| Re: PostScript and IDL, why object graphics ? [message #11870] |
Thu, 28 May 1998 00:00 |
Martin Schultz
Messages: 515
Registered: August 1997
|
Senior Member |
|
|
others wrote [...]
Thanks for this discussion!! Just two days ago I sneeked into one of
these "technical seminars" that RSI gives these days, and I must say I
was really amazed by the live_tool and object graphics capabilities.
*BUT* in my work I completely depend on being able to produce
publishable figures (and even with direct graphics this is often a major
step after having a the data on the screen, especially when maps are
involved). One aspect that is too often forgotten in all this object
oriented stuff is the "reproducibility" of a plot. E.g. I make a plot of
some preliminary data, and of course it is nice to optimize this plot
with the help of a few mouse clicks, i.e. in object graphics. But then I
want to reproduce the exact same plot with the final data that comes in
a few weeks/months later. With conventional direct graphics, I only have
to rerun my program and that's it. With object graphics, after rerunning
my program I would have to go through all this clicking again, and it is
very likely that I forget a few important clicks. And this situation
becomes even worse when I have to process not only one data set but
order 100. Then I don't want one single click in order to prevent RSI
("repetative stress syndrome", a big issue here at Harvard ;-). Now, I
must admit, I haven't really used IDL object graphics too much, but my
experience from EXCEL and others which presumably have a similar
underlying concept warns me. I guess one solution would be to have some
kind of program generator from object graphics. If I understand the live
tools correctly, one can use them in the same manner as e.g. the plot
command. If there were a way to generate a program from the final plot
(this would generate a call to live_plot with all the view parameters
that are necessary to reproduce exactly the same graph: scaling, line
thicknesses, types, symbols, etc.) it would help a lot -- but of course
only if there was also a true ps output capability.
As far as WYSIWYG is concerned, I think the direct graphics routines
could also be improved. One solution would be to have a "screenps"
device which would produce the plot exactly as on the postscript page
but with reduced resolution (yes, this is what ghostview does). You
might not be able to read all your labels any longer, but at least you
could position them correctly without having to
set_plot,'ps',file='bla.ps'
myprogram
set_plot,'X'
[unix prompt]gs bla.ps
n times.
Martin
PS: (don't read if you love your mouse) what's upsetting me about object
graphics (or better its interface with the user) is also what upsets me
about windows and why I like unix so much. Instead of
foreach name ( thisplot[abc]*.ps thatplot*3*.ps )
ps2epsi $name
end
(with csh under unix) you need zillions of mouseclicks and ctrl key
strokes to do the same in windows. This example may serve to illustrate
more clearly what I meant above.
------------------------------------------------------------ -------
Dr. Martin Schultz
Department for Earth&Planetary Sciences, Harvard University
109 Pierce Hall, 29 Oxford St., Cambridge, MA-02138, USA
phone: (617)-496-8318
fax : (617)-495-4551
e-mail: mgs@io.harvard.edu
Internet-homepage: http://www-as.harvard.edu/people/staff/mgs/
------------------------------------------------------------ -------
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
