It is actually a correct claim, if you read the article I mentioned. It does not attempt to represent impossible things: When exact representation is impossible it represents a range that includes the correct value. The size of the range (and thus the loss of precision) is tracked, and grows far more slowly with repeated calculations than floating point numbers do. It really is a brilliant idea.
-James On Tuesday, May 3, 2016 at 8:03:01 PM UTC-5, Jason Felice wrote: > > Just to point out, the OP claimed the number format has no rounding or > precision problems, which is not possible. Think about ways to represent > pi for a bit! > > -Jason > > > On Tue, May 3, 2016 at 5:51 PM, James Elliott <brun...@gmail.com > <javascript:>> wrote: > >> Heh! I just read about these in an email from ACM, and found your post by >> searching this Group for UNUM. It looks like not yet, but perhaps we should >> do something about that? >> >> The article which excited me: >> http://ubiquity.acm.org/article.cfm?id=2913029 >> >> >> On Friday, November 6, 2015 at 8:23:35 AM UTC-6, Richard Davies wrote: >>> >>> Unum is a number representation system that is a superset of IEEE >>> integers and IEEE floats which avoids many problems (especially in floating >>> point arithmetic) as it has no rounding, no overflow to infinity, no >>> underflow to zero, and is safe to parallelize. >>> >>> I was wondering if anyone has implemented or is interested in >>> implementing this in Clojure/ClojureScript (or Java/JavaScript)? >>> >>> There is an existing reference implementation written in Mathematica and >>> it appears a port has already been done to Python: >>> https://github.com/jrmuizel/pyunum >>> >>> For more details of unum see >>> https://www.youtube.com/watch?v=jN9L7TpMxeA >>> >>> https://www.crcpress.com/The-End-of-Error-Unum-Computing/Gustafson/9781482239867 >>> http://arith22.gforge.inria.fr/slides/06-gustafson.pdf >>> >>> >>> -- >> You received this message because you are subscribed to the Google >> Groups "Clojure" group. >> To post to this group, send email to clo...@googlegroups.com >> <javascript:> >> Note that posts from new members are moderated - please be patient with >> your first post. >> To unsubscribe from this group, send email to >> clojure+u...@googlegroups.com <javascript:> >> For more options, visit this group at >> http://groups.google.com/group/clojure?hl=en >> --- >> You received this message because you are subscribed to the Google Groups >> "Clojure" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to clojure+u...@googlegroups.com <javascript:>. >> For more options, visit https://groups.google.com/d/optout. >> > > -- You received this message because you are subscribed to the Google Groups "Clojure" group. To post to this group, send email to clojure@googlegroups.com Note that posts from new members are moderated - please be patient with your first post. To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en --- You received this message because you are subscribed to the Google Groups "Clojure" group. To unsubscribe from this group and stop receiving emails from it, send an email to clojure+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
