🚀 go-pugleaf

In <93018.171821U35395@uicvm.uic.edu> C. M. Sperberg-McQueen <U35395@uicvm.uic.edu> writes:
>Robert Hartman writes:
>> Yes, that's the crux.  What I mean by "exactly" is zero output from
>> diff.  One off-the-cuff way to rate the degree of conformance to the
>> standard is by the ratio of diff output lines to original (text) lines
>> for a large and complicated set of reference documents.
>
>I have *real* problems with the notion that an SGML processor should be
>expected or required (possibly even problems with the notion that it
>should be *allowed*) to preserve distinctions in the input form which
>the standard says quite explicitly do not affect the processing of the
>document.  Yes, it's handy in at least some cases (I use SGML comments
>for information I would just as soon keep around for the moment).  But
>the standard says very clearly that an end-tag inferred by the parser
>behaves the same way an end-tag explicitly given in the input stream.
>If we require SGML processors to preserve that distinction from import
>to export, we really are requiring something more than SGML support from
>them.
>
>-C. M. Sperberg-McQueen
> ACH / ACL / ALLC Text Encoding Initiative
> University of Illinois at Chicago
>

I think we have to make a distinction, as Michael does, between
the requirements of ISO 8879 and what I will call user requirements.
ISO 8879 does not require, for example, preservation of record
ends caused by markup (e.g., starting each attribute on a new line).
However, as a user, you may have a requirement *on your editor* to
preserve record ends caused by markup.  An editor that does not
preserve markup record ends (and that otherwise conforms) is a
conforming SGML application, but it may not meet your requirements,
which is between you and it.

The idea that a test of conformance would be the degree of
change between input and output ignores the above paragraph and
the specifics of the standard.

I feel well within my rights to state the user requirement that
an SGML editor preserve *all* aspects of the input document,
whether required by ISO 8879 or not, but I cannot condemn an
editor that does not meet those requirements as non conforming.

Eliot Kimber                      Internet:  drmacro@ralvm13.vnet.ibm.com
Dept E14/B500                     IBMMAIL:   USIB2DK9@IBMMAIL
Network Programs Information Development     Phone: 1-919-543-7091
IBM Corporation
Research Triangle Park, NC 27709