| Re: Make the plots invisible [message #82005] |
Tue, 13 November 2012 08:41  |
![DavidF[1] is currently offline](theme/default/images/xoffline.png.pagespeed.ic.XRkd1fkXye.png "DavidF[1] is currently offline")
DavidF[1]
Messages: 94
Registered: April 2012
|
Member |
|
|
I wrote a few minutes ago:
> And, to call the documentation "abysmal" is to be WAY too generous.
Things could be worse. When I got to work this morning, my boss,
who is trying to learn the entirely new and revamped ENVI GUI interface,
points out that the "User Guide" (i.e., tutorial as to how to use
the new interface) is a whopping two and a half pages long!
Cheers,
David
|
|
|
|
| Re: Make the plots invisible [message #82007 is a reply to message #82005] |
Tue, 13 November 2012 08:08  |
David Fanning
Messages: 11724
Registered: August 2001
|
Senior Member |
|
|
Paul van Delst writes:
> Anyone who reads this newsgroup knows there are, uh, issues with certain
> aspects of creating graphics in NG (e.g. altogether now: COLORBAR! :o)
If the color bar doesn't work, the graphics system
doesn't work, as far as I'm concerned. I mean, come
on, we are trying to do science here!
I'm still not sure it is possible, two and a half
years after this system first came out, to compare
two images side-by-side.
I literally spent three hours yesterday, trying
to create a simple contour plot (which no one
has yet helped me with, probably because it is
impossible to create). So, clearly, these
things are not "intuitive". And, to call the
documentation "abysmal" is to be WAY too
generous.
I really don't know what to make of people telling
me these things "work well". What, exactly, are
they doing with the damn things!? It makes me
feel like an idiot when I can't get one simple
thing to work without literally spending hours and
hours trying to figure it out. You'd think it was
the remote control to the TV the way I struggle
with it! :-(
Cheers,
David
--
David Fanning, Ph.D.
Fanning Software Consulting, Inc.
Coyote's Guide to IDL Programming: http://www.dfanning.com/
Sepore ma de ni thue. ("Perhaps thou speakest truth.")
|
|
|
|
| Re: Make the plots invisible [message #82008 is a reply to message #82007] |
Tue, 13 November 2012 07:41 |
Paul Van Delst[1]
Messages: 1157
Registered: April 2002
|
Senior Member |
|
|
Hello,
I'd like to toss my $1.40 in ....
From my perspective, the concept of the IDL new graphics functions are
exactly what I was looking for. I can add (or remove) things to them as
I want, and the usual qualitative investigative things one wants to do
with a data plot (zoom, translate, rotate) tools are finally built in
(without having to resort to an iTool). [Ed: Hey, I can jump on the OO
bandwagon just as fast as the next guy...]
And they're relatively easy to use -- although the keyword usage could
be applied more consistently across function (e.g. vertical_alignment
and horizontal_alignment is used for some, whereas for others the latter
is referred to as just "alignment"). I think the IDL folks have
discovered the right path for a better user interface to their object
graphics engine.
Finally, the output just looks much nicer than anything I ever produced
with DG.
Anyone who reads this newsgroup knows there are, uh, issues with certain
aspects of creating graphics in NG (e.g. altogether now: COLORBAR! :o)
but I'm encouraged by the IDL folk's attention to these issues over the
last year or two (e.g. MarkP and ChrisT regularly reply to this
newsgroup, updates are more frequent addressing user concerns, etc).
But the big hairy 800 pound primate in the room is the real time
execution speed of NG -- which must be addressed to make NG be taken
seriously (and by that I mean for day-to-day data visualisation, not the
"produce the final graphic for publication" usage). Until the speed of
the NG plots rivals that of DG, I reckon a lot of folks aren't going to
use them. And people new to IDL will likely ditch it for something
faster. And FWIW, not all of us have the latest greatest hardware out
there - my nearly 3-year old desktop linux box was a low-to-medium-end
machine (i.e. relativly slow) when I got it.
It's tangential, but:
http://edition.cnn.com/2012/11/12/tech/web/video-loading-stu dy
I really hope the NG speed up in v8.2.2 is at least O(1) to bring it
somewhat closer to DG speeds.
Here endeth my monologue.
cheers,
paulv
On 11/12/12 15:47, Mark Piper wrote:
> On Sunday, November 11, 2012 1:23:52 PM UTC-7, David Fanning wrote:
>>
>> Totally amazing, considering it's a function graphics routine! ;-)
>>
>
> Hi David,
>
> In fact, NG work very well for me, and for most of the people I talk
> to. We continue to improve them with every IDL release. The biggest
> issue I've had is that plots with large numbers of points (roughly
> 1e6) are slow to render, but we've greatly improved this for 8.2.2,
> due in January.
>
> If you find problems with NG, please let us know. Constructive
> criticism will help us make them better.
>
> mp
|
|
|
|
|
|
|
|
|
|
| Re: Make the plots invisible [message #82026 is a reply to message #82025] |
Mon, 12 November 2012 13:16 |
DavidF[1]
Messages: 94
Registered: April 2012
|
Member |
|
|
Mark Piper writes:
> In fact, NG work very well for me, and for most of the people I talk to. We continue to improve them with every IDL release. The biggest issue I've had is that plots with large numbers of points (roughly 1e6) are slow to render, but we've greatly improved this for 8.2.2, due in January.
I agree they continue to improve with every release, but I am still too
gun-shy to adopt them for serious work. Sounds like we talk to different
people about function graphics. I run into few people who use them,
and those that do almost uniformly complain about how slow they are.
But, I hear fewer complaints about how they appear on the display. Most of
the problems now appear to be in creating file output from them. Lines
that suddenly grow to 10 times their normal size, and that kind of thing.
I should probably spend some time updating this page on contour
plots for IDL 8.2.1. I'd be curious to see how much things have improved
in both functionality and speed:
http://www.idlcoyote.com/cg_tips/compcont.php
Cheers,
David
|
|
|
|
| Re: Make the plots invisible [message #82027 is a reply to message #82026] |
Mon, 12 November 2012 12:47 |
Mark Piper
Messages: 198
Registered: December 2009
|
Senior Member |
|
|
On Sunday, November 11, 2012 1:23:52 PM UTC-7, David Fanning wrote:
>
> Totally amazing, considering it's a function graphics routine! ;-)
>
Hi David,
In fact, NG work very well for me, and for most of the people I talk to. We continue to improve them with every IDL release. The biggest issue I've had is that plots with large numbers of points (roughly 1e6) are slow to render, but we've greatly improved this for 8.2.2, due in January.
If you find problems with NG, please let us know. Constructive criticism will help us make them better.
mp
|
|
|
|
|
|
| Re: Make the plots invisible [message #82041 is a reply to message #82040] |
Sun, 11 November 2012 11:48 |
d.poreh
Messages: 406
Registered: October 2007
|
Senior Member |
|
|
On Sunday, November 11, 2012 6:23:40 AM UTC-8, David Fanning wrote:
> dave poreh writes:
>
>
>
>> I want to make some plots and make the plots invisible, that means just i want to save the plots like this:
>
>>
>
>> p = plot(x(*,i),color='blue',thick=2)
>
>>
>
>> p.save, 'plot1.jpg'
>
>> but without a window pops up,
>
>
>
> Set the BUFFER keyword.
>
>
>
> Cheers,
>
>
>
> David
>
>
>
>
>
>
>
> --
>
> David Fanning, Ph.D.
>
> Fanning Software Consulting, Inc.
>
> Coyote's Guide to IDL Programming: http://www.dfanning.com/
>
> Sepore ma de ni thue. ("Perhaps thou speakest truth.")
Cool, i just put BUFFER=1 and it works,
Thank you very much,
Cheers,
Dave
|
|
|
|
| Re: Make the plots invisible [message #82043 is a reply to message #82041] |
Sun, 11 November 2012 06:23 |
David Fanning
Messages: 11724
Registered: August 2001
|
Senior Member |
|
|
dave poreh writes:
> I want to make some plots and make the plots invisible, that means just i want to save the plots like this:
>
> p = plot(x(*,i),color='blue',thick=2)
>
> p.save, 'plot1.jpg'
> but without a window pops up,
Set the BUFFER keyword.
Cheers,
David
--
David Fanning, Ph.D.
Fanning Software Consulting, Inc.
Coyote's Guide to IDL Programming: http://www.dfanning.com/
Sepore ma de ni thue. ("Perhaps thou speakest truth.")
|
|
|
|
| Re: Make the plots invisible [message #82092 is a reply to message #82014] |
Wed, 14 November 2012 09:19 |
David Fanning
Messages: 11724
Registered: August 2001
|
Senior Member |
|
|
David Fanning writes:
>> I should probably spend some time updating this page on contour
>> plots for IDL 8.2.1. I'd be curious to see how much things have improved
>> in both functionality and speed:
>>
>> http://www.idlcoyote.com/cg_tips/compcont.php
>
> Ok, I give up! I spent a couple of hours trying to
> get this plot to work in IDL 8.2.1, and there is
> no way I can get the damn colorbar to work! I have
> NO idea what I am doing wrong. I am using code
> straight out of the documentation and I don't
> EVER see a taper on my color bar.
>
> Can anyone help? Just replicate the contour
> plot shown in that article in IDL 8.2.1.
Seriously? No help whatsoever!?
OK, then I think three conclusions are possible.
1. Despite claims to the contrary, no one outside
of Exelis uses function graphics.
2. If people do use function graphics, they can't
figure out how to do this simple contour plot
example with them.
3. The example is so trivial that no one wants to
embarrass me further by pointing out my obvious
inability to understand the very simplest things.
If conclusion 3 is the true conclusion, then I
urge you to come forward anyway. I'm going to
blame everything on Coyote, of course, so no permanent
damage can be done. ;-)
Here is what I have so far:
Code: http://www.idlcoyote.com/misc/compare_contour_plots.pro
Pic: http://www.idlcoyote.com/misc/idl_8.2.1_contour.png
Issues to resolve:
1. Contour plot doesn't honor POSITION keyword.
2. Contour plot doesn't use correct colors.
3. Only half of box axes in place.
4. Color bar has correct colors, but no taper.
5. Color bar not labeled correctly.
Cheers,
David
--
David Fanning, Ph.D.
Fanning Software Consulting, Inc.
Coyote's Guide to IDL Programming: http://www.dfanning.com/
Sepore ma de ni thue. ("Perhaps thou speakest truth.")
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
