Re: Market for an IDL to Java Translator? [message #55054] |
Tue, 31 July 2007 03:55  |
mattf
Messages: 13 Registered: January 2007
|
Junior Member |
|
|
On Jul 30, 3:11 pm, CyclingGuy <ericdav...@netzero.net> wrote:
> I'm trying to gauge the need for an IDL to Java Translator.
>
> My proposed translator would take IDL source code and produce Java
> Source code. Initially (at least), it would be limited to IDL code
> that contained no GUI widgets and made no calls to shared libraries.
>
> My thought is to offer the translator as a service rather than a
> product. A customer would upload their code to a web site, and get
> back a trial jar file (a java executable that would run for a limited
> period of time) in return. If they were satisfied with the execution
> of the jar file (correctness and speed), they would be able to
> purchase the generated Java source code.
>
> My questions are (to those that would be interested in such a
> service):
> 1) How many lines of code do you have that you would be interested in
> converting to Java?
> 2) Would you be interested in this service if it was offered if it
> was offered at a price of $0.10/line of code?
> 3) ... at a price of $0.05/line of code?
> 4) ... at a price of $0.02/line of code?
> 5) ... at a price of $0.01/line of code?
> 6) Would you be more interested in an IDL to C translator?
>
> Thank you,
> Eric.
I have a feeling that the intersection of the set of professional IDL
users and professional Java users is pretty small, so there would be a
strong global constraint on the size of the market... FWIW, I'd guess
that the intersections between IDL and Fortran, or IDL and C/C++, or
even IDL and Python would be bigger.
|
|
|
|
|
Re: Market for an IDL to Java Translator? [message #55132 is a reply to message #55057] |
Tue, 31 July 2007 12:21  |
Paul Van Delst[1]
Messages: 1157 Registered: April 2002
|
Senior Member |
|
|
Brian Larsen wrote:
> Eric,
>
> while this is an interesting idea I think going the other way is what
> people are most interested in. A translator from C or Fortran (yuk)
> to IDL would be something I would pay for, but I imagine this being a
> nearly impossible task to make "good" array based IDL code from c or
> fortan.
Over the years I've converted a good portion of my old IDL code in Fortran95. Everyone has
a f95 compiler (or can get one for free), but not as many people have (or can afford) IDL
licenses. (Wasn't IDL originally written in Fortran -- well, FORTRAN -- way back in the
old olden days on VMS systems?)
Creating on-the-fly things in Fortran is obviously more difficult i.e. it's strongly typed
and you have to allocate arrays as required. In IDL you would do
m = 100 & n = 24
x = FLTARR(100,24)
y = x
whereas in f95 you would do,
REAL, PARAMETER :: zero = 0.0
REAL, ALLOCATABLE :: x(:), y(:)
INTEGER :: m, n
m = 100; n = 24
ALLOCATE( x(m,n), y(m,n) )
x = zero
y = x
Apart from the type declarations, same dog different leg.
(although, Fortran2003 doesn't do OO like IDL does. :o)
> While from a nice platform independent language (IDL) to another such
> language (java) w/o all the gui bells and whistles doesn't sell for
> me.
Agreed.
cheers,
paulv
--
Paul van Delst
|
|
|