lists.lugod.org
http://lists.lugod.org/mailman/listinfo/vox-tech
Re: [vox-tech] tell gnuplot to use a single X11 window
On Mon 23 May 05, 3:18 PM, Dylan Beaudette [EMAIL PROTECTED] said:
On Monday 23 May 2005 03:03 pm, Peter Jay Salzman wrote:
On Mon 23 May 05, 2:55 PM, Dylan Beaudette [EMAIL PROTECTED] said:
Hi everyone,
is there any good way to get gnuplot to send its output to a single
window, such that any successive plots update an existing X11 window?
i have looked over the gnuplot man page, and can't seem to find what i am
looking for.
while it is a little ridiculous, i like to keep track of my progress
while writting with this little script:
while `true`; do wc -w main.tex | awk '{print $1}' word_count ;\
echo set ylab 'words'; set xlab 'minutes'; \
plot 'word_count' with lines notitle | gnuplot -persist;\
sleep 60; done
this produces a nice little graph updated once a minute. however, gnuplot
on linux produces a new window with every plot. on OSX it is possible to
keep all output in a single window* .
* there is a bit of a difference on OSX, as an application called
AquaTerm is being used to render the output, not X11.
any ideas on how to do this with regular X11?
thanks!
I had the same problem. My fix was to write a small perl script to wrap
around gnuplot.
The script writes a small gnuplot script and then invokes gnuplot with the
script.
When it comes time for a new plot, the script rewrites the gnuplot script,
kills all instances of gnuplot, and re-runs gnuplot with the new gnuplot
script.
It works surprisingly well.
Pete
PS- The data is output of a time evolution partial differential equation.
It creates a movie of the time evolution of a quantum wavefunction. Very
cool.
Cool!
thanks for the tip Pete. I will give it a try tonight (this thing is due in a
couple of hours!) when i have some time.
any demos of the output that you mentioned?
Thanks,
Yeah, sure thing. Download it here:
http://www.dirac.org/p/dylan.tar.bz2
My Perl wrapper is called plotter. You'll see it. You'll also see the
directory which is the output of my PDE solver. It's called data.22 (I
ran my PDE solver many many hundreds of times to cover a parameter space in
order to study the behavior of the PDE, so I resorted to uninspired names
for my data directory).
To see what options are available for my plotter, type:
./plotter --help
./plotter -h
./plotter -?
To see the plotter in action, try this:
./plotter -d data.22/ -x 1e-14 -y 4e14
The -x option sets the horizontal axis.
The -y option sets the vertical axis.
The -d option sets which directory to take the plot data from.
The -d option sets which directory to take the plot data from.
If the plot goes too fast, you can slow it down with the -p (pause) option,
which takes an integer argument:
./plotter -d data.22/ -x 1e-14 -y 4e14 -p1
The -x and -y options can have the form of:
-x from,to -x 2,100(x axis goes from 2 to 100)
-x ,to -x ,200 (x axis goes from 0 to 200)
-x from,-x 10, (x axis goes from 10 to last datapoint)
-x from -x 5(x axis goes from 5 to last datapoint)
It's loaded with options. You can change the point size and even convert
your graphs to eps files, suitable for including into your latex
dissertation. ;-)
Hope you find it useful. It was a lot of fun to write.
Pete
PS- I'm not the best Perl hacker, but I can always get the job done.
--
Every theory is killed sooner or later, but if the theory has good in it,
that good is embodied and continued in the next theory. -- Albert Einstein
GPG Fingerprint: B9F1 6CF3 47C4 7CD8 D33E 70A9 A3B9 1945 67EA 951D
___
vox-tech mailing list
vox-tech@lists.lugod.org
http://lists.lugod.org/mailman/listinfo/vox-tech
