| Re: IDL 7.0 Preview Review [message #56825 is a reply to message #56744] |
Wed, 14 November 2007 09:55  |
Michael Galloy
Messages: 1114
Registered: April 2006
|
Senior Member |
|
|
I'd say the best part was meeting people from the newsgroup in person
at the reception. We need to get together more often.
See below for more comments...
On Nov 13, 10:47 pm, David Fanning <n...@dfanning.com> wrote:
> Anyway, the new IDL Workbench sorta looks like the old IDLDE, but
> not quite, and its WAY cooler than anything else anyone could
> think of, and it works the same on every platform God ever made,
> which is a HUGE advantage to the IDL developers, as they only
> have to maintain one code base now. (Presumably freeing up some
> time to work on something else, so we will get more for our
> maintenance dollars, and so forth.)
>
> One of the developers showed us a few of its tricks. (A still
> secret incantation will allow the user to hook up an RSS feed
> of this very newsgroup into the Workbench environment. A feature
> some people may not see as much of an advance, come to think of it.)
I was surprised to learn there was a developer's RSS feed from the ITT
VIS website that is displayed there now. Check out:
http://www.ittvis.com/feeds/devnet.xml
Now how about ones for new code contrib entries, news articles, tech
tips, etc.? Trying to figure out if there were any new posts in the
beta forum was a pain. How about an RSS feed?
> Anyway, LOTS of new features that certainly got the presenters
> pretty darned excited. Of course, the cynical programmer sitting
> beside me pointed out that he had been doing everything they were
> showing up there for *at least* the last five years with EMACS and
> IDLWAVE mode. So I guess you could think of the IDL Workbench
> as EMACS-lite, if you like. Anyway, it does some nifty things.
And, as an Emacs user, what program couldn't be described as "Emacs-
lite"?
> Of course, if you are too old to change (it is a possibility for
> some of the gray hairs sitting in the audience), those of you working
> on UNIX platforms can still have access to the normal IDL command line.
> If you are a Windows user, well, you are used to being told what to do
> anyway. And, I'm SURE you are going to like it!
I still don't know why there is no Windows command line. I expect it
is because they think there is no demand for it. Windows users should
ask for this if they are interested in it.
> Everyone was careful to point out that there was a bit of a learning
> curve with the new Workbench, but that in a couple of days you would
> be so overjoyed with the new features that learning would actually
> be fun again. Ronn Kling has written a little book, entitled
> "Navigating the IDL Workbench" that I suppose is yours for the asking.
> The second chapter is ominously entitled "Don't Panic!!!", but it is
> short enough to read while you eat a burrito, so that's encouraging.
As you indicate below, Emacs has a steep learning curve as well. I'm
fairly confident that the Eclipse learning curve is much easier on new
users while still providing approximately the same (or more)
functionality.
> After the IDL Workbench demo we got an update on where IDL is headed
> in the next 3-4 years. The target audience "doesn't want to read a
> manual", "wants something that looks like his Java or C++ environment",
> "wants to focus on the task at hand", and "wants to get away from
> programming". And--I was thinking--is dumb as a stone, but they didn't
> say that.
I was a bit worried about that part too. But since I type words on my
computer instead of just drag-and-dropping all day, maybe this will
insure my position in the ruling elite.
> Anyway, expect more features that "modernize how users interact
> with IDL". This would include more GUI-based tools, drag and drop
> functionality, and what is going to be called "Interactive Ad-Hoc
> Analysis", which is, well, different in some way than what we do
> now. "We want to change the way you interact with IDL!", they said.
> Hey, I'm all for that. The less thinking I have to do, the better,
> I say. They did offer some examples of how one might set graphics
> and other visualization properties from the IDL command line that
> looked to me a lot like what MatLab allows you to do now.
>
> We did get some exciting news. The IDL GUI-Builder is history!
> Having that god-awful code languish in infamy is a step forward,
> I think. :-)
I almost stood up and clapped.
> The bad news is, we are going to make another stab at a ToolBuilder
> sometime in the future. (There is hope we might get something useable
> on this, the fourth, or is it fifth?, iteration.) Anyway, Eclipse is
> set up for this kind of thing, so we are going to have another go.
> This one will be neat. No event handling, no state structure, no
> coding. Everything will just work! (I asked how data you might want
> to do something with would get passed around in this "program", but
> those details haven't been worked out yet. In conversations later,
> I learned that people were actually thinking about this and there
> was talk about having data "show up" in the "right scope" somehow.
> At least it gave me some ideas for how I could make data "show up
> in the right scope" even now!)
I suppose we'll just have to wait and see. The only "tool builder"
type application I've ever liked was the QT Designer and I still hand
code instead of using it. I'm probably not who they are building it
for, though. But at least have it build code that I can fix.
> After the Workshop presentation, we were treated to a demonstration
> of an IDL program written at LASP, using iTool functionality. Just
> about the entire thing had been coded up by an undergraduate programmer
> who "didn't know a darn thing about IDL when he started." The point of
> the presentation was, apparently, to demonstrate how easy it is to
> build tools using iTools functionality.
Or, once you spend some time figuring it out, how easy it is to build
tools using iTools functionality.
> Of course, they couldn't quite work out how to build the iTool they
> wanted to build (and who here would cast the first stone?), so they
> opted for putting all their functionality in a single pane of an
> existing iTool. Still, it worked pretty good. Then we learned that
> "We use common blocks to pass data. It's probably not ideal, but it was
> one of the ways we could get around some iTool limitations." Whatever.
> I doubt a professional programmer could have done much better,
> and my hat is off to the kid.
Yes, it looked pretty useful.
> Next up, Ronn Kling showed us some of the neat things he has
> coded up for his web page and in his annual Christmas cards.
> He is always doing *something* interesting. We didn't get to
> look at code, of course, and the "how?" was pretty sketchy,
> but sometimes just knowing something is possible is enough
> to give you encouragement.
Ronn definitely showed some cool demos. I've used his trick of
creating another view to show the magnifying glass view of an an image
over the top of the full size view of the image, but I was surprised
to hear him say they did it because "device, /copy was unacceptably
slow" for their application.
> Finally, an ITTVIS technical support engineer gave us a
> run-down on what's happening on the ITTVIS web page. We
> learned about some neat new things on the IDL Code
> Contrib page, including (and I have not downloaded this
> yet myself) an IDL program for viewing and reading GRIB
> files. (My cynical programming friend and I both raised
> our eyebrows, pretty much simultaneously, so that is a
> VERY good sign.)
>
> I detected a new and refreshing openness on the part of
> IDL developers at this meeting to interact and engage
> customers in what they really wanted. There is an openness
> to working with 3rd party developers and consultants
> that I have never seen before. I certainly came away
> feeling that the ITTVIS people are smart, capable,
> and ready to listen to user feedback.
I did too and it was encouraging.
> Oddly, I don't seem to run into the customer they are
> apparently writing code for in my day to day IDL dealings,
> but he or she must be out there. They are sure spending a lot
> of time and money going in this direction.
These might not be *current* customers.
> I'm going to have a look at the new IDL Workshop. God knows
> I've tried to learn EMACS for the last 4-5 months and I *still*
> have to look up how to do a Search-and-Replace every time I
> want to do it. My fingers just don't want to learn that dance.
> The IDL Workshop may be just what the doctor ordered. At least
> I hope so.
Like I said before, I think the learning curve on the Workbench will
be easier on new users. I think debugging is more intuitive and
powerful on the Workbench as well. IDL-Wave probably has more features
as well as leveraging all the native Emacs features.
> The IDL 7.0 CD was signed off on just today. It will be available
> from the ITTVIS web page sometime near the end of November, and
> will be shipped to users with valid maintenance contracts about
> two weeks after it is released on the web page.
I can't wait. I really like the new features and use the Workbench
regularly for real work, but I need some bugs fixed from the beta...
Mike
--
www.michaelgalloy.com
Tech-X Corporation
Software Developer II
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
