| Re: Object Graphics + Convolution with Point Spread Function [message #37772] |
Wed, 28 January 2004 08:18 |
David Fanning
Messages: 11724
Registered: August 2001
|
Senior Member |
|
|
Nuno Oliveira writes:
> By the way, you make you're your applications for a
> specific operating system or you code your programs so they can run either
> in Unix and Windows?
Within the constraints of IDL, and with God's grace,
my programs run everywhere.
Cheers,
David
P.S. Let's just say I wouldn't put too much faith in
the marketing literature, and I would test like a
banshee. :-)
--
David W. Fanning, Ph.D.
Fanning Software Consulting, Inc.
Phone: 970-221-0438, E-mail: david@dfanning.com
Coyote's Guide to IDL Programming: http://www.dfanning.com/
Toll-Free IDL Book Orders: 1-888-461-0155
|
|
|
|
| Re: Object Graphics + Convolution with Point Spread Function [message #37773 is a reply to message #37772] |
Wed, 28 January 2004 07:34  |
Nuno Oliveira
Messages: 75
Registered: October 2003
|
Member |
|
|
It is not complete decided but the question is that there is a strong
intention (and it is not my personal option or decision) to run it in
Virtual Machine in Unix. By the way, you make you're your applications for a
specific operating system or you code your programs so they can run either
in Unix and Windows?
"David Fanning" <david@dfanning.com> wrote in message
news:MPG.1a816ea4454a24db9897ba@news.frii.com...
> Nuno Oliveira writes:
>
>> So, if I want volume visualisation in Virtual Machine running in Unix
does
>> it mean that I have a big problem!!!
>
> It would appear so. Have you had a chat with the people
> you forked over all that money to? :-)
>
> Cheers,
>
> David
> --
> David W. Fanning, Ph.D.
> Fanning Software Consulting, Inc.
> Phone: 970-221-0438, E-mail: david@dfanning.com
> Coyote's Guide to IDL Programming: http://www.dfanning.com/
> Toll-Free IDL Book Orders: 1-888-461-0155
|
|
|
|
|
|
| Re: Object Graphics + Convolution with Point Spread Function [message #37775 is a reply to message #37774] |
Wed, 28 January 2004 03:32 |
Nuno Oliveira
Messages: 75
Registered: October 2003
|
Member |
|
|
So, if I want volume visualisation in Virtual Machine running in Unix does
it mean that I have a big problem!!!
"David Fanning" <david@dfanning.com> wrote in message
news:MPG.1a8019a4cc60c4a59897b9@news.frii.com...
> Nuno Oliveira writes:
>> When you say volume visualization it's easier with IDLgrVolume class
object.
>> is there another way?
>
> I wouldn't think so.
>
> Cheers,
>
> David
>
> --
> David W. Fanning, Ph.D.
> Fanning Software Consulting, Inc.
> Phone: 970-221-0438, E-mail: david@dfanning.com
> Coyote's Guide to IDL Programming: http://www.dfanning.com/
> Toll-Free IDL Book Orders: 1-888-461-0155
|
|
|
|
| Re: Object Graphics + Convolution with Point Spread Function [message #37784 is a reply to message #37775] |
Tue, 27 January 2004 05:46 |
David Fanning
Messages: 11724
Registered: August 2001
|
Senior Member |
|
|
Nuno Oliveira writes:
> Comes to me another few questions "object related". I'm used to build
> programs with functions: when a portion of code, with a limited number of
> variables, repeats in the program several times then I make a function. And
> what about the objects? When do you say "well this can be an class object
> defined with these class methods"?
Whenever I find myself passing info structures all over God's
green earth, the notion of an object appeals to me. That is to
say, nearly every time I write a widget program.
> And what are the requirements to build class objects and methods in IDL?
Open-mindedness, perseverance, a willingness to challenge the
IDL newsgroup received wisdom. A taste for adventure might help.
Basically the same things you need to be a successful learning
any foreign language. :-)
> When you say volume visualization it's easier with IDLgrVolume class object.
> is there another way?
I wouldn't think so.
Cheers,
David
--
David W. Fanning, Ph.D.
Fanning Software Consulting, Inc.
Phone: 970-221-0438, E-mail: david@dfanning.com
Coyote's Guide to IDL Programming: http://www.dfanning.com/
Toll-Free IDL Book Orders: 1-888-461-0155
|
|
|
|
| Re: Object Graphics + Convolution with Point Spread Function [message #37787 is a reply to message #37784] |
Tue, 27 January 2004 02:56 |
Nuno Oliveira
Messages: 75
Registered: October 2003
|
Member |
|
|
Comes to me another few questions "object related". I'm used to build
programs with functions: when a portion of code, with a limited number of
variables, repeats in the program several times then I make a function. And
what about the objects? When do you say "well this can be an class object
defined with these class methods"?
And what are the requirements to build class objects and methods in IDL?
When you say volume visualization it's easier with IDLgrVolume class object.
is there another way?
Cheers,
Nuno.
"David Fanning" <david@dfanning.com> wrote in message
news:MPG.1a7ef563ba4ea31e9897b7@news.frii.com...
> No, no. You misunderstand. I *love* object oriented programs.
> Nearly all the programs I write for clients are object oriented.
> They are slick, easy to maintain, easy to write (once you have
> a decent library of routines) and elegant.
>
> Of course, if you are doing anything at all in 3D space,
> object graphics are hard to beat.
>
> Cheers,
>
> David
> --
> David W. Fanning, Ph.D.
> Fanning Software Consulting, Inc.
> Phone: 970-221-0438, E-mail: david@dfanning.com
> Coyote's Guide to IDL Programming: http://www.dfanning.com/
> Toll-Free IDL Book Orders: 1-888-461-0155
|
|
|
|
| Re: Object Graphics + Convolution with Point Spread Function [message #37794 is a reply to message #37787] |
Mon, 26 January 2004 10:45 |
Paul Van Delst[1]
Messages: 1157
Registered: April 2002
|
Senior Member |
|
|
Nuno Oliveira wrote:
>
> "David Fanning" <david@dfanning.com> wrote in message
> news:MPG.1a7c38977949f04e9897b4@news.frii.com...
>> P.S. The three to one ratio is a bonus. The typical ratio
>> of object graphics code to direct graphics code is ten to one. :-)
>
> I get surprised with all these arguments. I still learning. but.
>
> I heard somewhere that object oriented languages where more easy and
> practical to use. Or you just don't like the way IDL's graphic objects are
> build?
In my experience, it depends on your definition of "easy" and "practical". And those
definitions can change for the same programmer from one project to another.
The object-graphic-y-ness or direct-graphic-y-ness of a particular visualisation tool in
IDL is something I care not a whit about. I just want to look at my data using PLOT,
SURFACE, CONTOUR, etc etc. and interactively muck about with the plot to prettify it. If
those tools are implemented using OG (which they bloody well should be) then it's
transparent to me -- and that's how I like it. If it takes more than, oh, say, 5-10
minutes to put something together to view said data (e.g. automatically reading in the
file format du jour), then I'm not interested.
iTools may be the solution, but I'm still using v5.5 and the nightmarish memories of the
"insight" tool from when IDL OG was first introduced has made me not so keen to upgrade
just for that. I'm amazed that the first thing that didn't happen when OG was introduced
was to transition the "regular" plotting tools (PLOT, SURFACE, CONTOUR, etc etc) to OG.
For things like, e.g. PLOT, cantankerous users such as myself would notice no difference
until we read the on-line help, or accidently clicked in the plot window, and saw all the
extra neato things available -- e.g. using the mouse to adjust axis and plot titles, or
legends (position, font, etc etc)...or clicking on a particular line in a series of
PLOT/OPLOTs so I could change the colour/PSYM of just that line (updating the legend entry
at the same time of course)... before we created a PostScript output file including the
changes. Matlab users have doing this sort of thing, "easily and practically", for about
10 years or so.
Apologies for the whinge.
paulv
--
Paul van Delst
CIMSS @ NOAA/NCEP/EMC
|
|
|
|
| Re: Object Graphics + Convolution with Point Spread Function [message #37795 is a reply to message #37794] |
Mon, 26 January 2004 08:59 |
David Fanning
Messages: 11724
Registered: August 2001
|
Senior Member |
|
|
Nuno Oliveira writes:
> I heard somewhere that object oriented languages where more easy and
> practical to use. Or you just don't like the way IDL's graphic objects are
> build?
No, no. You misunderstand. I *love* object oriented programs.
Nearly all the programs I write for clients are object oriented.
They are slick, easy to maintain, easy to write (once you have
a decent library of routines) and elegant.
But that doesn't mean they don't require a lot more code
up front. They do. The payoff is on the back end, once you
have a number of building blocks for creating functionality.
I have no objection at all to object graphics, except that
they are mostly overkill and unnecessary for the kinds of
programs my clients want. For me, direct graphics objects
are much simpler to use and write.
Of course, if you are doing anything at all in 3D space,
object graphics are hard to beat.
Cheers,
David
P.S. I have recently had occasion to see a 2D application
that was written in object graphics and I must admit, I was
greatly impressed with the look and feel. In fact, it was
spectacular! I couldn't come close to it it's look and feel
in direct graphics (at least until RSI gives us some decent
direct graphics fonts). But having some idea of the work
involved, I'm not sure the polish was worth the elbow grease. :-)
--
David W. Fanning, Ph.D.
Fanning Software Consulting, Inc.
Phone: 970-221-0438, E-mail: david@dfanning.com
Coyote's Guide to IDL Programming: http://www.dfanning.com/
Toll-Free IDL Book Orders: 1-888-461-0155
|
|
|
|
| Re: Object Graphics + Convolution with Point Spread Function [message #37797 is a reply to message #37795] |
Mon, 26 January 2004 07:43 |
Nuno Oliveira
Messages: 75
Registered: October 2003
|
Member |
|
|
"David Fanning" <david@dfanning.com> wrote in message
news:MPG.1a7c38977949f04e9897b4@news.frii.com...
> P.S. The three to one ratio is a bonus. The typical ratio
> of object graphics code to direct graphics code is ten to one. :-)
I get surprised with all these arguments. I still learning. but.
I heard somewhere that object oriented languages where more easy and
practical to use. Or you just don't like the way IDL's graphic objects are
build?
|
|
|
|
| Re: Object Graphics + Convolution with Point Spread Function [message #37804 is a reply to message #37797] |
Sat, 24 January 2004 07:09 |
David Fanning
Messages: 11724
Registered: August 2001
|
Senior Member |
|
|
Neil writes:
> Simulation of shapes viewed with optical systems normally culminates
> with the final image being convoluted with the point spread function
> of the optical equipment. This just blurrs the image, according to the
> laws of diffraction, thus accurately representing real effects in
> nature. This convolution is of course a one line operation in IDL and
> easy to implement in Direct Graphics.
>
> Does anyone know if there an equivalant equally effective way to do
> the convolution in Object Graphics?
>
> If this is possible it would significantly enhance scene simulation
> software for optical systems.
I guess I could be missing something here, but if this is
a one line solution in direct graphics, then it can't
possibly be more than a three line solution in object graphics.
In pseudo code:
1. Get the data out of the object. (GetProperty)
2. Apply the convolution. (Convol)
3. Put the data back into the object. (SetProperty)
Cheers,
David
P.S. The three to one ratio is a bonus. The typical ratio
of object graphics code to direct graphics code is ten to one. :-)
--
David W. Fanning, Ph.D.
Fanning Software Consulting, Inc.
Phone: 970-221-0438, E-mail: david@dfanning.com
Coyote's Guide to IDL Programming: http://www.dfanning.com/
Toll-Free IDL Book Orders: 1-888-461-0155
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
