Subject:
|
Re: LDraw DAT certification for clones?
|
Newsgroups:
|
lugnet.cad.dev.org.ldraw
|
Date:
|
Mon, 3 Jan 2000 01:28:45 GMT
|
Viewed:
|
653 times
|
| |
|
|
In lugnet.cad.dev.org.ldraw, Jacob Sparre Andersen writes:
> Todd (not admin@lugnet.com :-) wrote:
>
> > The definition could say things like: a program which
> > writes LDraw DAT files -must- output LDU coordinates as
> > decimal integers or IEEE floating-point numbers, -may-
> > output these numbers to high precision, but -should- round
> > them to at most 2 significant digits past the decimal
> > point.
>
> I am not sure that the BP7 floating-point format exactly
> matches the IEEE format, but I will look it up.
Borland Pascal, Borland C/C++, and Microsoft C/C++ all use the IEEE format
for floating point numbers (float, double and extended are all IEEE defined)
because these are the formats supported by the 80x87 line math coprocessors.
(Even though Pentiums no longer have a separate coprocessor chip, the math
processing is still done by a separate subsystem)
And if you mean for numbesr to be rounded, there should be a test against
truncation. Truncation of decimal places causes nearly all known math
errors when using IEEE values because the IEEE format is inherently bad at
storing certain values.
|
|
Message is in Reply To:
 | | Re: LDraw DAT certification for clones?
|
| (...) Considered, yes. Done something, no. (...) We try to do that already. (...) How do we do that? (...) I am not sure that the BP7 floating-point format exactly matches the IEEE format, but I will look it up. (...) Yes. Also for easy testing of (...) (25 years ago, 14-Dec-99, to lugnet.cad.dev.org.ldraw)
|
3 Messages in This Thread:
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
