Re: data with metadata

"*** Hendrickson" <***.hendrickson@xxxxxxx> wrote in message
news:lGVHj.56066$D_3.24843@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

Gary Scott wrote:

It would be of tremendous value tho if there was a standardized database
interface for Fortran.

Doesn't the C interoperability feature do this? Essentially any
set of library routines that can be described in terms of the C
features can be called from Fortran. I think most of the major
databases have a C interface.

Uh, well, there's a question of whether C interoperability can ever
do anything. For most people out there at the present time, 32-bit
Windows is all there is in the world of computing. When you want to
interface to something in that environment, it almost inevitably
means to a *.DLL with STDCALL interfaces.

Now, Intel is waiting for someone else to make all the mistakes
before it considers STDCALL to be compatible with BIND(C), and I
don't know about the other commercial vendors, but I sent a bug
report to Andy about this and never heard back (although this was
during a period of inactivity, which seems to have ended) and the
gfortran group can't seem to consider it to be a bug or at least
a fixable bug. In tracking down a bug with Mozilla-based browsers
I ran across the dreaded bug 915:

https://bugzilla.mozilla.org/show_bug.cgi?id=915

The discussion there shows how an active open source project can
get hung up on an obvious crippling bug and not do anything about
it for a decade! I am hoping that gfortran doesn't experience
the same sort of problems with STDCALL interoperability, but for
sure it could happen, and others who are waiting for someone to
make the first move could stay backed up indefinitely as well.

Until movement it made by someone, BIND(C) will be hamstrung on the
most popular platform at the current time.

--
write(*,*) transfer((/17.392111325966148d0,6.5794487871554595D-85, &
6.0134700243160014d-154/),(/'x'/)); end


.