From Newsgroup: comp.lang.forth
On 5/11/26 05:20, minforth wrote:
Am 06.05.2026 um 23:14 schrieb Krishna Myneni:
Since David's original webpages no longer appear to be accessible I
have placed his last version (v0.5.5) of his IEEE 754 floation-point
word set(s) on my website. A subset of files from his folder, relevant
to the above proposal, are located at the link below.
https://ccreweb.org/documents/programming/forth/standards/forth-200x/
proposals/ieee-fp/dnw/
Please see the readme.txt file in this directory for an explanation of
its contents.
I am really impressed by all the good work put into these tests.
Also since the standard Forth test suite for fp-numbers is only
rudimentary.
Just for completeness (and because it is a bit out of my realm):
Are there any differences between the Forth test results and the results obtained by C standard libraries math.h and float.h ?
I'm not sure if David checked the Forth test results against any current
C library results. That's on my list of things to do after I understand David's tests a little better. For example, I have my own IEEE FP
comparison tests for special values and his test code also does
comparison tests, but I need to understand his notation before I can see whether or not our tests are identical.
Now that I have studied David's proposal a bit, I'm quite impressed with
it. I think it could be a good starting point for an optional IEEE Forth
word set. I also have a proposal on forth-standard.org, based largely on discussions I had with David, with the intent of starting out with a
smaller subset of his words. Mine doesn't have as much in depth research
as his proposal, and the terminology is not as precise. I'm thinking
about how to revise mine to align better with his larger proposal.
--
Krishna
--- Synchronet 3.22a-Linux NewsLink 1.2