| GUI Builder limited to Windows platform (LONG) [message #13329] |
Tue, 10 November 1998 00:00  |
David Foster
Messages: 341
Registered: January 1996
|
Senior Member |
|
|
I wanted to post a response I received from Mark Goosman, the
IDL Product Manager, about my concern that the new GUI Builder is
meant only for the Windows platforms. As far as I can tell, they
have no plans on extending this functionality to other platforms.
For a company that makes cross-platform compatibility one of its
main selling points, I think this is inexcusable. Does anyone else
agree?
I probably wouldn't even use the Builder myself, but I know a lot
of groups that are just beginning to develop widget programs, and
this tool would be invaluable to them.
Anyway, here is Mark's response. I've prefaced my comments by %
(pun intended). First, I want to thank Mark for taking the
time to send me a very thorough response.
====== Response from RSI's Mark Goosman, IDL Product Manager =======
Dave:
Your Email to Jennifer Lodder was forwarded on to me. We are a small
enough company that Email to just about any account finds it's way to a
rightful recipient.
I understand, and to a certain extent, share your comments regarding the
IDL GUIBuilder being limited to the Microsoft Windows platforms. There
are several issues that went into this decisions.
The primary decision was the additional effort required to support the
IDL GUIBuilder on all supported IDL platforms. This would have required
development and testing on each of the Motif versions of Sun, HP-UX,
SGI, IBM/AIX, DEC Unix, in addition to the Macintosh. We also have some
very verbal OpenVMS customers who expect the same level of support. This
level of resource is beyond the scope of Research Systems and would have
prevented us from releasing IDL 5.2 in a reasonable time frame. We also
had a large number of other major features for IDL 5.2 that could not
have been implemented had we not limited the IDL GUIBuilder to the
Windows platform. I'm not sure if you've seen the total list of new
features for IDL 5.2 but we've added several new data types (16 bit
unsigned, 64 bit signed, and 64 bit unsigned), support for several new
data file formats (HDF-EOS, DICOM, PNG, DXF, and others), very large
files, and several other features requested by our user community.
% So it really boils down to a question of priority. Either get
% version 5.2 released, or spend resources implementing the Builder
% on other platforms. Tough decision. My first question would be: If
% supplying a GUI Builder for the platforms that use IDL is "beyond
% the scope of Research Systems" then why was the Builder created
% for the Windows platforms in the first place? And my next question
% would be: How many other platform specific tools (and perhaps
% upgrades!) will we have to look forward to?
We are finding that most Unix users of IDL do have access to Windows
machines. The recent addition of FlexLM support to IDL for Windows means
that anyone with an Unix floating license can set up IDL for access on
their Windows platforms. This offers a solution to a large percentage of
our Unix customer base. One major goal of the IDL GUIBuilder was that
the generated code be cross-platform compatible.
% No. I would agree that most Unix users have access to Windows
% machines, but it is not true that most Unix users have floating
% license; at least not here at UCSD, where this is definitely
% not true. For a group such as ours, floating point licenses
% are much more expensive, so upgrading the licenses would not
% be an acceptable option.
One other major aspect of the decision was the demand for GUI
development tools on the Windows platform compared to the same market
for Unix. There are several major companies offering these types of
tools (Microsoft Visual Basic, Symantec Cafe, etc). There are, however,
fewer products for Unix users reflecting less of a demand.
% If this is the kind of reasoning that RSI's marketing used to
% reach these decisions then we are really in trouble! The demand
% for GUI builders for Visual Basic, C and Pascal reflect a need
% for such tools by those creating commercial applications that
% they intend to market. My impression is that most of the users
% of IDL are writing in-house applications for their own research
% purposes. The software is to be used locally, and not distributed.
% There is absolutely no reason to assume that those users of IDL
% using Windows machines have any more need for a GUI Builder than
% those who use Unix machines.
%
% So as I now understand it, what it *really* boils down to is
% "market demand".
Any decision on new features and functionality in IDL is based on
providing the best solution to the IDL user community. I hope it is
apparent that RSI makes every effort to prioritize resources in a way
that makes IDL most effective in solving the needs of our users and
customers.
% Don't get me wrong. I love IDL, I would be a great salesman for
% it. But it concerns me when I see a company begin to favor support
% for one platform over another, based on an analysis of "market
% demands". This is the first time I have seen this with RSI. I
% have seen this many times with other companies, and the results
% are always unfortunate for those who don't happen to be using the
% most common platform/hardware/software/etc.. If RSI decides to
% support a particular platform, then SUPPORT it!
%
% Thank-you for your valuable time; I really appreciate your
% very thoughtful response.
%
% Dave
Please let me know if you have any questions or additional comments.
Best regards,
Mark Goosman
********************************************
Mark Goosman
IDL Product Manager
Research Systems, Inc.
4990 Pearl East Circle
Boulder, CO 80301 USA
Tel: 303-413-3966
Fax: 303-786-9909
Email: mgoosman@rsinc.com
WWW: http://www.rsinc.com
*************************************
--
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~
David S. Foster Univ. of California, San Diego
Programmer/Analyst Brain Image Analysis Laboratory
foster@bial1.ucsd.edu Department of Psychiatry
(619) 622-5892 8950 Via La Jolla Drive, Suite 2240
La Jolla, CA 92037
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~
|
|
|
|
| Re: GUI Builder limited to Windows platform (LONG) [message #13440 is a reply to message #13329] |
Thu, 12 November 1998 00:00  |
mirko_vukovic
Messages: 50
Registered: January 1998
|
Member |
|
|
In article <Y8v22.7529$q15.230893@news.san.rr.com>,
"Pete Riley" <uk2@scibernet.com> wrote:
> Hi Dave,
>
stuff deleted
> But, if UNIX isn't
> paying the bills, and you've got a 450 MHz Pentium II with 1/2 a gig of
> memory and 50 gigs of space then you rule the world...at least in my little
> mind.
>
> -Pete Riley
>
I was unpleasantly surprised when I was not able to get a piece of high-end
electro-magnetic software for a DEC Alpha. It was made for WIntel, and other
Unixes, but not for DEC. Similarly, the software we use for 3D magnetic
modeling is first released for WIntel and only then for the Unix boxes.
Puzzled, frustrated, I inquired. And the answer was simple enough, as if
they were quoting Mr. Gates from his Economist article (a couple of months
back). Any PC running WIntel will run their codes. But each Unix vendor's
Unix is different, and requires additional programming effort.
Thus in some sense, the Unix machine vendors have placed themselves in that
unenviable position, with all of their best versions of Unix. Thus, for
commercial written software, WIntel may be the platform of choice, and the
workstations are left to internally written software.
Possybly the only thing that can bring Unix back to favor with the software
developers is Java. At a recent Plasma physics conference (Maui, Hawaii :-))
one company was showing a modeling code written in C and the front end in
Java. Actually, written in F90, translated and compiled into C.
My two breaths
Mirko
-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
|
|
|
|
|
|
| Re: GUI Builder limited to Windows platform (LONG) [message #13450 is a reply to message #13329] |
Wed, 11 November 1998 00:00 |
Pete Riley
Messages: 12
Registered: June 1998
|
Junior Member |
|
|
Hi Dave,
The GUI builder is slick, but it doesn't really save you any time
programming (as long as you can type more than 10 words a minute). It can't
hide the nuts and bolts...only make them look like something else. In a
previous incarnation, WIDED (which is certainly not a fair comparison,
considering how much work has gone into the GUI builder), I would use a lot
to get a quick look at how my widgets would appear on the screen. But, I
would still go back to simple programming for the final product. (A bug in
the READF in the beta version made me stop using 5.2b regularly, but I
suspect that I'll still continue to write from scratch even with the
corrected release).
On your remarks about platform independence, when I first used IDL (version
1) it only came in VMS flavour ....not very platform independent! And since
RSI is a company, they surely must pitch for 'the market'.....after all,
they're not gn-doing it for free. I don't believe they ever supported BE?
And, strictly speaking IDL is not really platform independent - they only
provide versions for most of the common operating systems. Moreover, you
know that any serious code must be 'nudged' to run on different platforms
(I'm particularly annoyed at the moment by having to try to get rid of a
bunch of DOS calls in an old programme I'm updating). I would not be
surprised if VMS was dropped in the near future (but hopefully NASA/Goddard
is loud?). If Jobs doesn't keep the iiiiiiiiiiiii-mprovements to the Mac,
then that could quite easily fall by the wayside. Neither, of course, would
be a big deal if you're not using either OS (and at least 95% (depending on
how you define 'users') aren't). What is somewhat disturbing is the
transition from UNIX as being the mainstay of IDL to Windows, which the
introduction of the GUI builder is obviously signalling. But, if UNIX isn't
paying the bills, and you've got a 450 MHz Pentium II with 1/2 a gig of
memory and 50 gigs of space then you rule the world...at least in my little
mind.
-Pete Riley
|
|
|
|
| Re: GUI Builder limited to Windows platform (LONG) [message #13457 is a reply to message #13329] |
Wed, 11 November 1998 00:00 |
David Foster
Messages: 341
Registered: January 1996
|
Senior Member |
|
|
David Foster wrote:
>
> I wanted to post a response I received from Mark Goosman, the
> IDL Product Manager, about my concern that the new GUI Builder is
> meant only for the Windows platforms. As far as I can tell, they
> have no plans on extending this functionality to other platforms.
>
<long response snipped>
I wanted to follow this up with a sincere apology to Mark Goosman.
It is important that everyone know that his reply to me was a
personal communication, and he did not intend to have it posted
for the world to see.
After a discussion with him I realize that I should have responded
to him personally, and then perhaps posted the results of this
exchange. Posting his reply without his permission was a serious
breach of netiquette and I apologize.
After talking with Mark I have a better idea of how hard it is for
RSI to prioritize development efforts given limited resources. And
to be fair, I am much happier having support for Dicom, unsigned
integers, etc. etc. in IDL 5.2 than I would have been with a
GUI builder for my Unix box.
Sincerely,
David Foster
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~
David S. Foster Univ. of California, San Diego
Programmer/Analyst Brain Image Analysis Laboratory
foster@bial1.ucsd.edu Department of Psychiatry
(619) 622-5892 8950 Via La Jolla Drive, Suite 2240
La Jolla, CA 92037
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~
|
|
|
|
| Re: GUI Builder limited to Windows platform (LONG) [message #13481 is a reply to message #13329] |
Tue, 17 November 1998 00:00 |
mirko_vukovic
Messages: 50
Registered: January 1998
|
Member |
|
|
In article <72p19k$jna$4@gwdu19.gwdg.de>,
pit@uni-sw.gwdg.de wrote:
> In article <72evlf$116$1@nnrp1.dejanews.com>,
> mirko_vukovic@notes.mrc.sony.com writes:
>
>> Puzzled, frustrated, I inquired. And the answer was simple enough, as
>> if they were quoting Mr. Gates from his Economist article (a couple of
>> months back). Any PC running WIntel will run their codes. But each
>> Unix vendor's Unix is different, and requires additional programming
>> effort.
>
>> Thus in some sense, the Unix machine vendors have placed themselves in
>> that unenviable position, with all of their best versions of Unix.
>> Thus, for commercial written software, WIntel may be the platform of
>> choice, and the workstations are left to internally written software.
>
> And, shortly speaking, it is almost complete nonsense. Almost any
> decent Unix program compiles on many flavours of Unix. There is more
> than enough Software to run on, e.g. Dec, HP, AIX, SGI and Intel-Unix,
> where you don't get a Win-version.
>
> In general, if you have one unix-version, porting it to another flavour
> of unix is very little effort. If I remember correctly, a programmer of
> Informix wrote about the port to Linux: "Basically, it was just a
> 'make'.."
>
> Peter
>
From my (limited) contacts with commercial scientific/engineering software
developers, even C code is only "close to fully compatible" between various
vendors. I presume that this is not a Herculean task in itself, but the
testing probably is.
In addition, I was told, the GUI's are not fully compatible and thus
need additional effort.
Respectfully,
Mirko Vukovic
-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
|
|
|
|
| Re: RSI's priorities [message #13487 is a reply to message #13448] |
Mon, 16 November 1998 00:00 |
krist
Messages: 6
Registered: April 1994
|
Junior Member |
|
|
Just a few thoughts while I'm waiting for an IDL program to finish executing:
I've been using IDL since 1986, when it was at V1.0 and graphics output meant
a Tektronix terminal and an HP plotter (I had to write my own workstation
display and Postscript drivers, before RSI came out with them). I have used
IDL as my primary programming and data analysis environment since then (even
during the dark days at work where using IDL to do something was essentially
forbidden because of all of the development effort in IRAF/STSDAS).
I have developed something of a love/hate relationship with IDL over the years.
I love the ability to work directly and interactively with my data and do
things with one line of code that would take many in other languages; I can
appreciate the power and simple beauty of the WHERE function; and I can
respect the utility of widgets.
However, some things drive me crazy. It seems with every major release, much
of my code and the code in libraries I use (like the Astronomy User's Library)
breaks because RSI changed a keyword, changed the functionality of a procedure,
or altered the behavior of some feature. The utmost worst time I ever had
with IDL was with widgets. It got to the point that every minor release
changed the way widgets behaved, breaking code right and left. It seems to me
the same has been happening with objects. They were implemented before they
were really well thought out. I would never recommend IDL for mission critical
work unless its use was frozen to a specific version. It's just not safely
backwards compatible.
Cost is also a factor, and it hurts the further widespread use of IDL in more
than one way. I myself cannot afford the $1895 for a Windows or Linux license
(yes people, the price is going up after January 1, not down), and that doesn't
factor in the cost of maintenance contracts. I know more than one place where
IDL is still at version 4 or so, because the cost is just too high. And of
course, when people start using new features like objects, these places are
left out, unable to use other's code, futher souring there relationship with
IDL.
Give me a stable, moderate cost environment and I'll be happy (well, content).
What we need is a Turbo Pascal sort of thing (low cost, does the job) to shake
some sense into RSI. Volume, volume, volume!
And don't get me started on dongles...
John Krist
------------------------------------------------------------ --------------
The opinions expressed above are my own and not necessarily those of my
employer (STScI) or its employer (AURA) or its employer (NASA) or its
employer (the Government) or its employer (me). Oops, I guess they are!
|
|
|
|
| Re: GUI Builder limited to Windows platform (LONG) [message #13488 is a reply to message #13329] |
Mon, 16 November 1998 00:00 |
pit
Messages: 92
Registered: January 1996
|
Member |
|
|
In article <72evlf$116$1@nnrp1.dejanews.com>,
mirko_vukovic@notes.mrc.sony.com writes:
> Puzzled, frustrated, I inquired. And the answer was simple enough, as
> if they were quoting Mr. Gates from his Economist article (a couple of
> months back). Any PC running WIntel will run their codes. But each
> Unix vendor's Unix is different, and requires additional programming
> effort.
> Thus in some sense, the Unix machine vendors have placed themselves in
> that unenviable position, with all of their best versions of Unix.
> Thus, for commercial written software, WIntel may be the platform of
> choice, and the workstations are left to internally written software.
And, shortly speaking, it is almost complete nonsense. Almost any
decent Unix program compiles on many flavours of Unix. There is more
than enough Software to run on, e.g. Dec, HP, AIX, SGI and Intel-Unix,
where you don't get a Win-version.
In general, if you have one unix-version, porting it to another flavour
of unix is very little effort. If I remember correctly, a programmer of
Informix wrote about the port to Linux: "Basically, it was just a
'make'.."
Peter
--
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~
Peter "Pit" Suetterlin http://www.uni-sw.gwdg.de/~pit
Universitaets-Sternwarte Goettingen
Tel.: +49 551 39-5048 pit@uni-sw.gwdg.de
-- * -- * ...-- * -- * ...-- * -- * ...-- * -- * ...-- * -- * ...-- * --
Come and see the stars! http://www.kis.uni-freiburg.de/~ps/SFB
Sternfreunde Breisgau e.V. Tel.: +49 7641 3492
____________________________________________________________ ______________
|
|
|
|
|
|
|
|
| Re: GUI Builder limited to Windows platform (LONG) [message #13492 is a reply to message #13329] |
Mon, 16 November 1998 00:00 |
Struan Gray
Messages: 178
Registered: December 1995
|
Senior Member |
|
|
Mirko , mirko_vukovic@notes.mrc.sony.com writes:
> On a related thread, while the programming features of IDL
> are growing, I would hope that the interactive features
> would do the same. Insight may be one attempt, but still
> very limited.
For me, the real benefit of the object orientation is actually
during interactive use. The fact that objects are globally accessible
and encapsulated makes it much easier to mix fully developed,
'production' code and new routines which are still at the
brainstorming stage.
I still do a lot of ad-hoc programming for one-off cases, but
anything that does get used more than a few times quickly migrates
from being a 'temp.pro' file to a full-blown widget, especially if
there are parameters I want to play with interactively (and these can
be algorithmic parameters as well as functional ones). Not
surprisingly, other people are much more willing to use my routines if
I provide them in a packaged form with widget front ends, and now I
can use those widgets myself while mucking about at the command line
with the same data.
I like IDL and try to evangelise it in my lab. RSI deserve credit
for staying multi-platform in a non-trivial way and they do seem to
try and address users' concerns. Prices are always emotive, and
personally I would love to see a cheaper way of distributing completed
applications, but on the other hand I am aware of the dangers of
insufficient cashflow to small businesses who are trying to innovate
and I also can't help wondering how much of the price of IDL is made
up of fixed licensing fees for things like OpenGL that can't be
reduced.
My biggest worry is that there is a tendency for known bugs to
persist forever. My favourite whinge is the lack of tools to monitor
memory usage, but the immortal contour plot problems and the
non-drawing of large offshore islands in direct graphics are other
examples of hugely annoying issues that seem to have been simply
ignored on the way to implementing sexier new features. There are
similar unncessary limitations built into the current object graphics
routines (the IDLgrTrackball object is the least object-like object
I've ever seen and a total pain in the arse to use), and I fear that
they too will be ignored or left to web-archives of known workarounds
to sort out rather than being fixed properly.
Such casual disregard for 'minor bugs' is less of a problem in a
community of IDL gurus who are essentially programming for themselves,
but tilts the learning curve to unnecessarily steep angles for novice
and casual users, which makes life harder for both myself and the RSI
sales force.
Struan
|
|
|
|
| Re: RSI's Priorities (was: GUI Builder...) [message #13498 is a reply to message #13448] |
Sun, 15 November 1998 00:00 |
mirko_vukovic
Messages: 50
Registered: January 1998
|
Member |
|
|
In article <gurman-1411982301570001@goodgulf.nascom.nasa.gov>,
gurman@gsfc.nasa.gov (Joseph B. Gurman) wrote:
> In article <364DDB7F.C8531805@io.harvard.edu>, Martin Schultz
> <mgs@io.harvard.edu> wrote:
>
>> David Fanning wrote:
>>
>>> Brian Jackel (jackel@danlon.physics.uwo.ca) writes:
>>>
>>> [...]
>>>>
>>>> New bells and whistles at $1500 a pop or
>>>> Freeze development and deliver an ``IDL Classic'' at
>>>> $500 per head.
>>>
> [Long comments omitted so my new client will let me post!]
>
> Frankly, I just pulled down the objects manual and the 5.1 changes
> manual the other day for the first time, to try to see if it was a better
> way to write MPEG movies than I was using (should I cross post to that
> thread?). Applying Occam's razor (the one that slits your throat if you
> try doing things elegantly when you already have three job titles and a
> family to boot), I will probably decide to make do with what I know how to
> do instead.... but I'm intrigued. Maybe next time.
>
Unfortunately, it takes much more than a weekend to appreciate the usefullness
of OOP. I was fortunate enough to have enough liberty at my previous and
current place of work, to have had the time to learn the rudiments of what
(I think) is the proper way of OOPing.
On a related thread, while the programming features of IDL are growing,
I would hope that the interactive features would do the same. Insight may
be one attempt, but still very limited.
The ideal, as I see it, would be a GUI type tool, where you would click and
drag data to appropriate functions (or use some of them to create the data).
If input were necessary, a dialog window would pop up. The interface would
be able to compile user functions and procedures, and create rudimentary,
non-intelligent input/output dialogs for them as well.
Since data analysis is typically a sequence of actions,
each action may be represented by connecting lines between data objects (I am
talking here arrays and vectors as objects, nothing fancy), and the functions
that operate
on them. A time sequence of these connections would then be transformed
into an IDL program for the user to tweak.
It seems to me that all of this should be implemented in IDL or Java. (I have
never programmed in Java). That should make it platform independent.
Now how 'bout that?! Off I go to implement this!!!
Yeah right :-)
cheers,
Mirko
-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
|
|
|
|
| Re: RSI's Priorities (was: GUI Builder...) [message #13500 is a reply to message #13448] |
Sat, 14 November 1998 00:00 |
gurman
Messages: 82
Registered: August 1992
|
Member |
|
|
In article <364DDB7F.C8531805@io.harvard.edu>, Martin Schultz
<mgs@io.harvard.edu> wrote:
> David Fanning wrote:
>
>> Brian Jackel (jackel@danlon.physics.uwo.ca) writes:
>>
>> [...]
>>>
>>> New bells and whistles at $1500 a pop or
>>> Freeze development and deliver an ``IDL Classic'' at
>>> $500 per head.
>>
[Long comments omitted so my new client will let me post!]
Frankly, I just pulled down the objects manual and the 5.1 changes
manual the other day for the first time, to try to see if it was a better
way to write MPEG movies than I was using (should I cross post to that
thread?). Applying Occam's razor (the one that slits your throat if you
try doing things elegantly when you already have three job titles and a
family to boot), I will probably decide to make do with what I know how to
do instead.... but I'm intrigued. Maybe next time.
More to the point, the folks here who _do_ have time all prefer
objects, and have written some very powerful analysis tools built around
them. (See, for example,
http://orpheus.nascom.nasa.gov/~zarro/idl/maps.html .)
And they're even scientists.... and even got their degrees more than 10
years ago. Guess I'll go with the opinions of those who write the most and
best code here.
Joe Gurman
P.S. And on the subject of MPEG movies and 3 x m x n arrays (why did RSI
have to make that the default for MPEG_PUT if you want color?), I'm
spoiled by a machine that's fast enough that I don't really have to worry
about the speed difference between intelligent and silly index ordering.
Now that's spoiled.
P.P.S. And yes, sigh, scientists do work weekends. I understand Ken Burns
used the "The intellect of man is forced to chose / Perfection of the life
or of the work" line of Yeats in his show on Frank Lloyd Wright. I wonder
if he used the next line: "and if it take the second must refuse a
heavenly mansion, raging in the dark."
--
Joseph B. Gurman / NASA Goddard Space Flight Center / Solar Physics Branch /
Greenbelt MD 20771 / work: gurman@gsfc.nasa.gov /other: gurman@ari.net
Government employees are still not allowed to hold opinions while at work,
so any opinions expressed herein must be someone else's.
|
|
|
|
| Re: RSI's Priorities (was: GUI Builder...) [message #13502 is a reply to message #13448] |
Sat, 14 November 1998 00:00 |
davidf
Messages: 2866
Registered: September 1996
|
Senior Member |
|
|
Brian Jackel (jackel@danlon.physics.uwo.ca) writes:
> Another important issue is that most of us researchers
> don't have much experience with object oriented programming.
> If I give some IDL code to someone who knows Fortran/Pascal/C
> they can probably see what's going on and make whatever
> changes they need. But objects? Forget it. Again, life
> is short and procedural programming works well enough for
> most of us.
>
> Hmmm, I'm sounding like an old codger now...
Brian, I've been known to feel a little old from time to
time too, but I can assure you that the people who want
you to think object oriented programming is hard are all
after your job and are at least 10 years younger than you. :-)
It is NOT hard. In fact, the wonderful thing about it
is just how EASY it is. Sigh... I've really got to
get this next book finished. The only thing that prevents
me is knowing that anyone who spends a hour reading the
book and learning about objects will soon be writing
programs that are about 10 times better than mine. :-(
You're right, Brian. Object programming *is* too hard. I
wouldn't even look into it. :-)
Cheers,
David
----------------------------------------------------------
David Fanning, Ph.D.
Fanning Software Consulting
E-Mail: davidf@dfanning.com
Phone: 970-221-0438, Toll-Free Book Orders: 1-888-461-0155
Coyote's Guide to IDL Programming: http://www.dfanning.com/
|
|
|
|
|
|
| Re: RSI's Priorities (was: GUI Builder...) [message #13504 is a reply to message #13448] |
Sat, 14 November 1998 00:00 |
Martin Schultz
Messages: 515
Registered: August 1997
|
Senior Member |
|
|
David Fanning wrote:
> Brian Jackel (jackel@danlon.physics.uwo.ca) writes:
>
> [...]
>>
>> New bells and whistles at $1500 a pop or
>> Freeze development and deliver an ``IDL Classic'' at
>> $500 per head.
>
> Objects have absolutely revolutionized the way I write
> programs. I think they will have as big an effect on me
> as widgets did when they were first introduced. I don't
> think I've written a program for pay in the last six
> months that didn't have at least one object in it. And
> I especially like them with direct graphics. They are
> fast, they print nicely, and they are unbelievably
> powerful.
>
> Could I do the same thing without objects? Probably, if
> I were clever enough. But objects just make it so darn
> easy. :-)
>
> I cast my vote more more bells and whistles like this
> one!
>
> Cheers,
>
> David
>
> ----------------------------------------------------------
> David Fanning, Ph.D.
> Fanning Software Consulting
> E-Mail: davidf@dfanning.com
> Phone: 970-221-0438, Toll-Free Book Orders: 1-888-461-0155
> Coyote's Guide to IDL Programming: http://www.dfanning.com/
I claim that this newsgroup is biased !! Sure enough, personally
I rather take it as my own fault that I have *not* been using objects in
IDL yet -- there are just always so many things that have to be finished
on time that I never get around learning about them (in fact, I haven't
even fully digested all the information that David gives on widget
programming in his book yet ... sigh ... if I only had time to improve
my EXPLORE program ... end sigh). But: e.g. in our research group there
about 10 people using IDL, and 1.5 of them participate in or even read
this newsgroup. Most others (and I would say that's probably typical for
scientists) have some basic knowledge in programming (sure, FORTRAN
(=f77) mostly), but not much beyond that. They are quite happy if they
can use e.g. a readdata.pro to retrieve some ASCII data from a file, and
they rejoice when they learn about the where function. Most programs
written in our groups start out as 5-20 lines of code, hand-tailored for
one specific plot. And most programs stay that way, too. I am already
happy if my colleagues make use of some of the library routines that
I have written or gathered from various sources, but if I were ever to
tell them: "Here's this perfectly handy and powerful routine, and it's
even object-oriented!" they would just thank me very politely but never
speak to me again ;-) There once was a german chancellor who said
"Entscheidend ist, was hinten rauskommt" (free translation: "it's the
plot, stupid! [not the program]"), and he sure was right on this one.
But if David ever writes another book on object oriented IDL programming
techniques, I may well become one of his customers!
Best regards,
Martin
--
------------------------------------------------------------ -------
Dr. Martin Schultz
Department for Engineering&Applied 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/
------------------------------------------------------------ -------
|
|
|
|
| Re: RSI's Priorities (was: GUI Builder...) [message #13507 is a reply to message #13448] |
Sat, 14 November 1998 00:00 |
Brian Jackel
Messages: 34
Registered: January 1998
|
Member |
|
|
Robert S. Mallozzi wrote:
> I agree - objects and pointers greatly increased IDL's
> functionality and ease of use. They, as widgets did, take
> IDL to a new level. I often find myself thinking,
> what's the best way to handle this data...use a pointer, of
> course. In fact, it's interesting to think of how I ever
> got along without them :-)
>
> Also, I feel there is no doubt about it - object programming
> is, in most cases, far superior to the "old" procedural method.
> Code reuse is one of the great advantages - write the object once,
> then *forget* about the internals, and just use it. That said,
> of course there is still the need for fast development and
> analysis, and objects might not be useful for this type of
> "quick and dirty" work.
Pointers are definitely handy, especially for getting large
amounts of data between widgets. As for objects, well...
Perhaps the problem is that most of my work is "quick and dirty"!
For example, I've just finished deblurring some satellite
images. This required
1) Figuring out how to read in a peculiar data format
2) Cleaning the data and calculating statistics
3) Trying several different deblurring algorithms
4) Creating a bunch of figures for a report
The whole process was very interactive and iterative. Now
I'm (sort of) done and have a set of working routines and
some example scripts. It might be nice to go back and clean
everything up, but time is short. Sure it could be nicely
packaged with objects, but I can't see how objects would
have helped me during the development phase.*
Another important issue is that most of us researchers
don't have much experience with object oriented programming.
If I give some IDL code to someone who knows Fortran/Pascal/C
they can probably see what's going on and make whatever
changes they need. But objects? Forget it. Again, life
is short and procedural programming works well enough for
most of us.
Hmmm, I'm sounding like an old codger now ("In my day,
all we had were zeros and ones, and sometimes we didn't
even have ones!"). Ten years ago as an undergrad I'd
learn new programming and scripting languages just because
they were cool. These days I'm more interested in physics,
and just want tools that work.
There does seem to be a clear distinction between people
involved in
1) Research/analysis
2) Development of ``production'' level code
For what it's worth, I'd say every one of the half-dozen
IDL users in our group fall into category 1.
Brian Jackel
*A general ``image'' object might have been useful, but I'm
not sure how a sufficiently generic thing could be created.
A satellite image is a completely different beast from the
image produced by an all-sky camera. I'll just stick with a
2D array, and do whatever pruning is required on a case-by
case basis.
|
|
|
|
| Re: RSI's Priorities (was: GUI Builder...) [message #13509 is a reply to message #13448] |
Sat, 14 November 1998 00:00 |
mallors
Messages: 76
Registered: November 1997
|
Member |
|
|
In article <MPG.10b6b35a5dca0660989722@news.frii.com>,
davidf@dfanning.com (David Fanning) writes:
> Brian Jackel (jackel@danlon.physics.uwo.ca) writes:
>
>> However, I'm watching recent developments with objects and
>> object graphics with some dismay. All of the data analysis
>> and display that I've needed to do has been easy to accomplish
>> without any of these newfangled object thingies. Seriously,
>> I've managed to maintain some moderate sized programs without
>> using objects. Given the existence of objects it is quite
>> likely that I'll get around to using them eventually, but I
>> can't help but see this as a choice between
>>
>> New bells and whistles at $1500 a pop or
>> Freeze development and deliver an ``IDL Classic'' at
>> $500 per head.
>
> Objects have absolutely revolutionized the way I write
> programs. I think they will have as big an effect on me
> as widgets did when they were first introduced. I don't
> think I've written a program for pay in the last six
> months that didn't have at least one object in it. And
> I especially like them with direct graphics. They are
> fast, they print nicely, and they are unbelievably
> powerful.
>
> Could I do the same thing without objects? Probably, if
> I were clever enough. But objects just make it so darn
> easy. :-)
>
> I cast my vote more more bells and whistles like this
> one!
I agree - objects and pointers greatly increased IDL's
functionality and ease of use. They, as widgets did, take
IDL to a new level. I often find myself thinking,
what's the best way to handle this data...use a pointer, of
course. In fact, it's interesting to think of how I ever
got along without them :-)
Also, I feel there is no doubt about it - object programming
is, in most cases, far superior to the "old" procedural method.
Code reuse is one of the great advantages - write the object once,
then *forget* about the internals, and just use it. That said,
of course there is still the need for fast development and
analysis, and objects might not be useful for this type of
"quick and dirty" work.
Regarding the recent discussions of the cost of IDL,
it is a bit disconcerting that the price of the software
really prohibits me from getting it for home use, and I believe
is restricting the possible widespread use that IDL might
enjoy if it was lower cost.
I think stand-alone IDL executables would certainly be a step
in the right direction for increasing the user base and thus
(hopefully) reducing the cost of the software. If people
could see the power of IDL without having to buy the
interpreter, it seems to me that there would be some
developers that might say "Gee, now that I see how powerful
the language is, maybe I should think about buying it to do
some development...". So, how about it, RSI - give us
the capability to distribute IDL executables for the
world to see!
--
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~
Robert S. Mallozzi 256-544-0887
Mail Code ES 84
Work: http://www.batse.msfc.nasa.gov/ Marshall Space Flight Center
Play: http://cspar.uah.edu/~mallozzir/ Huntsville, AL 35812
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~
|
|
|
|
| Re: RSI's Priorities (was: GUI Builder...) [message #13514 is a reply to message #13448] |
Fri, 13 November 1998 00:00 |
davidf
Messages: 2866
Registered: September 1996
|
Senior Member |
|
|
Brian Jackel (jackel@danlon.physics.uwo.ca) writes:
> However, I'm watching recent developments with objects and
> object graphics with some dismay. All of the data analysis
> and display that I've needed to do has been easy to accomplish
> without any of these newfangled object thingies. Seriously,
> I've managed to maintain some moderate sized programs without
> using objects. Given the existence of objects it is quite
> likely that I'll get around to using them eventually, but I
> can't help but see this as a choice between
>
> New bells and whistles at $1500 a pop or
> Freeze development and deliver an ``IDL Classic'' at
> $500 per head.
Objects have absolutely revolutionized the way I write
programs. I think they will have as big an effect on me
as widgets did when they were first introduced. I don't
think I've written a program for pay in the last six
months that didn't have at least one object in it. And
I especially like them with direct graphics. They are
fast, they print nicely, and they are unbelievably
powerful.
Could I do the same thing without objects? Probably, if
I were clever enough. But objects just make it so darn
easy. :-)
I cast my vote more more bells and whistles like this
one!
Cheers,
David
----------------------------------------------------------
David Fanning, Ph.D.
Fanning Software Consulting
E-Mail: davidf@dfanning.com
Phone: 970-221-0438, Toll-Free Book Orders: 1-888-461-0155
Coyote's Guide to IDL Programming: http://www.dfanning.com/
|
|
|
|
|
|
|
|
| Re: RSI's Priorities (was: GUI Builder...) [message #13546 is a reply to message #13448] |
Thu, 19 November 1998 00:00 |
LC's No-Spam Newsread
Messages: 18
Registered: September 1997
|
Junior Member |
|
|
On 16 Nov 1998, Peter Suetterlin wrote:
> Brian Jackel <jackel@danlon.physics.uwo.ca> writes:
>> There does seem to be a clear distinction between people involved in
>> 1) Research/analysis
>> 2) Development of ``production'' level code
> Brian, I do wholehartedly agree. I'm also from group 1, and really
> don't have a need for all those fancy things. Even widgets are a very
> rare thing for me....
I fully agree. Me and all my colleagues are all "group 1". Although it
does not mean that we may need well written and documented code for
some "repetitive" analysis.
I've been using IDL as a "scratchpad" programming environment for quick
data manipulation and visualization, and also for fast development of
concurrent variants of an algorithm (the final version to be coded in
Fortran). Advantages for this wrt other s/w environments should be
obvious.
I've *NEVER* used widgets myself (in IDL, all my GUI programming was
HTML forms ... for the rest I'm quite keen on command line mode !) except
in one instance recently, when I had to *demonstrate* a given look-and-feel
to a colleague ... which has to do a thing in Visual Basic (external
constraint) and is an absolute beginner with GUIs.
> Same for IDL, it looks like they want to make some sort of visual basic
> out of it. Together with the prices, this makes me heavily think about
> switching to ANA or Midas.
By "Midas" do you mean ESO/MIDAS ?
Will that be still kept alive for long ?
I agree that RSI policy is scaring many people away. We ourself survive with a
few oldish (4.0) licenses and are not interested in spending lot of money for
a so called "maintenance contract" (at the moment we are keeping alive an
old workstation which was our first license manager, because it looks a waste
to spend money for a maintenance contract just to be allowed to move the
license file - for a perfectly legitimate license set - on a new machine).
We would definitely prefer to buy several low cost "basic IDL" licenses (and
do not have license contention "ehi you down the corridor, have you finished
using IDL ?") for a frozen version.
I've seen several international astronomical projects (space and ground based)
rejecting to consider IDL as their development environment ("oh yes it's
quite good, but it's commercial, it costs money, and where will it go")
--
------------------------------------------------------------ ----------
nospam@ifctr.mi.cnr.it is a newsreading account used by more persons to
avoid unwanted spam. Any mail returning to this address will be rejected.
Users can disclose their e-mail address in the article if they wish so.
|
|
|
|
| Re: RSI's priorities [message #13551 is a reply to message #13448] |
Wed, 18 November 1998 00:00 |
David Ritscher
Messages: 30
Registered: August 1995
|
Member |
|
|
It seems like a simple way to solve the problem of the breaking of
old routines would be to add an IDL and PV-Wave functionality of
internal version control. There would be a variable called either
from MAIN or from within a routine, similar to the on_error command,
of a form:
version_functionality, 'IDL 5.1' or
version_functionality, 'Wave 6.0'
This would cause the interpreter to execute all commands according to
the functionality of that release.
This would mean a bit more internal overhead within IDL and PV-Wave,
but would mean something I wrote would always continue to work, and it
would also mean these firms could get much more adventurous with future
developments without worrying about losing their old customers.
They could even make radical paradigm shifts, for example, make
for i=0, 40000 do print,i
print something on the screen :-)
Even though changing this old faithful friend would break many of our
old programs.
David Ritscher
--
Cardiac Rhythm Management Laboratory
Department of Medicine
University of Alabama at Birmingham
B168 Volker Hall - 1670 University Boulevard
Birmingham AL 35294-0019
Tel: (205) 975-2122 Fax: (205) 975-4720
Email: davidNO.ritscherSPAM@bigfoot.com
|
|
|
|
|
|
|
|
| Re: RSI's priorities [message #13562 is a reply to message #13448] |
Wed, 18 November 1998 00:00 |
David Kastrup
Messages: 33
Registered: February 1998
|
Member |
|
|
davidf@dfanning.com (David Fanning) writes:
> John Krist (krist@nemesis.stsci.edu) writes:
>
>> [Long whinge snipped.]
>
>> It seems to me the same has been happening with objects.
>> They were implemented before they were really well thought out.
>
> Like most of *my* programs, as it turns out. :-(
>
> Ever try writing a book, building a desk, writing a long love
> letter. If things weren't implemented before they were really well
> thought out nothing much would be happening around here. :-)
That's the reason why one does well in complicated products (books,
programs, particularly user interface stuff) not properly fathomed in
advance to first implement them "organically" until they work, then
scrap the entire attempt and write the stuff from a clean slate again,
after one has a good overview of the matter. It almost always leads
to lots of woe if you huddle out the first ad-hoc design and bereave
yourself of that possibility.
--
David Kastrup Phone: +49-234-700-5570
Email: dak@neuroinformatik.ruhr-uni-bochum.de Fax: +49-234-709-4209
Institut f�r Neuroinformatik, Universit�tsstr. 150, 44780 Bochum, Germany
|
|
|
|
|
|
|
|
| Re: GUI Builder limited to Windows platform (LONG) [message #13606 is a reply to message #13329] |
Tue, 24 November 1998 00:00 |
R.Bauer
Messages: 1424
Registered: November 1998
|
Senior Member |
|
|
David Fanning wrote:
> David Foster (foster@bial1.ucsd.edu) writes:
>
>> I probably wouldn't even use the Builder myself, but I know a lot
>> of groups that are just beginning to develop widget programs, and
>> this tool would be invaluable to them.
>
> Possibly. But I'm still partial to a good book on the
> subject. :-)
>
> As it happens, I tried to use the Builder this afternoon.
> I was having problems coming up with some good ideas for
> an interface, so I thought I would sketch a few out with
> this tool.
>
> Granted, I didn't read anything about it. I just figured
> it would be sort of easy to use. But I found it confusing,
> and I found that I was futzing around with it so much that
> I wasn't getting anything done. (This in itself is not
> that unusual, to be fair.)
>
> I'll probably give it another try later, but I'm
> pretty sure the code it generates will not help
> novices write better widget programs. Faster? Maybe,
> if they can learn to use it. Easier to maintain? Not
> likely. But, hey, the more people are confused,
> the better my accountant and I like it. :-)
>
> Cheers,
>
> David
>
Dear David, I aggree to you without this point.
I believe with the builder its easier to align widget by positioning in
one base. I will use it only for designing a widget the event handling
is too much clicking.
Reimar
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
