We still need words to capture what is in the two tables though. The spec today does not say what the parser or unparser does when asked to process zero occurrences of an element (whether stipulated by maxOccurs or dfdl:occursCount), and that's the behaviour given in the tables and accompanying notes. Regards Steve Hanson Architect, IBM DFDL Co-Chair, OGF DFDL Working Group IBM SWG, Hursley, UK smh@uk.ibm.com tel:+44-1962-815848 From: Mike Beckerle <mbeckerle.dfdl@gmail.com> To: Steve Hanson/UK/IBM@IBMGB Cc: DFDL-WG <dfdl-wg@ogf.org> Date: 17/11/2014 20:02 Subject: Re: [DFDL-WG] Action 271: semantic of maxoccurs '0' I looked through the Spec at each use of the word maxOccurs. I saw nowhere that had to be modified if we stipulate that maxOccurs can be zero. Hence, I think the only thing we need to do is (1) explicitly state maxOccurs can be zero, and I suggest we also provide a suggested use case. I think the use case would be to accept a prior version of a DFDL-described data format, but have validation fail for elements that are not supposed to be present anymore in a revised version of that schema. I believe section 5.2.1 is the right place for this stipulation. The use case can go in a footnote perhaps. Mike Beckerle | OGF DFDL Workgroup Co-Chair | Tresys Technology | www.tresys.com Please note: Contributions to the DFDL Workgroup's email discussions are subject to the OGF Intellectual Property Policy On Mon, Oct 13, 2014 at 10:39 AM, Steve Hanson <smh@uk.ibm.com> wrote: 271 Use of maxOccurs '0' (All) 2/9: Legal in XSDL and DFDL. One use case in XSDL is when deriving a complex type by restriction, where it is used to indicate that an element in the base type must not appear. Another use case could be if a schema undergoes a version revision that removes elements; perhaps clearer if the removed element gets maxOccurs '0' rather than omitting it. Steve has seen an instance of its use in a DFDL schema. So if DFDL continues to support it, need to document the behaviour for the various occursCountKinds and what happens to separator. Also behaviour of occursCount expression that returns 0. Proposal needed. ... 23/9: No progress IBM has discussed this internally and has the following proposal. Infoset refers to the augmented infoset. OCK maxOccurs '0' Parsing Unparsing fixed No occurrences looked for in the data (1) No occurrences looked for in the infoset or written (2) implicit No occurrences looked for in the data (1) No occurrences looked for in the infoset or written (2) expression occursCount occurrences expected. If any are found it is a validation error. Any number of occurrences expected in the infoset and written. If any are found it is a validation error. parsed Any number of occurrences expected in the data. If any are found it is a validation error. Any number of occurrences expected in the infoset and written. If any are found it is a validation error. stopValue Any number of occurrences expected in the data. If any are found it is a validation error. Any number of occurrences expected in the infoset and written. If any are found it is a validation error. OCK Parsing and occursCount '0' Unparsing and no occurrences in infoset expression No occurrences looked for in the data. (1) No occurrences written. (2) Notes (1) If sequence has a separator, no separator is looked for in the data when parsing. (2) If sequence has a separator, no separator is output when unparsing. Regards Steve Hanson Architect, IBM DFDL Co-Chair, OGF DFDL Working Group IBM SWG, Hursley, UK smh@uk.ibm.com tel:+44-1962-815848 Unless stated otherwise above: IBM United Kingdom Limited - Registered in England and Wales with number 741598. Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU -- dfdl-wg mailing list dfdl-wg@ogf.org https://www.ogf.org/mailman/listinfo/dfdl-wg Unless stated otherwise above: IBM United Kingdom Limited - Registered in England and Wales with number 741598. Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU