[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

Re: RE: The limits of XML mean the limits of my data world

• From: Marcus Reichardt <u123724@gmail.com>
• To: Roger L Costello <costello@mitre.org>
• Date: Tue, 31 May 2022 09:48:31 +0200

I guess the "limits of XML mean the limits of my data world" title is
appropriate then :) I mean what's "hugely important" in the markup
world anyway? Would the runaway success and ultimate failure of HTML
as our best bet for a broadly applicable format for digital text
qualify?

Let me clarify a bit how SGML CONCUR works - theoretically, because
neither OpenSP nor sgmljs.net SGML supports it. It's an optional
feature of SGML after all ;(

So in SGML you can write markup such as

    <(A|B|C)X><(A)Y>bla<(B)Z> bla </(A)Y> bla</(B)Z></(A|B|C)X>

where "(A)" and "(A|B|C)" etc. are *name groups* of document type
names. SGML can have multiple DOCTYPE declarations in the prolog, and
you specify which document to parse/process with (by default, the
first document type aka the base document type is active). When SGML
parses a start- or end-element tag, it ignores those that have an
inactive document type. The concept of "active" document types extends
to link processes (think stylesheets and/or processing profiles)
sharing the namespace with document types. You can use this feature
for multiple "views" over documents, and quite elaborate content
processing pipelines; for example, you can generate/project a table of
content, and perform similar content apps.

So you can see that SGML doesn't really "represent" overlapping markup
as a component model; rather, it merely allows delimiting text spans
using overlapping tags. As to overlapping things in general other than
text, in the SGML context it might be instructive to look at HyTime,
where such things are represented as time-bounded events with lots of
ID references. It might also be of interest that Charles Goldfarb, the
creator of SGML, as a passionate Jazz musician also created the
Standard Music Description Language, an early attempt at digitally
encoding music.

Have a nice day,
Marcus
sgml.io

On 5/30/22, Roger L Costello <costello@mitre.org> wrote:
> Hi Peter,
>
>
>
> If I understand correctly, you are saying that in the following XML
> document:
>
>
>
>     <l>A’ was a merry man—took up the child:</l>
>
>     <l>“Yea,” quoth he, “dost thou fall upon thy face?</l>
>
>     <l>Thou wilt fall backward when thou hast more wit;</l>
>
>     <l>Wilt thou not, Jule?” and, by my holidame,</l>
>
>     <l>The pretty wretch left crying and said “Ay.”</l>
>
>
>
> one might want to insert additional markup around the quoted sentences.
> Something like this:
>
>
>
>     <l>A’ was a merry man—took up the child:</l>
>
>     <l><quote>“Yea,”</quote> quoth he, <quote>“dost thou fall upon thy
> face?</l>
>
>     <l>Thou wilt fall backward when thou hast more wit;</l>
>
>     <l>Wilt thou not, Jule?”</quote> and, by my holidame,</l>
>
>     <l>The pretty wretch left crying and said <quote>“Ay.”</quote></l>
>
>
>
> Is that the idea?
>
>
>
> I can see how that is a problem for XML. XML does not allow
> start-tag/end-tag pairs that are not properly nested.
>
>
>
> So the overlapping problem is this: Sometimes it's desirable to allow
> start-tag/end-tag pairs that are not properly nested.
>
>
>
> Is that right?
>
>
>
> Truthfully, the concept of allowing start-tag/end-tag pairs that are not
> properly nested doesn't seem like, in the universe of data concepts, a
> hugely important concept. I guess that I am not seeing its importance.
>
>
>
> Are there data concepts that are not expressible by XML and are hugely
> important concepts?
>
>
>
> By the way, here is the URL to Jeni's blog article:
> http://www.jenitennison.com/2008/12/06/overlap-containment-and-dominance.html
>
>
>
> /Roger
>
>
>

• References:
  • Re: RE: The limits of XML mean the limits of my data world
    • From: Roger L Costello <costello@mitre.org>