- From: <zongaro@ca.ibm.com>
- Date: Thu, 22 Mar 2001 14:52:42 -0500
- To: <xmlschema-dev@w3.org>
Hi Ashok, Thanks for your response. Although the "+" sign is prohibited from the mantissa, there doesn't appear to be any restriction against "+" in the exponent, which is the case I was asking about. Thus, it appears that both 1.5E+1 and 1.5E1 meet the criteria for the canonical representation, but my guess is that it was not intended that 1.5E+1 be permitted. In addition, the lexical values 1.0E0 and 1.000000000001E0 don't appear to break the rule against trailing zeros in the mantissa, so I still believe that they both meet the conditions for the canonical representation. That in spite of the fact that they both map to the same value in the value space. It will obviously be trickier to state rules that avoid this issue. Sorry if there's anything I've misunderstood. Thanks, Henry ------------------------------------------------------------------------ Henry Zongaro XML Parsers development IBM SWS Toronto Lab Tie Line 778-6044; Phone (416) 448-6044 mailto:zongaro@ca.ibm.com "Ashok Malhotra" <ashokma@microsoft.com>@w3.org on 2001/03/21 11:53:21 PM Please respond to "Ashok Malhotra" <ashokma@microsoft.com> Sent by: w3c-xml-schema-ig-request@w3.org To: Henry Zongaro/Toronto/IBM@IBMCA, <w3c-xml-schema-ig@w3.org>, <xmlschema-dev@s3.org> cc: Subject: RE: Canonical representation of float and double Section 3.2.3.2 of the spec says; "The canonical representation for float is defined by prohibiting certain options from the Lexical representation (§3.2.3.1). Specifically, the preceding optional "+" sign is prohibited from the mantissa. The exponent must be indicated by "E" and number representations must be normalized such that for non-zero numbers there is a single non-zero digit to the left of the decimal point. Leading and trailing zeroes are disallowed in the mantissa and leading zeroes are disallowed in the exponent. " So, the preceding + sign is prohibited and trailing zeroes are prohibited in the mantissa. This seems to be what you are asking for. Ashok -----Original Message----- From: zongaro@ca.ibm.com [mailto:zongaro@ca.ibm.com] Sent: Wednesday, March 21, 2001 11:24 AM To: w3c-xml-schema-ig@w3.org Subject: Canonical representation of float and double Hello, I posted the following to xmlschema-dev@w3.org yesterday, but I thought I should cross-post here, in case anyone on this list isn't monitoring the other. Thanks, Henry ------------------------------------------------------------------------ Henry Zongaro XML Parsers development IBM SWS Toronto Lab Tie Line 778-6044; Phone (416) 448-6044 mailto:zongaro@ca.ibm.com ---------------------- Forwarded by Henry Zongaro/Toronto/IBM on 2001/03/21 02:22 PM --------------------------- Henry Zongaro/Toronto/IBM@IBMCA@w3.org on 2001/03/20 02:49:01 PM Please respond to Henry Zongaro/Toronto/IBM@IBMCA Sent by: xmlschema-dev-request@w3.org To: xmlschema-dev@w3.org cc: Subject: Canonical representation of float and double Hello, The definition of the canonical lexical representation for the float data type in the Schema Datatypes PR is as follows ( http://www.w3.org/TR/xmlschema-2/#float): 3.2.4.2 Canonical representation The canonical representation for float is defined by prohibiting certain options from the Lexical representation (§3.2.4.1). Specifically, the exponent must be indicated by "E". Leading zeroes are prohibited in the exponent. For the mantissa, the preceding optional "+" sign is prohibited and the decimal point is required. Leading and trailing zeroes are prohibited subject to the following: number representations must be normalized such that there is a single digit to the left of the decimal point and at least a single digit to the right of the decimal point. Both lexical values 1.0E0 and 1.000000000001E0 appear to conform to the requirements of the canonical representation, but they both map to the same value in the value space. According to the definition of Canonical Lexical Representation (§2.3.1 - http://www.w3.org/TR/xmlschema-2/#canonical-lexical-representation), the mapping between the canonical representation and the value space should be one-to-one. I think the requirements on the canonical representation for float need to be more stringent. In addition, I believe it needs to be explicitly stated that the optional "+" preceding the exponent is prohibited in the canonical representation. The double type has these same problems. Thanks, Henry ------------------------------------------------------------------------ Henry Zongaro XML Parsers development IBM SWS Toronto Lab Tie Line 778-6044; Phone (416) 448-6044 mailto:zongaro@ca.ibm.com
Received on Thursday, 22 March 2001 14:52:00 UTC