Re: revised text and diagram

Thanks Michael.

On Wed, 12 Jan 2022 at 17:47, C. M. Sperberg-McQueen
<cmsmcq@blackmesatech.com> wrote:

> > Comments on the text please?
>
> In the paragraph on ixml grammars in XML, the text reads
>
> >    2. *ixml grammar XML file*. An ixml file specified in xml. Is there
> >    a schema of some ilk to which such an xml may tested? This
> >    expresses the language [alphabet?]  which may be used to constrain
> >    input files valid to that language.
>
> No official schema for the XML form of grammars exists; it is not hard
> to make one, and the schemas I have created are in the archive of this
> mailing list since I posted them in response the last time this question
> came up.  (Two schemas, one allowing extension attributes.)

Can I assume this will be in / referenced from the spec?
Suggest  *ixml grammar XML file*. An ixml file specified in xml and
valid to the appropriate schema.



>
> In the context of parsing and formal languages, 'alphabet' has a
> technical meaning, and it's better to avoid it in the final sentence
> above.  The grammar defines a language.  Possible alternatives to the
> final sentence:
>
>     This can be used to constrain the input.
>
>     This can be used to make explicit the structure implicit in the
>     input.
>
>     This defines a language and can be used to check and identify the
>     structure of input files in that language.

I prefer the last of these three, but none give any indication of compliance,
validity or some other way in which the *text input file* obeys the language
specified by the *ixml grammar file".   How about

"This defines a language and can be used to check the
   input files  against the language of the *ixml grammar file*"






>
> In describing the text input file I would steer clear of speaking about
> 'intent'.  Many uses of ixml will be parsing documents created for other
> purposes and often without any intent to follow any grammmar rules at
> all.

Suggest " A *text input file* Written by an ixml user which may be checked
against the *ixml grammar file* "

(On the assumption that if written for some other purpose it is
outside the scope
of the spec)

>
> There are no pragma files in ixml as currently specified, and no
> proposals for pragma files.  I would delete that box and that section of
> the text.

I disagree, though I will if all mention of pragmas disappears from
the meetings?
I am guessing that some wording will be drafted for pragmas in the near future?
Isn't that what you and Tom are working towards?



>
> I don't understand the description of the ixml processor; if it takes
> three inputs, what are they?  If it takes four, what are they?

Inputs are
One of xml or text grammar
input file
[pragma 'file' to be removed, wording on xml grammar to  include pragmas.]

These were the four files.

Proposed update on 1. *ixml grammar*

1.  An ixml grammar, expressed in either text (valid to the spec) or in
   an XML version thereof. Either may express the grammar to be used
   for processing a *text input file* into output. Optionally, pragmas may
be included to affect the grammar.

 Updated * ixml processor* definition.

* ixml processor*

An application taking as input a grammar and *text input file* and
producing the output. Its task is to validate the text input file
against the given grammar and if valid produce the *ixml output* following the
 grammar . Errors detected may [shall?] be produced and
shown to the user.

Unless otherwise directed, suggest 'shall' be used for errors.


I'll produce rev 0.e tonight.

regards



-- 
Dave Pawson
XSLT XSL-FO FAQ.
Docbook FAQ.

Received on Wednesday, 12 January 2022 18:54:54 UTC