Blogging the IDL User's Group Meeting

Blogging the IDL User's Group Meeting

Post by David Fann » Fri, 26 Feb 2010 01:52:22


olks,

An IDL newsgroup first: Live Blogging! from the IDL User's Group
meeting, 24 Feb 2010, Boulder, CO.

Kirk Bennel, Chief Technology Officer, "the best release we
have had in the last 10 years, maybe even ever." I think he
may be right about this. It *is* exciting.

"Expect to see not just marketing people, but people who
build the product on the road, talking to customers."

"Focusing on Software for Science."

;*******************************************************

Bill Okubo, ID Product Manager.

It wasn't a dream, which is what I thought when I woke up
this morning. IDL 8 really does have new graphics programs
that will change everything about how graphics are created
and used in IDL. Bill says, "Power of object graphics, in
an easy to use syntax." YES!! Even scientists will be able
to use these routines. And better yet, half the Coyote
Library can be retired forever.

Language enhancements:

!NULL arrays, scalars, etc.
Object garbage collection (reference counted)
List and Hash variables
FOREACH operator
Negative array indices
ISA function
Operator overloading (Yikes!)

Graphics

Simple, direct graphics like syntax for producing object
graphics plots. YEAH!!!!
Not everything here in this realize, but lots of it is

Workbench

Simplified the interface. Have gone back to combined
command line/console. Simple, unclutter look, with
toolbar functionality (that can be shrunk or hidden).
Looks very nice.

Help System

Revamped again! But maybe this time they have it right,
or mostly right. More emphasis on examples (click and run).

Other

DIMENSION keyword for Mean and Moment
New plot symbols.
Integrated PDF output. (Not ready yet, but will be there
by release (May?)


Beau Legeer, Technical Manager for IDL/ENVI

Top Ten Features in IDL 8.0

1. Integrated command line/console. Content assist. Colors
in console output. Completion tab for finding history
commands.

2. Automatic garbage collection.
Pointers and objects that fall out of scope are
automatically cleanup up.

3. FOREACH control statement. Works with arrays, lists,
and hashes (hash still comming).

4. !NULL Null pointers, objects, variables, free memory.
Null arrays possible. VERY easy syntax: n = [] of n{}.

5. Dot for method calls.

obj -> DoIt() becomes obj.doit()

(Ed. Note:, this is said to be "bomb-proof", but I
haven't had a chance to test this yet.)

6. Dot notation for Get/Set Property methods.

obj -> GetPropery, NAME=name & Print, name

becomes:

Print, obj.name
obj.name = 'Coyote'

To use, objects must inherit IDL_Object.
(Ed. Note. GREAT!!)

7. !Color 147 color names! Forget about FSC_Color()!

8. Operator overloading. All IDL operators, including HELP and PRINT.

9. Lists and Hash tables

List - collection of different data types, access using array syntax
Hash - like a list, but with string "keys" to access. (Not in current
build, but will be in 8.0)

10. New Graphics. Resizeable windows, when change property automatic
updating, no worries about color decomposition, layout controls
(better !P.Multi). Buttons on window allow output to standard
formats, including PostScript, printer, JPEG, PNG, etc.
(Ed. Note. This is how this should be done!)





--
David Fanning, Ph.D.
Fanning Software Consulti
 
 
 

Blogging the IDL User's Group Meeting

Post by R.G. Stock » Fri, 26 Feb 2010 02:02:45


......


wow, sounds great. Thanks for posting this stuff.

 
 
 

Blogging the IDL User's Group Meeting

Post by chri » Fri, 26 Feb 2010 16:47:43

n 24 Feb., 17:52, David Fanning < XXXX@XXXXX.COM > wrote:
> List and Hash variable>
> FOREACH operat>r
> Negative array indi>es
> ISA func>ion
> Operator overloading (Yi>es>)
>
> Gra>hi>s
>
> imple, direct graphics like syntax for producing >bject
> graphics plots.>YEAH!!!!
> ot everything here in this realize, but lot> o> it is
>
>>Wo>kbench
>
> implified the interface. Have gone back >o combined
> ommand line/console. Simple, unclutte> look, with
> oolbar functionality (that can be shrun> or hidden).
> o>ks>very nice.
> >> >elp System
>
> evamped again! But maybe this time they>have it right,
> r mostly right. More emphasis on examples >cl>ck and r>n)>
>
> Other
>
> DIMENSION keyword fo> Mean and Moment
> gt;New plot symbols.
> Integrated PDF output. (Not ready yet> but will be there
> gt;>by release (May?)
>
> Beau Legeer, Technical>Ma>ager for IDL/ENVI
>
> Top Te> F>atures in IDL 8.0
>
> 1. Integrated command line/console. C>ntent assist. Colors
> n console output. Completion t>b for finding >is>ory
> ommands.
>
> 2. Automa>ic garbage collection.
> ointers and objects th>t fall out of scope are
> gt;ut>matically cleanup up.
>
> 3. FOREACH control statement. >orks with arrays, lists,
> nd ha>he> (hash still comming).
>
> 4. !NULL Null pointers, object>, variables, free memory.
> ull arrays possible. VERY >as> syntax: n = [] of n{}.
> >> >. Dot fo> method calls.
>
> bj -> D>It() ecomes obj.doit()
>
> Ed. Note:, this is >aid to be "bomb-proof", but I
> haven't>ha> a chance to test this yet.)
>
> 6. Dot notati>n >or Get/S>t Property methods.
>
> bj -> Ge>Pr>pery, NAME=na>e > Print, name
>
> gt;ecomes:
>
> rint> o>j.name
> bj.name = 'Coyote'
>
> o >se, objects must inheri> I>L_Object.
> Ed. Note. GREAT!!)
>
> 7. !Color 4> c>lor names! Forget about FSC_Color()!
>
> 8. Operator overloading. All >DL>operators, including HELP a>d >RINT.
>
> 9. Lists and Hash tables
>
> ist - collection of differ>nt data types, access using array syntax
> ash - like a list, but wit> string "keys" to access. (Not >n >urrent
> uild, but will be in 8.0)
>
> 10. New Graphics. Resizea>le windows, when change property automatic
> updating, no worri>s about color decomposition, layout controls
> (better !P.Mul>i). Buttons on window allow output to standard
> format>, including PostScript, printer, JPEG, PNG, etc. >> > (E>. Note. This is how thi> should be done!)
>
> --
> David >anning, Ph.D.
> Fanning Software Consulting, Inc.
> Coyote'> Guide to IDL Programming:http://www.dfanning.com/
> Sepore ma de ni thue. ("Perhaps thos speakest truth.")

Good news. I'm looking forward :)

Maybe Python has to wait....

Cheers

CR
 
 
 

Blogging the IDL User's Group Meeting

Post by ben.bighai » Fri, 26 Feb 2010 23:10:22


Hi David,

Thanks for sharing the exciting details. Sounds like you are having
fun and the the rest of us might have some soon, too!

Cheers,
Ben
 
 
 

Blogging the IDL User's Group Meeting

Post by Dave Pore » Fri, 26 Feb 2010 23:27:28

n Feb 24, 8:52m, David Fanning < XXXX@XXXXX.COM > wrote:
> List and Hash variable>
> FOREACH operat>r
> Negative array indi>es
> ISA func>ion
> Operator overloading (Yi>es>)
>
> Gra>hi>s
>
> imple, direct graphics like syntax for producing >bject
> graphics plots.>YEAH!!!!
> ot everything here in this realize, but lot> o> it is
>
>>Wo>kbench
>
> implified the interface. Have gone back >o combined
> ommand line/console. Simple, unclutte> look, with
> oolbar functionality (that can be shrun> or hidden).
> o>ks>very nice.
> >> >elp System
>
> evamped again! But maybe this time they>have it right,
> r mostly right. More emphasis on examples >cl>ck and r>n)>
>
> Other
>
> DIMENSION keyword fo> Mean and Moment
> gt;New plot symbols.
> Integrated PDF output. (Not ready yet> but will be there
> gt;>by release (May?)
>
> Beau Legeer, Technical>Ma>ager for IDL/ENVI
>
> Top Te> F>atures in IDL 8.0
>
> 1. Integrated command line/console. C>ntent assist. Colors
> n console output. Completion t>b for finding >is>ory
> ommands.
>
> 2. Automa>ic garbage collection.
> ointers and objects th>t fall out of scope are
> gt;ut>matically cleanup up.
>
> 3. FOREACH control statement. >orks with arrays, lists,
> nd ha>he> (hash still comming).
>
> 4. !NULL Null pointers, object>, variables, free memory.
> ull arrays possible. VERY >as> syntax: n = [] of n{}.
> >> >. Dot fo> method calls.
>
> bj -> D>It() ecomes obj.doit()
>
> Ed. Note:, this is >aid to be "bomb-proof", but I
> haven't>ha> a chance to test this yet.)
>
> 6. Dot notati>n >or Get/S>t Property methods.
>
> bj -> Ge>Pr>pery, NAME=na>e > Print, name
>
> gt;ecomes:
>
> rint> o>j.name
> bj.name = 'Coyote'
>
> o >se, objects must inheri> I>L_Object.
> Ed. Note. GREAT!!)
>
> 7. !Color 4> c>lor names! Forget about FSC_Color()!
>
> 8. Operator overloading. All >DL>operators, including HELP a>d >RINT.
>
> 9. Lists and Hash tables
>
> ist - collection of differ>nt data types, access using array syntax
> ash - like a list, but wit> string "keys" to access. (Not >n >urrent
> uild, but will be in 8.0)
>
> 10. New Graphics. Resizea>le windows, when change property automatic
> updating, no worri>s about color decomposition, layout controls
> (better !P.Mul>i). Buttons on window allow output to standard
> format>, including PostScript, printer, JPEG, PNG, etc. >> > (E>. Note. This is how thi> should be done!)
>
> --
> David >anning, Ph.D.
> Fanning Software Consulting, Inc.
> Coyote'> Guide to IDL Programming:http://www.dfanning.com/
> Sepore ma de ni thue. ("Perhaps thos speakest truth.")

Wow!! Sounds great. I can wait.
Cheers
Dave
 
 
 

Blogging the IDL User's Group Meeting

Post by David Fann » Fri, 26 Feb 2010 23:42:55

Ben Tupper writes:


For nearly 25 years, I have been trying to "fix" IDL,
either by writing programs that work the way I think
they should work, or by teaching people how to work
around IDL's shortcomings. I've grown old and cynical.

I really didn't expect to feel much different after
this week's User Group meeting. But I have to say that
what I have seen in the past two days is not only good
news for the casual IDL user, but is *great* news
for IDL developers.

I don't want to set expectations too high. This is
just the first release in a new direction, after all,
and I haven't really had a chance to play with the
code and understand how it all works together.
There are obvious pieces still missing. But this step
is just absolutely in the right direction. For me,
it provides tools that I can see myself using every
single day. That hasn't happened for a LONG time.

I said yesterday that one reason I haven't yet
written that 3rd Edition of my book is that it is
difficult to write while popping Prozac for
depression. If you don't see a future for the
software, why write about it? But this changes
everything. I can see myself writing with Happy
Thoughts once again. That, in itself, is
enormously comforting. :-)

Cheers,

David

P.S. The other thing cheering me up today is the
fact that I will no longer have to support
half the programs in the Coyote Library! Maybe
I can give Coyote the boot and hire that young
thing with the doe eyes to handle the lighter
load. :-)

--
David Fanning, Ph.D.
Fanning Software Consulting, Inc.
Coyote's Guide to IDL Programming: http://www.yqcomputer.com/
Sepore ma de ni thui. ("Perhaps thou speakest truth.")
 
 
 

Blogging the IDL User's Group Meeting

Post by kBob » Sat, 27 Feb 2010 00:20:41

n Feb 24, 9:52m, David Fanning < XXXX@XXXXX.COM > wrote:
> List and Hash variable>
> FOREACH operat>r
> Negative array indi>es
> ISA func>ion
> Operator overloading (Yi>es>)
>
> Gra>hi>s
>
> imple, direct graphics like syntax for producing >bject
> graphics plots.>YEAH!!!!
> ot everything here in this realize, but lot> o> it is
>
>>Wo>kbench
>
> implified the interface. Have gone back >o combined
> ommand line/console. Simple, unclutte> look, with
> oolbar functionality (that can be shrun> or hidden).
> o>ks>very nice.
> >> >elp System
>
> evamped again! But maybe this time they>have it right,
> r mostly right. More emphasis on examples >cl>ck and r>n)>
>
> Other
>
> DIMENSION keyword fo> Mean and Moment
> gt;New plot symbols.
> Integrated PDF output. (Not ready yet> but will be there
> gt;>by release (May?)
>
> Beau Legeer, Technical>Ma>ager for IDL/ENVI
>
> Top Te> F>atures in IDL 8.0
>
> 1. Integrated command line/console. C>ntent assist. Colors
> n console output. Completion t>b for finding >is>ory
> ommands.
>
> 2. Automa>ic garbage collection.
> ointers and objects th>t fall out of scope are
> gt;ut>matically cleanup up.
>
> 3. FOREACH control statement. >orks with arrays, lists,
> nd ha>he> (hash still comming).
>
> 4. !NULL Null pointers, object>, variables, free memory.
> ull arrays possible. VERY >as> syntax: n = [] of n{}.
> >> >. Dot fo> method calls.
>
> bj -> D>It() ecomes obj.doit()
>
> Ed. Note:, this is >aid to be "bomb-proof", but I
> haven't>ha> a chance to test this yet.)
>
> 6. Dot notati>n >or Get/S>t Property methods.
>
> bj -> Ge>Pr>pery, NAME=na>e > Print, name
>
> gt;ecomes:
>
> rint> o>j.name
> bj.name = 'Coyote'
>
> o >se, objects must inheri> I>L_Object.
> Ed. Note. GREAT!!)
>
> 7. !Color 4> c>lor names! Forget about FSC_Color()!
>
> 8. Operator overloading. All >DL>operators, including HELP a>d >RINT.
>
> 9. Lists and Hash tables
>
> ist - collection of differ>nt data types, access using array syntax
> ash - like a list, but wit> string "keys" to access. (Not >n >urrent
> uild, but will be in 8.0)
>
> 10. New Graphics. Resizea>le windows, when change property automatic
> updating, no worri>s about color decomposition, layout controls
> (better !P.Mul>i). Buttons on window allow output to standard
> format>, including PostScript, printer, JPEG, PNG, etc. >> > (E>. Note. This is how thi> should be done!)
>
> --
> David >anning, Ph.D.
> Fanning Software Consulting, Inc.
> Coyote'> Guide to IDL Programming:http://www.dfanning.com/
> Sepore ma de ni thue. ("Perhaps thos speakest truth.")

Lets us not forget all those great presentation on day two - all very
good and inspirational. I glade to be an IDL programmer. Especially,
David's talk about Map Projections. Maybe in their new Agile
programming environment, ITTVIS will give us new up to date (PROJ4)
mapping routines by IDL 8.1 release or as a patch.


Kelly Dean
Fort Collins, CO
 
 
 

Blogging the IDL User's Group Meeting

Post by savoi » Sat, 27 Feb 2010 00:45:53

kBob < XXXX@XXXXX.COM > writes:



That might be a reason to stick with IDL.

Matt

--
Matthew Savoie - Scientific Programmer
National Snow and Ice Data Center
(303) 735-0785 http://www.yqcomputer.com/
 
 
 

Blogging the IDL User's Group Meeting

Post by wlandsma » Sat, 27 Feb 2010 01:34:15


Thanks for the detailed and optimistic report. I know I shouldn't
be greedy but was there any news about planned future enhancements to
IDL widgets? Thanks, -Wayne
 
 
 

Blogging the IDL User's Group Meeting

Post by David Fann » Sat, 27 Feb 2010 01:46:12

David Fanning writes:


I've been informed by Coyote's attorney this morning that
I can take this idea and "put it where the sun don't shine".
Apparently Coyote has an iron-clad contract through at least
2025. Oh, well, we can't have everything, I guess. :-(

Cheers,

David



--
David Fanning, Ph.D.
Fanning Software Consulting, Inc.
Coyote's Guide to IDL Programming: http://www.yqcomputer.com/
Sepore ma de ni thue. ("Perhaps thos speakest truth.")
 
 
 

Blogging the IDL User's Group Meeting

Post by David Fann » Sat, 27 Feb 2010 01:49:05

Wayne Landsman writes:


No news, but repeated requests from the audience,
especially during Tuesday's open forum. I think
ITTVIS acknowledges the need, but--as always--
has a lot on their plate. In my view, they
can't be doing graphics this nice without seeing
clearly that this in how widgets should work, too.

Personally, I think they understand the need. When
they might get it done is anyone's guess. I didn't
hear anyone give anything away, even in private
conversations.

Cheers,

David



--
David Fanning, Ph.D.
Fanning Software Consulting, Inc.
Coyote's Guide to IDL Programming: http://www.yqcomputer.com/
Sepore ma de ni thue. ("Perhaps thos speakest truth.")
 
 
 

Blogging the IDL User's Group Meeting

Post by Spon » Sat, 27 Feb 2010 03:52:27


I say go for it!
Nothing wrong with doe eyes - if they're fresh!

Good news, thanks for keeping us posted!

Regards,
Chris
 
 
 

Blogging the IDL User's Group Meeting

Post by chri » Sat, 27 Feb 2010 16:26:10

I hope, that they will also give the opportunity to create real
executable binaries are intend to do this in future releases. Make_rt
is not a real option.

Cheers

CR
 
 
 

Blogging the IDL User's Group Meeting

Post by pp » Sat, 27 Feb 2010 19:25:03


I would find that very surprising. IDL programs run on virtual
machines, so it would be a major fundamental change to make it a
native compiler. It is the same in other VM languages, as far as I
know, such as Python and Java, for instance. Even though Java uses
strong static types (much easier to compile), and has the JIT compiler
(to make some parts into native binaries), no Java program runs
without the VM.

make_rt produces a more complete solution than can be done in Java,
since it includes the VM, so it can be used directly (though there are
the IDL VM limitations, which exist because IDL is not free). The way
to distribute ready programs in Java, jar files, is more similar to
making IDL save files. Since everybody has the Java VM, distributing
jar files is more practical than distributing IDL save files. But IDL
will never be as ubiquitous as Java, since it is neither free or a
general purpose language.
 
 
 

Blogging the IDL User's Group Meeting

Post by vasil » Sun, 07 Mar 2010 01:38:53

Any hope for multi-core support? Most desktops nowadays come with
multiple cores and for many tasks in data analysis simple task farming
would help enormously!