Re: IDLv5.4

[Craig Markwardt <craigmnet(at)cow.physics.wisc.edu>]

Greetings again--

Thanks to David and Jeffrey for sending me examples of SAVE files from
IDL version 5.4.

I can confirm that these files are not in a format compatible with
versions earlier than IDL 5.4.  Both files that were sent to me by
David and Jeffrey had a format version of "6".  For comparison, other
IDL 5's produce a file version of "5" and IDL 4 doesn't store a file
version.  There is at least one new record type, and the format of
some existing record types seems to have changed.  More research on
this is needed.

Therefore I can pretty safely conclude that this change was a
*conscious* one by RSI, and not an incidental bug.  It is virtually
guaranteed that files produced by IDL 5.4 will not be backward
compatible.

On the political side, I think all RSI guaranteed was that SAVE files
would be *forward* compatible, so they haven't broken any promises.
However they really should have withheld this change until the next
major revision, and especially not a minor revision.  The IDL 4 -> 5
conversion didn't even break backward compatibility I'm pretty sure.

And finally on the more positive side, I am still working on
documenting my code.  It can read and write SAVE files in the older
format, and in principle can be extended to support both formats.

Craig

Craig Markwardt <craigmnet@cow.physics.wisc.edu> writes:

> I've been following this discussion with some interest.  I've been
> slowly developing and documenting some tools to open, read,
> interrogate, and write IDL save files, separate from IDL itself.  For
> example I have an IDL procedure while can print a "directory" of the
> contents of a save file.  This might also be nice for programmers who
> would want to save and restore variables with SAVE files, and don't
> want to be stuck with the SAVE and RESTORE commands.
> 
> I would imagine that people would be interested in trying this
> out.  Is that true?
> 
> I would also like to see some examples of this IDL 5.4 behavior.
> Maybe I can figure out where it's going wrong.  Email of some short
> files would be fine.
> 
> Craig
> 
> 
> Jeffrey Newmark <newmark@eitv2.nascom.nasa.gov> writes:
> > Hi Ben and others,
> > 
> > I am having this problem with data, any array at all! I am
> > aware of problems previously with restoring complex structures, code
> > etc., but for many versions now I have not had the problem of
> > restoring a simple data array with an older version of IDL.
> > 
> > Ben Tupper wrote:
> > > 
> > > Hello,
> > > 
> > > I'm probably reminding you of something you already know, but only
> > > data in SAVEd files are IDL version independent.  Code is not version
> > > independent.  This behavior is noted in the online documentation for
> > > SAVE.
> > > 
> > > Are your saved sets of code or data?  If it's data that you can't
> > > RESTORE, then we all better be listening.  (How many funky
> > > colortables have I saved as *.pal?)
> > > 
> > > Ben
> > > 
> > > Jeffrey Newmark wrote:
> > > 
> > > > Hi All,
> > > >
> > > > I was wondering if anyone has had problems reading save sets
> > > > made with IDL V5.4 on older versions? Is this a feature? I
> > > > can read old save sets on the new version. If indeed the savesets
> > > > are not backwards compatible, that is very annoying!
> > > >
> > > > IDL> print,!version
> > > > { alpha OSF unix 5.4 Sep 25 2000      64      64}
> > > >
> > > > thanks,
> > > > Jeff Newmark
> > > > SOHO NASA/GSFC
> > > 
> > > --
> > > Ben Tupper
> > > 248 Lower Round Pond Road
> > > POB 106
> > > Bristol, ME 04539
> > > 
> > > Tel: (207) 563-1048
> > > Email: PemaquidRiver@tidewater.net
> > 
> > -- 
> > ------------------------------------------------------------------
> > 
> > Dr. Jeffrey Newmark
> > SOHO - EIT Team Scientist
> 
> -- 
> --------------------------------------------------------------------------
> Craig B. Markwardt, Ph.D.         EMAIL:    craigmnet@cow.physics.wisc.edu
> Astrophysics, IDL, Finance, Derivatives | Remove "net" for better response
> --------------------------------------------------------------------------

-- 
--------------------------------------------------------------------------
Craig B. Markwardt, Ph.D.         EMAIL:    craigmnet@cow.physics.wisc.edu
Astrophysics, IDL, Finance, Derivatives | Remove "net" for better response
--------------------------------------------------------------------------