Re: Terminology proposals

On Tue, 25 Jan 2022 at 15:28, Bethan Tovey-Walsh <accounts@bethan.wales> wrote:

> A few notes:
>
> ixml processor.  Suggest delete the bracketed para - not a terminology item? Part of the spec?
>
>
> The bracketed paragraph is there because I think it’s useful to make explicit the limits of the previous sentence: a conformant processor must output vxml for *a* valid parse tree. This could be confusing, since it’s not immediately obvious that a processor doesn’t have to output any *specific* parse tree from an ambiguous parse.

A finer level distinction IMHO? Query if we can have 'descriptive'
bits in the spec? Always an element
of contention for a W3C document?
  OK to leave it in though.


>
> *ixml input string* used, but not defined.
>
>
> Placeholder definition added -  I daresay we’ll want to make changes after discussion.

Tks. Will wait for update Bethan.


>
> Is this resulting in a non-accepting state of the parse?
>
>
> I don’t think we entirely agreed yet whether a partial parse is a categorical failure or not in ixml. It’s not essential to the definition to say whether this is an accept state or not, I don’t think; but we could add that information if we make a decision on the status of partial parses in ixml. (Apologies if I’m misremembering, and we *did* make a decision already.)

I've no memory of such a decision.
By definition isn't it a fail /non-accepting finish?

>
> OK, I don't understand this use of prefix. Seems inappropriately
> overloaded here?
>
>
> I was mainly trying to clarify the original doc’s definition of “prefix recognition” by giving the technical definitions of “prefix” and “proper prefix”. These are very useful for defining a partial parse.

OK, I'll pass and let others comment.


>
> I like "visible XML", but less enamoured of "vxml", since it makes it sound like it is some special version of XML. I think it should just be "XML" (and it's the only place where XML is involved, as output).
>
>
> Pragmatically, I’d say that people *will* abbreviate “visible XML” to “vxml”. If we’re going to use the former (I’m glad you like it!), we may as well embrace the latter.

+1


>
> As Norm said, I think it’s essential to have a term for the XML which is output by ixml, as distinct from the XML you might want to generate downstream, which might have features such as schema inclusions, namespace declarations, and so on - which are (currently) not supported by ixml.

I've no mental model of this. I'll pass, wait for others to comment (@ndw?)


>
> I don’t think that the implication that this is a special kind of XML is a bad thing. It helps with explaining why things like schema inclusion aren’t permitted, because it draws a clear distinction between the output of the ixml process (which is a sort of intermediary format, intended to be a structured representation of the content of the input string) and the ultimate output you might want when you’re finished (which will often involve further processing of the ixml output, e.g. by adding other features to the XML (such as schemas and namespaces), or by transforming into another format entirely, etc.).

Is that outside the scope of ixml? Seems to be 'what I do with the
"intermediate" output mentioned?
Hence my deficient mental model?


>
> Maybe “visible markup” would be a better term, though? With e.g. “vMarkup” as an abbreviation?

-1 compared to vxml? (catchy)

>
> I would prefer just "ixml grammar", since an ixml grammar only exists to be an input to an ixml processor.
>
>
> Makes sense - let me know if the edits I’ve made work for you.

I'm left asking which one is this? I pass.

>
> "constructed from" doesn't work for me. I would prefer "that uses an ixml grammar to recognise an input string". But there is the possibility of confusion here, because we also use "ixml parser" to mean the thing that parses the ixml, rather than the input string (even though, by elaboration, an ixml parser¹ is an ixml parser².)

Is this a bit too deep for an initial set of terms? Agree it happens,
but (I think) only in the world of
processor writers?

regards



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

Received on Tuesday, 25 January 2022 16:41:51 UTC