Related articles |
---|
Summary of responses - Conversion to C++ sivar@SanDiego.NCR.COM (SSR) (1994-08-01) |
Re: Summary of responses - Conversion to C++ wws@renaissance.cray.com (1994-08-04) |
Re: Summary of responses - Conversion to C++ bobduff@dsd.camb.inmet.com (1994-08-06) |
Re: Summary of responses - Conversion to C++ jan@neuroinformatik.ruhr-uni-bochum.de (1994-08-10) |
Re: Summary of responses - Conversion to C++ sastdr@unx.sas.com (1994-08-17) |
Newsgroups: | comp.compilers |
From: | jan@neuroinformatik.ruhr-uni-bochum.de (Jan Vorbrueggen) |
Keywords: | arithmetic, design |
Organization: | Institut fuer Neuroinformatik, Ruhr-Universitaet Bochum, Germany |
References: | 94-08-018 94-08-057 |
Date: | Wed, 10 Aug 1994 13:30:16 GMT |
wws@renaissance.cray.com (Walter Spector) writes:
Of all the things you have mentioned, isn't the above mentioned lack
of a decimal data type the most problematic? Decimal data types are
are fundamental and important in most COBOL applications. I have a
hard time seeing how this could be done efficiently in C/C++.
[Cobol doesn't require decimal arithmetic, although it makes the
implementation easier. Careful use of fixed point with scaling should
be adequate. -John]
And indeed, DEC says transitioning COBOL support from VAXen (which have
decimal arithmetic in hardware) to AXPs (which don't) was made easy because
the AXP's 64 bit integers have enough precision for the COBOL standard that
they can be used to implement decimal arithmetic.
Jan
--
Return to the
comp.compilers page.
Search the
comp.compilers archives again.