[wg11] Part 28: Incompatible revision of the MAP directive

[Dave Loffredo]

Heidi,

If I understand these plans correctly, I am concerned from a WG11
perspective.  In particular, the line below:

HP> This change in effect makes it IMPOSSIBLE to build a post
HP> processor that can accept an arbitrary configuration file, and
HP> process XML data according to the resulting XML schema.  In
HP> general, the post processor would need specific knowledge of how
HP> to perform any/all such MAPped transforms.

If this is true, than P28 will not be a standard that is implemented,
but rather just a paper mechanism for "blessing" random XML schemas as
STEP by means of an unimplementable config language transformation.

We will no longer have common basis for exchange among our suite of
STEP APs as with P21, only custom point-to-point translators for XML
schemas that cannot be simply transformed into each other.

						- Dave





Original message below ------------------------------

>>>>> "HP" == Heidi L Preston <hpreston at ebmail.gdeb.com> writes:

HP> In response to several ballot comments, the Part 28 editing meeting has
HP> decided that no post processor will be required to accept an arbitrary
HP> configuration file, and process XML data according to the resulting XML
HP> schema.  Rather, the configuration language will be used to formally define
HP> the derivation of particular (perhaps normative) XML schemas from the
HP> corresponding EXPRESS schemas.  Germany has requested a more powerful MAP
HP> configuration directive to specify the mapping of an arbitrary EXPRESS
HP> defined data type to an arbitrary XML schema data type, e.g., "julian_date"
HP> to "xsd:dateTime" or "address" to CEFACT "Address".  The editing meeting
HP> proposes to accept this comment, and to change the current MAP directive to
HP> support it.

HP> This change in effect makes it IMPOSSIBLE to build a post processor that
HP> can accept an arbitrary configuration file, and process XML data according
HP> to the resulting XML schema.  In general, the post processor would need
HP> specific knowledge of how to perform any/all such MAPped transforms.

HP> This is not a problem for post processors for specific derived XML schemas,
HP> e.g., the proposed AP214 XML schemas, or the probably AP238 and AP239 XML
HP> schemas.  Those post processors will understand the interpretation of the
HP> (perhaps normative) XML schema.

HP> It is also not a problem for schema generator tools, in that generating the
HP> MAPped XML schema only does not require knowledge of how to perform the
HP> data transforms.

HP> We recognize that this situation will be incompatible with the previous
HP> requirements of the draft.  Does anyone have a problem with this
HP> generalization of the MAP directive?

HP> ***************************************************************************
HP> Heidi Preston
HP> Electric Boat Corporation
HP> 75 Eastern Point Road
HP> Groton, CT 06340
HP> (860) 433-8649