| Introducing FL [message #48144] |
Sat, 01 April 2006 06:44  |
FL
Messages: 17
Registered: April 2006
|
Junior Member |
|
|
Hi guys!
The first public snapshot of Fawlty Language, an IDL compatible compiler
is available. (IDL is a registered trademark of Research Systems Inc.)
FL binaries can be freely used and downloaded from the homepage
http://web.interware.hu/fl
Available packages:
fl_0.6-i686-linux.tar.gz (Linux, Pentium II.
kernel 2.4/2.6, glibc >= 2.2.5)
fl_0.6-amd64-linux.tar.gz (Linux, AMD Athlon64,
kernel 2.6, glibc >= 2.3.3)
and for the bravehearted:
fl_0.6-i686-mingw.zip (Windows 2000/XP/?, Pentium II)
Features:
- all language elements are supported
- multithreaded operators
- array operations use MMX/SSE/SSE2, if available
- module profiling
- line profiling
- about 280 library functions (more or less usable)
- true-color (24 bit) direct graphics devices: X, WIN, PS, PDF, Z
- run-time performance: for many programs, FL is faster than IDL
(eg. the empty loop is three times faster in FL :-)
Enjoy!
fl
(Földy Lajos)
ps: Hungarian names are in reverse order. Wait, no! English names are
in reverse order :-). So my name is Lajos Foldy in English.
|
|
|
|
| Re: Introducing FL [message #48214 is a reply to message #48144] |
Thu, 06 April 2006 09:50  |
FL
Messages: 17
Registered: April 2006
|
Junior Member |
|
|
yes, sure. After stabilizing the internals I will add linkimage or DLM
support.
regards,
lajos
On Thu, 6 Apr 2006, Maurizio Tomasi wrote:
> FL wrote:
>> In the next few months I will clean up code, fixing bugs and extending
>> usability. If you have any ideas about features that the language should
>> have, but never dared to ask for, now it's time to speak. Speak now or
>> forever hold your peace :-)
>
> Is FL going to implement LINKIMAGE and dynamic linked modules? It would
> be a good way for other users to contribute even without having the
> source code for FL!
>
> Maurizio.
>
|
|
|
|
| Re: Introducing FL [message #48223 is a reply to message #48144] |
Thu, 06 April 2006 00:14 |
Maurizio Tomasi
Messages: 7
Registered: December 2005
|
Junior Member |
|
|
FL wrote:
> In the next few months I will clean up code, fixing bugs and extending
> usability. If you have any ideas about features that the language should
> have, but never dared to ask for, now it's time to speak. Speak now or
> forever hold your peace :-)
Is FL going to implement LINKIMAGE and dynamic linked modules? It would
be a good way for other users to contribute even without having the
source code for FL!
Maurizio.
|
|
|
|
| Re: Introducing FL [message #48242 is a reply to message #48144] |
Tue, 04 April 2006 16:28 |
George N. White III
Messages: 56
Registered: September 2000
|
Member |
|
|
On Tue, 4 Apr 2006, Paul Van Delst wrote:
> Y.T. wrote:
> [...]
>> At my place of employ I can get Matlab for free (i.e. site-license) but
>> I'd be horribly worthless with Matlab because I'd be spending my time
>> re-re-relearning when and where to place a comma or a semicolon and
>> what was the syntax for a linear fit again? So I use GDL, because that
>> allows me to get something *done*. As in *now*.
>
> That, of course, is very important but....
>>
>> (I work for a federally funded research facility and you, the
>> tax-payer, are expending my salary. Do you really think you're getting
>> your money's worth out of the deal if I spend my time learning this
>> years fad-language?)
>
> Yes, actually (being a taxpayer who also happens to work at a federally
> funded research facility). Learning new programming languages is not unlike
> learning new spoken languages (except easier IMO). It broadens one's horizon
> to make different solution methodologies available. In some cases they may
> help your work, others not. I would be a foolish taxpayer if I expected you
> to not expend time learning new stuff that may make your job easier and more
> efficient. And, of course, there's the "personal improvement" aspect -- ya
> gotta be happy at what you do. :o)
>
> paulv
>
> p.s. I would love to get a matlab site license - even though I barely know
> enough matlab for the "hello world" chestnut. Until then, learning Ruby and
> Python while I wait for Fortran2003 compilers will have to do. :o)
I also work at a government funded research facility where we have had
Matlab longer than we have had IDL (since the days when Cleve Moler worked
for Ardent). I have coded a number of core algorithms we use in both
Matlab and IDL as a sort of crib sheet for post-docs (many have worked
with Matlab, few with IDL). For my work, Matlab suffers in comparison to
IDL because (like S+) it encourages you to use doubles, but (unlike S+) it
lacks missing value support which is important for the things I do where
doubles are appropriate. If I needed sparse arrays I wouldn't mind the
doubles so much, but these days I'm more interested in finding tools that
support the OpenEXR 16-bit floating pt. format.
All these languages handle 75-80% of the problems I have without
straining. Where they differ is in the interests of the user community
and in the ability to handle the other 20-25% of the problems without
asking for money to upgrade my hardware.
--
George N. White III <aa056@chebucto.ns.ca>
|
|
|
|
| Re: Introducing FL [message #48246 is a reply to message #48144] |
Tue, 04 April 2006 12:38 |
JD Smith
Messages: 850
Registered: December 1999
|
Senior Member |
|
|
> I use IDL now because we have a large base of IDL code and it
> would be a pain to translate it all to python or anything else.
> While I appreciate IDL's capabilities, I often cringe as I
> reinvent something that python has had built in for years. On
> the other hand, I cheer as I see RSI slowly adding features that
> most other languages have always had (see for, example,
> command_line_args and persistent command line history that appear
> in IDL 6.2). For basic usability and libraries, IDL just doesn't
> stack up. For specialized libraries and graphics, IDL is great
> (iTools aside, that is!).
>
> It looks like Python-IDL may be a wonderful way to mix the two,
> but I'm a little afraid to try IDL widgets with it (well, ok, a
> lot afraid!)
So it's not really a replacement, since it requires a functioning IDL
(or GDL?) around to communicate with.
There are two basic issues at hand, as Craig has been arguing:
1. Is it feasible to replicate even a small fraction of all the
MATH/WIDGET/OBJECT/GRAPHICS/3D/MAPPING/NUMERIC routines which RSI
has included in IDL?
2. Even if it is feasible, is it a good idea?
I tend to think the answer to both of these is "No", which puts me
squarely in the camp of migration rather than replication. If, as
Craig opines, we could create an IDL->XXX translator which works most
of the time, and allows legacy code like AstroLib to function
perfectly well, then we've managed some fraction of the battle. The
rest of the battle is to get functional replacements for most of those
weird little corners of IDL that none of us uses everyday, but every
one of us have used with great effect and amazement on rare occasion.
JD
|
|
|
|
| Re: Introducing FL [message #48250 is a reply to message #48144] |
Tue, 04 April 2006 08:45 |
Craig Markwardt
Messages: 1869
Registered: November 1996
|
Senior Member |
|
|
"Y.T." <ytyourclothes@p.zapto.org> writes:
> Craig Markwardt wrote:
> [good stuff mercilessly snipped]
>>
>> True, but that's kind of my point. *If* we could easily translate all
>> our IDL code to Python (say), we'd have the immediate benefit of a
>> stable and supported runtime environment, where the wheel had already
>
>
> You're aware of PyDL, aren't you?
I'm also aware that PyDL development dropped off after a few months.
That is a risk when you have one or a few people developing a large
package.
> Speaking only for myself, I write stuff in IDL because I know IDL.
> That's all. That's where my "inertia" comes from. A converter that
> would produce perfect python from my IDL routines would be worthless to
> me.
I totally appreciate your sentiment. In fact, I feel the same way.
But if the translation happened behind the scenes and transparently
every time you modified your IDL .pro file, why would you need to care?
I'm not saying you should "switch" (though that was my option "a"). I
think it's possible to make the IDL->python->bytecode compilation fast
and automatic enough that people could keep their IDL language, even
though they were actually using python behind the scenes.
Craig
--
------------------------------------------------------------ --------------
Craig B. Markwardt, Ph.D. EMAIL: craigmnet@REMOVEcow.physics.wisc.edu
Astrophysics, IDL, Finance, Derivatives | Remove "net" for better response
------------------------------------------------------------ --------------
|
|
|
|
| Re: Introducing FL [message #48253 is a reply to message #48144] |
Tue, 04 April 2006 08:41 |
Craig Markwardt
Messages: 1869
Registered: November 1996
|
Senior Member |
|
|
"Richard G. French" <rfrench@wellesley.edu> writes:
> On 4/3/06 10:17 PM, in article onzmj2ccf7.fsf@cow.physics.wisc.edu, "Craig
> Markwardt" <craigmnet@REMOVEcow.physics.wisc.edu> wrote:
>
>>
>> [*] - I do appreciate the impressive effort that the GDL and FL
>> writers have put in!
>
>
> Out of curiosity, are there any legal issues involved in reproducing the
> functionality of IDL? I know that you can't 'decompile' code, and I suspect
> that FL and GDL couldn't sell their products, but is there any gotcha with
> simply writing separate code that does exactly what IDL does, and giving it
> away for free?
I don't see why not, it's [still] a free country and world.
Craig
--
------------------------------------------------------------ --------------
Craig B. Markwardt, Ph.D. EMAIL: craigmnet@REMOVEcow.physics.wisc.edu
Astrophysics, IDL, Finance, Derivatives | Remove "net" for better response
------------------------------------------------------------ --------------
|
|
|
|
| Re: Introducing FL [message #48254 is a reply to message #48144] |
Tue, 04 April 2006 08:21 |
FL
Messages: 17
Registered: April 2006
|
Junior Member |
|
|
Well, I use FLTK (http://www.fltk.org). It's small, portable and
efficient. And it has FL in it's name :-) (although it is pure
coincidence)
As far as OS X is concerned, Mac's are very rare in this corner of
the world. (Is someone from Apple reading this newsgroup? :-)
regards,
lajos
On Tue, 4 Apr 2006, Ken Mankoff wrote:
>
> On Tue, 4 Apr 2006, FL wrote:
>> In the next few months I will clean up code, fixing bugs and
>> extending usability. If you have any ideas about features that the
>> language should have, but never dared to ask for, now it's time to
>> speak. Speak now or forever hold your peace :-)
>
> OS X port. Use Qt or wxWidgets so there are native windows and
> widgets rather than X11/Motif.
>
> -k.
>
>
|
|
|
|
| Re: Introducing FL [message #48256 is a reply to message #48144] |
Tue, 04 April 2006 08:16 |
mmiller3
Messages: 81
Registered: January 2002
|
Member |
|
|
>>>> > "FL" == FL <fl@interware.hu> writes:
> In the next few months I will clean up code, fixing bugs
> and extending usability. If you have any ideas about
> features that the language should have, but never dared to
> ask for, now it's time to speak. Speak now or forever hold
> your peace :-)
Please, an open source license. Otherwise I don't see how it can
compete against RSI's IDL.
Mike
|
|
|
|
| Re: Introducing FL [message #48257 is a reply to message #48144] |
Tue, 04 April 2006 08:08 |
mmiller3
Messages: 81
Registered: January 2002
|
Member |
|
|
>>>> > "Y" == Y T <ytyourclothes@p.zapto.org> writes:
> You're aware of PyDL, aren't you?
Be carefull with the capitalization and overlap between TLAs and
MLA - especailly when talking about "IDL".
There is PyDL, as in Python Interface Declaration Language
(www.prescod.net/pytypes) which is a different sort of IDL.
There is PyDL, as in a plotting package inspired by the plot
routine in RSI's IDL (vingilot.org/pydl).
There is also a link on the PyDL page to pyIDL. It says it is
based on PyDL "for IDL embedding in python, and has more
functionality." At that page
(www.its.caltech.edu/~mmckerns/software.html), it says pyIDL is
"based on Andrew McMurry's python-IDL"
Python-IDL (www.astro.uio.no/~mcmurry/python-idl) actually looks
somewhat promising. It provides a python module that lets python
code interact with an IDL session, as in
>>> import idl
>>> idl.ex('print, 1')
1
>>> idl.ex("a = 'Hello'")
>>> print idl.var.a
Hello
>>> print idl.fn.strlen("Hello")
5
> (I work for a federally funded research facility and you,
> the tax-payer, are expending my salary. Do you really think
> you're getting your money's worth out of the deal if I
> spend my time learning this years fad-language?)
As a (mostly) federally funded researcher, I had to learn IDL
when I moved to the medical field where it is the norm. In my
previous incarnation, python was by far the better choice because
it is 1) open source, and 2) much more capable.
I use IDL now because we have a large base of IDL code and it
would be a pain to translate it all to python or anything else.
While I appreciate IDL's capabilities, I often cringe as I
reinvent something that python has had built in for years. On
the other hand, I cheer as I see RSI slowly adding features that
most other languages have always had (see for, example,
command_line_args and persistent command line history that appear
in IDL 6.2). For basic usability and libraries, IDL just doesn't
stack up. For specialized libraries and graphics, IDL is great
(iTools aside, that is!).
It looks like Python-IDL may be a wonderful way to mix the two,
but I'm a little afraid to try IDL widgets with it (well, ok, a
lot afraid!)
Mike
P.S. As I typed this, I discovered that iTools can be mistyped
as iTrolls ;O
--
Michael A. Miller mmiller3@iupui.edu
Imaging Sciences, Department of Radiology, IU School of Medicine
|
|
|
|
| Re: Introducing FL [message #48258 is a reply to message #48144] |
Tue, 04 April 2006 08:07 |
mmiller3
Messages: 81
Registered: January 2002
|
Member |
|
|
>>>> > "Y" == Y T <ytyourclothes@p.zapto.org> writes:
> You're aware of PyDL, aren't you?
Be carefull with the capitalization and overlap between TLAs and
MLA - especailly when talking about "IDL".
There is PyDL, as in Python Interface Declaration Language
(www.prescod.net/pytypes) which is a different sort of IDL.
There is PyDL, as in a plotting package inspired by the plot
routine in RSI's IDL (vingilot.org/pydl).
There is also a link on the PyDL page to pyIDL. It says it is
based on PyDL "for IDL embedding in python, and has more
functionality." At that page
(www.its.caltech.edu/~mmckerns/software.html), it says pyIDL is
"based on Andrew McMurry's python-IDL"
Python-IDL (www.astro.uio.no/~mcmurry/python-idl) actually looks
somewhat promising. It provides a python module that lets python
code interact with an IDL session, as in
>>> import idl
>>> idl.ex('print, 1')
1
>>> idl.ex("a = 'Hello'")
>>> print idl.var.a
Hello
>>> print idl.fn.strlen("Hello")
5
> (I work for a federally funded research facility and you,
> the tax-payer, are expending my salary. Do you really think
> you're getting your money's worth out of the deal if I
> spend my time learning this years fad-language?)
As a (mostly) federally funded researcher, I had to learn IDL
when I moved to the medical field where it is the norm. In my
previous incarnation, python was by far the better choice because
it is 1) open source, and 2) much more capable.
I use IDL now because we have a large base of IDL code and it
would be a pain to translate it all to python or anything else.
While I appreciate IDL's capabilities, I often cringe as I
reinvent something that python has had built in for years. On
the other hand, I cheer as I see RSI slowly adding features that
most other languages have always had (see for, example,
command_line_args and persistent command line history that appear
in IDL 6.2). For basic usability and libraries, IDL just doesn't
stack up. For specialized libraries and graphics, IDL is great
(iTools aside, that is!).
It looks like Python-IDL may be a wonderful way to mix the two,
but I'm a little afraid to try IDL widgets with it (well, ok, a
lot afraid!)
Mike
P.S. As I typed this, I discovered that iTools can be mistyped
as iTrolls ;O
--
Michael A. Miller mmiller3@iupui.edu
Imaging Sciences, Department of Radiology, IU School of Medicine
|
|
|
|
| Re: Introducing FL [message #48259 is a reply to message #48144] |
Tue, 04 April 2006 07:33 |
Ken Mankoff
Messages: 158
Registered: February 2000
|
Senior Member |
|
|
On Tue, 4 Apr 2006, FL wrote:
> In the next few months I will clean up code, fixing bugs and
> extending usability. If you have any ideas about features that the
> language should have, but never dared to ask for, now it's time to
> speak. Speak now or forever hold your peace :-)
OS X port. Use Qt or wxWidgets so there are native windows and
widgets rather than X11/Motif.
-k.
|
|
|
|
| Re: Introducing FL [message #48260 is a reply to message #48144] |
Tue, 04 April 2006 06:53 |
Michael Wallace
Messages: 409
Registered: December 2003
|
Senior Member |
|
|
Y.T. wrote:
> You're aware of PyDL, aren't you?
Hmmm, Python Interface Declaration Language? I find SciPy
(www.scipy.org) much more interesting and more in line with this thread.
:-)
-Mike
|
|
|
|
| Re: Introducing FL [message #48261 is a reply to message #48144] |
Tue, 04 April 2006 06:22 |
FL
Messages: 17
Registered: April 2006
|
Junior Member |
|
|
Hi guys!
Now the April Fool's Day is over, so let's get back to work. People never
can decide whether I am joking or not, and this was the case here, too.
Announcing FL on April 1 was a joke, but FL itself is a serious project.
FL is really compatible with IDL. FL really runs faster than IDL. And FL
has features (eg PDF output, or line profiling) which IDL is missing.
Now you can run IDL programs under IDL, GDL and FL. That's good news.
Competition is good for the end users. Think of Intel and AMD. Without
AMD, our fastest CPU would be a 1.2 GHz Pentium III today (with a $1500
price tag :-) If FL's only effect will be that RSI creates a faster
virtual machine, then all of you will benefit from it.
In the next few months I will clean up code, fixing bugs and extending
usability. If you have any ideas about features that the language should
have, but never dared to ask for, now it's time to speak. Speak now or
forever hold your peace :-)
regards,
lajos
On Sat, 1 Apr 2006, FL wrote:
> Hi guys!
>
> The first public snapshot of Fawlty Language, an IDL compatible compiler
> is available. (IDL is a registered trademark of Research Systems Inc.)
>
> FL binaries can be freely used and downloaded from the homepage
>
> http://web.interware.hu/fl
>
>
> Available packages:
>
> fl_0.6-i686-linux.tar.gz (Linux, Pentium II.
> kernel 2.4/2.6, glibc >= 2.2.5)
>
> fl_0.6-amd64-linux.tar.gz (Linux, AMD Athlon64,
> kernel 2.6, glibc >= 2.3.3)
>
> and for the bravehearted:
>
> fl_0.6-i686-mingw.zip (Windows 2000/XP/?, Pentium II)
>
>
> Features:
>
> - all language elements are supported
>
> - multithreaded operators
>
> - array operations use MMX/SSE/SSE2, if available
>
> - module profiling
>
> - line profiling
>
> - about 280 library functions (more or less usable)
>
> - true-color (24 bit) direct graphics devices: X, WIN, PS, PDF, Z
>
> - run-time performance: for many programs, FL is faster than IDL
> (eg. the empty loop is three times faster in FL :-)
>
>
> Enjoy!
>
> fl
> (Földy Lajos)
>
>
> ps: Hungarian names are in reverse order. Wait, no! English names are
> in reverse order :-). So my name is Lajos Foldy in English.
>
|
|
|
|
| Re: Introducing FL [message #48262 is a reply to message #48144] |
Tue, 04 April 2006 05:23 |
Paul Van Delst[1]
Messages: 1157
Registered: April 2002
|
Senior Member |
|
|
Y.T. wrote:
> Craig Markwardt wrote:
> [good stuff mercilessly snipped]
>
>> True, but that's kind of my point. *If* we could easily translate all
>> our IDL code to Python (say), we'd have the immediate benefit of a
>> stable and supported runtime environment, where the wheel had already
>
>
>
> You're aware of PyDL, aren't you?
>
> Speaking only for myself, I write stuff in IDL because I know IDL.
> That's all. That's where my "inertia" comes from. A converter that
> would produce perfect python from my IDL routines would be worthless to
> me.
>
> At my place of employ I can get Matlab for free (i.e. site-license) but
> I'd be horribly worthless with Matlab because I'd be spending my time
> re-re-relearning when and where to place a comma or a semicolon and
> what was the syntax for a linear fit again? So I use GDL, because that
> allows me to get something *done*. As in *now*.
That, of course, is very important but....
>
> (I work for a federally funded research facility and you, the
> tax-payer, are expending my salary. Do you really think you're getting
> your money's worth out of the deal if I spend my time learning this
> years fad-language?)
Yes, actually (being a taxpayer who also happens to work at a federally funded research
facility). Learning new programming languages is not unlike learning new spoken languages
(except easier IMO). It broadens one's horizon to make different solution methodologies
available. In some cases they may help your work, others not. I would be a foolish
taxpayer if I expected you to not expend time learning new stuff that may make your job
easier and more efficient. And, of course, there's the "personal improvement" aspect -- ya
gotta be happy at what you do. :o)
paulv
p.s. I would love to get a matlab site license - even though I barely know enough matlab
for the "hello world" chestnut. Until then, learning Ruby and Python while I wait for
Fortran2003 compilers will have to do. :o)
--
Paul van Delst
CIMSS @ NOAA/NCEP/EMC
|
|
|
|
| Re: Introducing FL [message #48263 is a reply to message #48144] |
Tue, 04 April 2006 03:13 |
Ben Panter
Messages: 102
Registered: July 2003
|
Senior Member |
|
|
Y.T. wrote:
> Craig Markwardt wrote:
> [good stuff mercilessly snipped]
>> True, but that's kind of my point. *If* we could easily translate all
>> our IDL code to Python (say), we'd have the immediate benefit of a
>> stable and supported runtime environment, where the wheel had already
>
>
> You're aware of PyDL, aren't you?
Which I guess is again different to PDL?
http://pdl.perl.org/
So many flavours...
Ben
--
Ben Panter, Garching, Germany.
Email false, http://www.benpanter.co.uk
or you could try ben at ^^^^^^^^^^^^^^^
|
|
|
|
| Re: Introducing FL [message #48264 is a reply to message #48144] |
Mon, 03 April 2006 21:28 |
Y.T.
Messages: 25
Registered: December 2004
|
Junior Member |
|
|
Craig Markwardt wrote:
[good stuff mercilessly snipped]
>
> True, but that's kind of my point. *If* we could easily translate all
> our IDL code to Python (say), we'd have the immediate benefit of a
> stable and supported runtime environment, where the wheel had already
You're aware of PyDL, aren't you?
Speaking only for myself, I write stuff in IDL because I know IDL.
That's all. That's where my "inertia" comes from. A converter that
would produce perfect python from my IDL routines would be worthless to
me.
At my place of employ I can get Matlab for free (i.e. site-license) but
I'd be horribly worthless with Matlab because I'd be spending my time
re-re-relearning when and where to place a comma or a semicolon and
what was the syntax for a linear fit again? So I use GDL, because that
allows me to get something *done*. As in *now*.
(I work for a federally funded research facility and you, the
tax-payer, are expending my salary. Do you really think you're getting
your money's worth out of the deal if I spend my time learning this
years fad-language?)
cordially
Y.T.
--
Remove YourClothes before you email me.
|
|
|
|
| Re: Introducing FL [message #48265 is a reply to message #48144] |
Mon, 03 April 2006 20:51 |
Richard French
Messages: 173
Registered: December 2000
|
Senior Member |
|
|
On 4/3/06 10:17 PM, in article onzmj2ccf7.fsf@cow.physics.wisc.edu, "Craig
Markwardt" <craigmnet@REMOVEcow.physics.wisc.edu> wrote:
>
> [*] - I do appreciate the impressive effort that the GDL and FL
> writers have put in!
Out of curiosity, are there any legal issues involved in reproducing the
functionality of IDL? I know that you can't 'decompile' code, and I suspect
that FL and GDL couldn't sell their products, but is there any gotcha with
simply writing separate code that does exactly what IDL does, and giving it
away for free?
Dick French
PS This question is inspired in part by the ongoing suit about the DaVinci
Code, where the claim DOES involve $$$...
|
|
|
|
| Re: Introducing FL [message #48266 is a reply to message #48144] |
Mon, 03 April 2006 19:17 |
Craig Markwardt
Messages: 1869
Registered: November 1996
|
Senior Member |
|
|
JD Smith <jdsmith@as.arizona.edu> writes:
>> And finally, with this proliferation of "almost-IDL" clones (FL, IDL,
>> ANA, Yorick), I have to wonder why everybody has to re-invent the
>> wheel (see above). I see the GDL people meticulously toiling over
>> basic functionality, and it makes me wonder: wouldn't it have been
>> better to write an IDL-to-{Python,Ruby,Perl} translator, and get the
>> existing well tested script environment for free? And not only that,
>> but Python for science seem about to really take off. An infusion of
>> well-tested IDL libraries into that community might be enough
>> cross-fertilization to really get a non-proprietary version of an
>> IDL-like language going strong.
>
> Python has been set to "take off" in the astronomy field for about 8
> years, and it just hasn't really happened. In my view, it's just
> inertia. It took RSI 30 years to get to where they are, and they have
> built up an impressively large (if ungainly and not exactly stylish)
> bag of tricks.
> ...
True, but that's kind of my point. *If* we could easily translate all
our IDL code to Python (say), we'd have the immediate benefit of a
stable and supported runtime environment, where the wheel had already
been invented, and done well.
Then there are all sorts of possibilities. An individual could then
(a) translate all of their routines to Python and then just "switch"
to Python once and for all. Or (b) they could still think in IDL but
compile to Python whenever needed.
Python would gain a whole suite of great science software originally
written for IDL, and IDL users would gain the freedom to switch away
from a proprietary environment. Everybody would gain! (with very
little change in inertia)
I just don't think it's possible for the clone writers of GDL or FL
can sustain the effort long enough to write a complete new invention
of the wheel (interpretter environment plus library). [*] Why not let
the Python people do more than half of that?
Craig
[*] - I do appreciate the impressive effort that the GDL and FL
writers have put in!
--
------------------------------------------------------------ --------------
Craig B. Markwardt, Ph.D. EMAIL: craigmnet@REMOVEcow.physics.wisc.edu
Astrophysics, IDL, Finance, Derivatives | Remove "net" for better response
------------------------------------------------------------ --------------
|
|
|
|
| Re: Introducing FL [message #48270 is a reply to message #48144] |
Mon, 03 April 2006 14:59 |
b_gom
Messages: 105
Registered: April 2003
|
Senior Member |
|
|
Well, if this was an April fool's joke, it's above my head.
Anyhow, good work on a gargantuan project. Any plans for opening up the
source? If not, how do you plan on completing the project, in terms of
providing full support for all the internal functions, widgets, etc? It
seems to me that this would require either astronomical time scales, or
the effort of more than one mere mortal.
A brief sketch on the webpage showing the currently supported features
and the ones that are in development would be useful.
Brad
FL wrote:
> Hi guys!
>
> The first public snapshot of Fawlty Language, an IDL compatible compiler
> is available. (IDL is a registered trademark of Research Systems Inc.)
>
> FL binaries can be freely used and downloaded from the homepage
>
> http://web.interware.hu/fl
>
>
> Available packages:
>
> fl_0.6-i686-linux.tar.gz (Linux, Pentium II.
> kernel 2.4/2.6, glibc >= 2.2.5)
>
> fl_0.6-amd64-linux.tar.gz (Linux, AMD Athlon64,
> kernel 2.6, glibc >= 2.3.3)
>
> and for the bravehearted:
>
> fl_0.6-i686-mingw.zip (Windows 2000/XP/?, Pentium II)
>
>
> Features:
>
> - all language elements are supported
>
> - multithreaded operators
>
> - array operations use MMX/SSE/SSE2, if available
>
> - module profiling
>
> - line profiling
>
> - about 280 library functions (more or less usable)
>
> - true-color (24 bit) direct graphics devices: X, WIN, PS, PDF, Z
>
> - run-time performance: for many programs, FL is faster than IDL
> (eg. the empty loop is three times faster in FL :-)
>
>
> Enjoy!
>
> fl
> (Földy Lajos)
>
>
> ps: Hungarian names are in reverse order. Wait, no! English names are
> in reverse order :-). So my name is Lajos Foldy in English.
|
|
|
|
| Re: Introducing FL [message #48271 is a reply to message #48144] |
Mon, 03 April 2006 13:29 |
JD Smith
Messages: 850
Registered: December 1999
|
Senior Member |
|
|
> And finally, with this proliferation of "almost-IDL" clones (FL, IDL,
> ANA, Yorick), I have to wonder why everybody has to re-invent the
> wheel (see above). I see the GDL people meticulously toiling over
> basic functionality, and it makes me wonder: wouldn't it have been
> better to write an IDL-to-{Python,Ruby,Perl} translator, and get the
> existing well tested script environment for free? And not only that,
> but Python for science seem about to really take off. An infusion of
> well-tested IDL libraries into that community might be enough
> cross-fertilization to really get a non-proprietary version of an
> IDL-like language going strong.
Python has been set to "take off" in the astronomy field for about 8
years, and it just hasn't really happened. In my view, it's just
inertia. It took RSI 30 years to get to where they are, and they have
built up an impressively large (if ungainly and not exactly stylish)
bag of tricks. Since their recent emphasis has been in areas where
most astronomy people I know don't have too much interest (point and
click data exploration), things may change in the next 5 years. That
said, most people in this field are increasing their use of IDL now
(since prices have come down from quasi-unreasonable for the small
researcher).
JD
|
|
|
|
| Re: Introducing FL [message #48273 is a reply to message #48144] |
Mon, 03 April 2006 11:33 |
Foldy Lajos
Messages: 268
Registered: October 2001
|
Senior Member |
|
|
Hi,
On Mon, 3 Apr 2006, R.G. Stockwell wrote:
>
> PS how is FOS (Fawlty Operating System, a MSwin compatible system)
> coming along?
>
well, excellent idea, FOS means 'shit' in Hungarian :-) But I think MSwin
can't be surpassed.
regards,
lajos
|
|
|
|
| Re: Introducing FL [message #48277 is a reply to message #48144] |
Mon, 03 April 2006 09:19 |
news.qwest.net
Messages: 137
Registered: September 2005
|
Senior Member |
|
|
> "FL" <fl@interware.hu> wrote in message
> news:Pine.LNX.4.58.0603302022170.31657@bifur.rmki.kfki.hu...
> Hi guys!
> The first public snapshot of Fawlty Language, an IDL compatible compiler
> is available. (IDL is a registered trademark of Research Systems Inc.)
> (F�ldy Lajos)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Hmmm, the name does have "Fools" in it, leaving "dylaj".
Let's see, doesn't seem to be the hungarian word for april...
perhaps "adjyl" is april upside down (you know, the "d" becomes
a "p"... the "j' upside down is sorta "r" looking...)
Wow, this is an extremely elaborate april fools day joke.
The creation of a usenet profile going back to 1997...
...appears to actually be posted from Hungary....
making a webpage, ... and a putting up scads of code.
Wow. Bravo! Well Done, sir/madam!
Regards,
Sir Olaf Lop
:)
PS how is FOS (Fawlty Operating System, a MSwin compatible system)
coming along?
|
|
|
|
| Re: Introducing FL [message #48287 is a reply to message #48144] |
Sun, 02 April 2006 12:33 |
FL
Messages: 17
Registered: April 2006
|
Junior Member |
|
|
Hi,
I came from Mars, but it's a secret, don't tell anybody else :-)
(http://www.setileague.org/askdr/hungary.htm)
FL has nothing to do with GDL. It started as a learning project in 2000.
After twelve years of C programming, I decided to learn C++, and flex and
bison, too. The reverse Polish notation calculator from the bison manual
was the starting point, but it was too simple. I wanted to extend it, and
C or Fortran were too complex. IDL came to my mind, as a simple language
:-)
regards,
lajos
On Sun, 2 Apr 2006, Craig Markwardt wrote:
> FL <fl@interware.hu> writes:
>
>
>> Hi guys!
>>
>> The first public snapshot of Fawlty Language, an IDL compatible compiler
>> is available. (IDL is a registered trademark of Research Systems Inc.)
>>
>> FL binaries can be freely used and downloaded from the homepage
>>
>
> So this is not an April Fools joke?
>
> What is the relation to GDL?
>
> I'm a little confused that a person who has contributed to a few
> threads in the past few years has suddenly come out with an IDL
> "compiler". I.e. developing a parser, virtual machine, libraries,
> etc. is not a simple effort. Where did F"oldy Lajos come from? :-)
>
> And finally, with this proliferation of "almost-IDL" clones (FL, IDL,
> ANA, Yorick), I have to wonder why everybody has to re-invent the
> wheel (see above). I see the GDL people meticulously toiling over
> basic functionality, and it makes me wonder: wouldn't it have been
> better to write an IDL-to-{Python,Ruby,Perl} translator, and get the
> existing well tested script environment for free? And not only that,
> but Python for science seem about to really take off. An infusion of
> well-tested IDL libraries into that community might be enough
> cross-fertilization to really get a non-proprietary version of an
> IDL-like language going strong.
>
> Craig
>
>
>
>
|
|
|
|
| Re: Introducing FL [message #48288 is a reply to message #48144] |
Sun, 02 April 2006 12:31 |
biophys
Messages: 68
Registered: July 2004
|
Member |
|
|
Craig Markwardt wrote:
> FL <fl@interware.hu> writes:
>
>
>> Hi guys!
>>
>> The first public snapshot of Fawlty Language, an IDL compatible compiler
>> is available. (IDL is a registered trademark of Research Systems Inc.)
>>
>> FL binaries can be freely used and downloaded from the homepage
>>
>
> So this is not an April Fools joke?
I thought it was April Fools joke too. Lajos must have chosen the
release date on purpose. Thank GOD it is not a virus.:)
>
> What is the relation to GDL?
>
> I'm a little confused that a person who has contributed to a few
> threads in the past few years has suddenly come out with an IDL
> "compiler". I.e. developing a parser, virtual machine, libraries,
> etc. is not a simple effort. Where did F"oldy Lajos come from? :-)
>
> And finally, with this proliferation of "almost-IDL" clones (FL, IDL,
> ANA, Yorick), I have to wonder why everybody has to re-invent the
> wheel (see above). I see the GDL people meticulously toiling over
> basic functionality, and it makes me wonder: wouldn't it have been
> better to write an IDL-to-{Python,Ruby,Perl} translator, and get the
> existing well tested script environment for free? And not only that,
> but Python for science seem about to really take off. An infusion of
> well-tested IDL libraries into that community might be enough
> cross-fertilization to really get a non-proprietary version of an
> IDL-like language going strong.
>
> Craig
I'd like to see some thing like IDL Script Node for LabVIEW emergeing
from the community. Well, a faster and more versatile IDL-like language
is definitely the ultimate dream
|
|
|
|
| Re: Introducing FL [message #48289 is a reply to message #48144] |
Sun, 02 April 2006 12:15 |
biophys
Messages: 68
Registered: July 2004
|
Member |
|
|
thanks, Lajos
Now I am convinced that FL is really fast, faster than IDL6.2! It was
not a fair comparison in my previous post. That idl time_test was run
inside IDLDE and the fl time_test in gnome-terminal with transparency
enabled. It seems that enabling transparency does slow things down. Now
if I run both time_tests in gnome_terminal with no background. FL is
the clear winner! Very impressive!
IDL6.2 time_test3 0.89sec
FL0.6 time_test3 0.43sec
bp
FL wrote:
> Hi,
>
> On Sat, 1 Apr 2006, biophys wrote:
>
>> good job, Lajos! It does run as fast as idl6.2 with some of my old
>> code. i've noticed backingstore problems with my linux machine. device,
>> retain=2 does not work. i'd like to know how to set working dir/search
>> path/start_up file.
>>
>> cheers,
>> bp
>>
>
> Thanks!
>
> HELP, /LIB will list all system routines with available parameters.
> RETAIN for DEVICE is not allowed yet, but it is on my todo list.
>
> time_test needs write permission to the current directory. You can move
> around with CD, 'dir' (read the working dir with CD, CURR=curr & PRINT,
> curr). Search path is the normal !PATH. The start up file name can be
> given in the FL_STARTUP env. variable.
>
> So you can run time_test by
>
> CD, 'some_dir_with_write_permission'
> !PATH+=':/usr/local/rsi/idl/lib'
> TIME_TEST3
>
>
> regards,
> lajos
|
|
|
|
| Re: Introducing FL [message #48290 is a reply to message #48144] |
Sun, 02 April 2006 11:09 |
Craig Markwardt
Messages: 1869
Registered: November 1996
|
Senior Member |
|
|
FL <fl@interware.hu> writes:
> Hi guys!
>
> The first public snapshot of Fawlty Language, an IDL compatible compiler
> is available. (IDL is a registered trademark of Research Systems Inc.)
>
> FL binaries can be freely used and downloaded from the homepage
>
So this is not an April Fools joke?
What is the relation to GDL?
I'm a little confused that a person who has contributed to a few
threads in the past few years has suddenly come out with an IDL
"compiler". I.e. developing a parser, virtual machine, libraries,
etc. is not a simple effort. Where did F"oldy Lajos come from? :-)
And finally, with this proliferation of "almost-IDL" clones (FL, IDL,
ANA, Yorick), I have to wonder why everybody has to re-invent the
wheel (see above). I see the GDL people meticulously toiling over
basic functionality, and it makes me wonder: wouldn't it have been
better to write an IDL-to-{Python,Ruby,Perl} translator, and get the
existing well tested script environment for free? And not only that,
but Python for science seem about to really take off. An infusion of
well-tested IDL libraries into that community might be enough
cross-fertilization to really get a non-proprietary version of an
IDL-like language going strong.
Craig
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
