Is there a term for this?

[salmo.bytes]

I often find it useful to refer to elements in an XML document using a
limited XPath construct: the contacentated string of element
names only, from root to any particular node in the XML. Is there a
term for this
idea--for XPaths constructed from concatenated element names, without
attributes
included in any way?


/book[@name="xml_stuff"] ...not what I'm talking about

/book/chapter/paragraph/sentance/ ...this is what I mean, is there a
term for this?

 

[Malcolm Dew-Jones]

(e-mail address removed) wrote:

: I often find it useful to refer to elements in an XML document using a
: limited XPath construct: the contacentated string of element
: names only, from root to any particular node in the XML. Is there a
: term for this
: idea--for XPaths constructed from concatenated element names, without
: attributes
: included in any way?


: /book[@name="xml_stuff"] ...not what I'm talking about

: /book/chapter/paragraph/sentance/ ...this is what I mean, is there a
: term for this?

Perhaps an element path.

Or
an xpath expression

a location path

a compound location path

an abreviated compound location path

a predicate-less abreviated compound location path


and in this particular case,

an absolute predicate-less abreviated compound location path


Perhaps a "child-only path" or a "simple path".

 

[Philippe Poulard]

I often find it useful to refer to elements in an XML document using a
limited XPath construct: the contacentated string of element
names only, from root to any particular node in the XML. Is there a
term for this
idea--for XPaths constructed from concatenated element names, without
attributes
included in any way?


/book[@name="xml_stuff"] ...not what I'm talking about

/book/chapter/paragraph/sentance/ ...this is what I mean, is there a
term for this?

I personnally use the term "Canonical path" ; here is my definition :

==============
Canonical path
The ·canonical path· of a node is an XPath expression that starts
from the root node and descends from child to child until the node
(without using the "*" joker). The sole predicate allowed in a step is
the "indexing predicate", such as [n] where n is a non nul positive
integer ; this predicate is not involved if the step would return a
unique node. For example,
/xcl:active-sheet/xcl:logic[3]/xcl:if[2]/xcl:then/xclarse is a valid
·canonical path· . When a node test involved an attribute or an element,
the prefix used is not necessary the same used in the document, for
example when prefixes are redefined ; a canonical path can't be used by
an XPath engine safely if a namespace mapping is not provided.
==============

In order to process a canonical path safely with an XPath engine, I put
such canonical paths in a "path element", that hosts namespace mappings
needed in the expression.
--
Cordialement,

///
(. .)
-----ooO--(_)--Ooo-----
| Philippe Poulard |
-----------------------

 

[salmo.bytes]

Thanks. I've wanted to say "canonical xpath" myself.
But I thought that meant something else.

 

[C. M. Sperberg-McQueen]

I often find it useful to refer to elements in an XML document using a
limited XPath construct: the contacentated string of element
names only, from root to any particular node in the XML. Is there a
term for this
idea--for XPaths constructed from concatenated element names, without
attributes
included in any way?


/book[@name="xml_stuff"] ...not what I'm talking about

/book/chapter/paragraph/sentance/ ...this is what I mean, is there a
term for this?

SGML veterans frequently use the term 'fully qualified generic
identifier' for this. That term isn't in wide use now, but
I think I prefer it to 'canonical XPath'.

The first version of the TEI Guidelines (in a section
drafted by Steve DeRose) defines a similar method of
pointing to an element in which each step has a
position indicator; in XPath notation something like

/book[1]/chapter[4]/paragraph[32]/sentence[3]

The term given there is 'typed path' (as opposed to
the numeric path /1/4/32/3).

-C. M. Sperberg-McQueen
World Wide Web Consortium

 

[salmo.bytes]

Thank you everybody for your input.
It seems there is no commonly adopted term for an xpath
as I described above (an xpath containing only the concatenated element
names,
without including any possible attributes in any way).

For my own purposes, I now feel free to invent my own term, and refer
to this as an 'elemental xpath'

 

[Malcolm Dew-Jones]

(e-mail address removed) wrote:
: Thank you everybody for your input.
: It seems there is no commonly adopted term for an xpath
: as I described above (an xpath containing only the concatenated element
: names,
: without including any possible attributes in any way).

: For my own purposes, I now feel free to invent my own term, and refer
: to this as an 'elemental xpath'


which I guess is a more generalized version of the noble elemental xpath

/He/Ne/Ar/Kr/Xe/Rn

and closely related to the more complex possibilities of the molecular
xpath, e.g.

/H2O/NaCl


both being excellent paths for aspiring chemists to follow.