julian@cernvax.cern.ch (julian bunn) (05/02/91)
This is posted on behalf of Mike Metcalf. Please address replies to
him directly; see below for the full email address.
Julian
======================================================================
Now that Fortran 90 is finally adopted and only awaiting formal
approval from ISO and ANSI, some of you might be interested in
two reports I have recently written:
'A derived-data type for data analysis', describing how to
develop a dynamic derived-data type using a histogramming
package as an example.
'Why Fortran 90?', giving a summary of reasons as to why
Fortran 90 should be adopted in preference to FORTRAN 77, and
even C++.
Paper or e-mail copies are available from me.
Also, there was a question recently about the extent to which
recent changes to the standards document impact the books on Fortran 90
already published. As far as our "Fortran 90 Explained" is concerned,
this was audited against S8.115 before publication and almost all
the changes passed last summer (as contained in S19) were incorporated
at the last moment. Changes to the standard since then have been largely
of an editorial nature and remain without any consequence for the book.
Most of the rest have been to nail down better the interaction between
'old' and 'new' features (pointers in COMMON, structures in EQUIVALENCE),
or define better obscure cases such as the use of optional dummy arguments
as optional actual arguments to intrinsic functions. This means that only
a few minor changes are required in some places (since we do try to offer
a complete description of the whole language). No-one is likely
to write an incorrect program because the book appeared before the
final standard.
Mike Metcalf
CERN METCALF@CERNVM.CERN.CH
1211 Geneva 23
Switzerland