- From: Cokus, Michael S. <msc@mitre.org>
- Date: Mon, 2 Nov 2009 10:14:25 -0500
- To: "antoine.mensch@odonata.fr" <antoine.mensch@odonata.fr>, "public-exi-comments@w3.org" <public-exi-comments@w3.org>
Hi Antoine, >thanks for spending the time clarifying the rationale for the EXI WG >position about floating point representation. It seems that this >position is backed by extensive testing, so it is clear that the >decimal-based representation is both space and processing efficient in >many cases, especially when the text representation of the data is >required at some point in the processing. However, we do have quite a >few machine-to-machine use cases where we intend to use EXI as an >alternative to hand-written binary encoding. In these cases, I believe >the base-2 representation may still be more efficient (assuming we use a >typed interface that will not require a conversion to the decimal >represention). You mention in your mail that EXI would support the >replacement of the default EXI representation by the IEEE one. Will that >be through the Data Representation Map mechanism, or as a new standard >option? This would be handled using the Data Type Representation Map. > >Thanks again for your time. > >Cheers > >Antoine Mensch Thanks very much, --mike Mike Cokus The MITRE Corporation 757-896-8553; 757-826-8316 (fax) 903 Enterprise Parkway, Hampton, VA
Received on Monday, 2 November 2009 15:15:07 UTC