| Re: HDF5 and IDL [message #38321] |
Thu, 04 March 2004 02:31  |
hcp
Messages: 41
Registered: August 1995
|
Member |
|
|
In article <MPG.1ab03f8b981bcf679896d5@news.frii.com>, David Fanning <david@dfanning.com> writes:
|> Organizations that write software for free spend even less
|> money and resources (is it possible!?) on technical writers
|> than reputable gouge-the-public software companies. So even
|> though there is an "official" release of new software, that
|> doesn't mean anyone can use it. Typically there are bugs,
|> the documentation appears to have been out-sourced to a
|> small African country, and sometimes you just run into
|> unexpected "features" that leave you utterly perplexed for
|> days. When you know no one is getting rich, you are more
|> tolerant. ("Harry probably knocked this out over his corned
|> beef sandwich the other day. Let's cut him some slack.")
This may be an accurate description of many free software projects.
It is rather unfair to imply that HDF5 is like this, though. HDF5 has
its problems: it is large, baroque, hard to compile the Fortran 9x
interface and has taken a while to stabilize. However it is a large,
centrally organized, widely used product and has sufficient documentation
that a sizeable chunk of the Pacific northwest is denuded of trees
every time someone prints it out.
Hugh
--
============S=u=p=p=o=r=t===D=e=b=i=a=n===http://www.debian. org==========
Dr. Hugh C. Pumphrey, School of GeoSciences
The University of Edinburgh, EDINBURGH EH9 3JZ, Scotland
OBDisclaimer: The views expressed herein are mine, not those of UofE.
============================================================ =============
|
|
|
|
| Re: HDF5 and IDL [message #38323 is a reply to message #38321] |
Thu, 04 March 2004 02:25  |
hcp
Messages: 41
Registered: August 1995
|
Member |
|
|
In article <51ee06c9.0403031720.1a4b56c1@posting.google.com>, reiche@stout.physics.ucla.edu (opa_rumple) writes:
|> Currently I am working on a postprocessor program in IDL reading HDF5
|> files. Everything was alright when I was using version 1.4.5 of the
|> HDF5 to link against my program, which is producing the HDF5 file.
|> However I recently upgraded to 1.6.2. Since then the IDL routines do
|> not recognize any datasets anymore (it still does groups and the file
|> itself). The version of the HDF in IDL is 1.4.3.
There was a substantial change in HDF5 between 1.4.x and 1.6.x, such that
files generated with 1.6.x are not readable by programs linked against 1.4.x.
(The reverse is OK: programs using 1.6.x can read older files.)
|> Is there a way to fix this problem, e.g. but upgrading IDL's HDF5
|> libraries or in creating the file in the other program, using the
|> latest version.
The easiest thing is to downgrade your HDF5 back to 1.4.x until such time as
a new release of IDL appears which is built against 1.6.x. I believe that if
you approach RSI directly, they will give you a loadable module linked against
1.6.x .
Hugh
--
============S=u=p=p=o=r=t===D=e=b=i=a=n===http://www.debian. org==========
Dr. Hugh C. Pumphrey, School of GeoSciences
The University of Edinburgh, EDINBURGH EH9 3JZ, Scotland
OBDisclaimer: The views expressed herein are mine, not those of UofE.
============================================================ =============
|
|
|
|
| Re: HDF5 and IDL [message #38327 is a reply to message #38323] |
Wed, 03 March 2004 18:17 |
David Fanning
Messages: 11724
Registered: August 2001
|
Senior Member |
|
|
opa_rumple writes:
> Currently I am working on a postprocessor program in IDL reading HDF5
> files. Everything was alright when I was using version 1.4.5 of the
> HDF5 to link against my program, which is producing the HDF5 file.
> However I recently upgraded to 1.6.2. Since then the IDL routines do
> not recognize any datasets anymore (it still does groups and the file
> itself). The version of the HDF in IDL is 1.4.3.
>
> Is there a way to fix this problem, e.g. but upgrading IDL's HDF5
> libraries or in creating the file in the other program, using the
> latest version.
Organizations that write software for free spend even less
money and resources (is it possible!?) on technical writers
than reputable gouge-the-public software companies. So even
though there is an "official" release of new software, that
doesn't mean anyone can use it. Typically there are bugs,
the documentation appears to have been out-sourced to a
small African country, and sometimes you just run into
unexpected "features" that leave you utterly perplexed for
days. When you know no one is getting rich, you are more
tolerant. ("Harry probably knocked this out over his corned
beef sandwich the other day. Let's cut him some slack.")
I'm telling you this not to enlighten you to the real world,
but to point out why you are not likely to find the latest
up-to-date version of a library in commercial software. Not
enough hours in the day, mostly. And let'em fix their
own damn bugs. Why should I be doing it for them! :-)
Not to mention when paying customers find things don't
work they, uh, complain a bit more vociferously.
HDF5 is important to RSI. They will get the latest libraries
built into IDL as soon as they can. I wouldn't look for
"updates". Rather, I would wait (and pray) for the regular
releases to feature it. (Sometime this summer, is what I hear
for the next IDL release. I've no idea about HDF5 1.6.2.)
Cheers,
David
P.S. Let's just say if I were going to be cutting edge,
I'd be getting on the Mac bandwagon. But I have enough
problems as it is, so....
--
David Fanning, Ph.D.
Fanning Software Consulting
Coyote's Guide to IDL Programming: http://www.dfanning.com/
|
|
|
|
| Re: HDF5 and IDL [message #38405 is a reply to message #38323] |
Thu, 04 March 2004 10:37 |
reiche
Messages: 7
Registered: November 2003
|
Junior Member |
|
|
Thanks for your help.
Fortunately I didn't deinstall my previous version of HDF5 and the
installation on our Beowulf Cluster is the also compatible with IDL.
Sven
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
