Re: [xml-dev] Does the XML syntax have an underlying data model?

[Peter Flynn <peter@silmaril.ie>]

On 04/16/2016 10:18 PM, Jim Tivy wrote:
> [...] I still find discussions of CDATA and PCDATA confusing and the
> idea of "entities" "parameterized entities" etc flew in the face of
> the relational modeling meaning of the words. Many of the founders of
> XML came from a document background and I come from a
> database/datamodel background.

I think this has a lot to do with it. I still have to explain to new
users that XML does not have "records" and "fields" per se, even though
some XML looks as if it has.

As the database engineer confronted with XML said, "I remember when all
this was fields" :-)

> But in the end, XML is a great spec for human readable documents - so
> perhaps you don't want some database dude in there at the get go
> screwing it up. I suppose, however, because of this new technical
> people may find the concepts of XML arcane - otherworldly - I did.

Having come from a programming job I found SGML a little hard to get to
grips with too, but I had a background in printing to help me. Now I
show new users a chunk of document XML like this and explain that it's
what XML was designed for: its use for data was an afterthought.

<p><mls unit="MS folio" n="108a"/> seiss for braine in churaig. Da
r&aacute;mma do iurn aithlegtha i ndib <mls unit="MS folio" n="108a
25"/> lamaib d&oacute; ic tescad in mara immi sain chan. No rannad
bledmila <lb n="12630"/> in mara immi do bressaib na ruadrama
s&uacute;as i n-ardda clethi. <ex>co</ex> tuittitis i frithdrui&ndot;g
na conaire cetna a dochum <ex>co m</ex>b&iacute;tis ar medon in churaig.
C<ex>o</ex> tibeds<ex>o</ex>m a ma&ndot;g&fdot;&aacute;itbiud gaire for
aird impu. <ex>co</ex> <ex>n</ex>digsed tre&sdot;ess cona nirt nonbair
dar ginchr&aacute;es d&oacute;. Comtar &eacute;cnaig <mls unit="MS
folio" n="108a 30"/> a thruim &ampersir; a glainene for ettegail dar
commur a chraes &ampersir; a br&aacute;get. <lb n="12
635"/> Samla beolo fidchuaich indara s&uacute;il d&oacute; fria chind
dianechtai r. acht mod ma 'dres<ex>ed</ex> corr in t&sdot;&uacute;il
anaill do ind a gr&uacute;ade. Samalta comchubrech daim amr&iacute;ata
ce<ex>ch</ex> finna sescnidi imluathi acarb sinis <mls unit="MS folio"
n="108a 35"/> tria fortrennaib a chr&uacute;adchind fri fonascad
&ndot;ga&iacute;the. Ba&iacute; da<ex>no</ex> L&aacute;eg coa
thuarascbail &ampersir; tairderc for aird. &ampersir; dorone in laed.
<lb n="12640"/></p>

Some of the documents have elements nesting a dozen levels deep and more
in cases of multiple comparative readings of manuscripts (that also puts
the hat on "XML is human-readable" :-)

///Peter