| Re: Compiling IDL ... ever likey ? [message #5611 is a reply to message #5518] |
Sun, 21 January 1996 00:00   |
thompson
Messages: 584
Registered: August 1991
|
Senior Member |
|
|
James Tappin <sjt> writes:
> It stikes me that the biggest technical problem would be handling the way IDL
> is able to change the type of variables without the user knowing explicitly
> that it is doing so. ...
Actually, there's nothing to keep a "compiled" version of an IDL program from
acting the same way as in the normal version of IDL. Such a compiled
executable would consist of two parts:
1. A program section, which contains a restricted version of the basic IDL
executable. It would be able to do everything IDL can do, except compile
procedures, or accept commands from a command line.
2. A data section, which contains the IDL procedures to be executed, in a
binary interpreted format--presumably the same format that IDL stores the
procedure in a SAVE file.
Such an object would act just like an executable--i.e. it would be a single
file that one would simply run--but it would preserve all the
interpretive-language advantages that IDL currently has. It would still be
IDL.
Some people might be disappointed that the performance would be the same.
There would not be the performance increase that compiled binaries typically
enjoy.
The principal objection to such a scheme would be that such an executable would
only be runable on the platform it was compiled for. If one wanted to be able
to run the program on a variety of platforms--e.g. Windows, Solaris, MacOS,
OpenVMS, etc.--one would have to compile it separately for each of those
platforms. Probably, most interest in a compiler would be for the MSWindows
and MacOS platforms.
William Thompson
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
