Re: [xsl] How would you like a new axis? time::

Subject: Re: [xsl] How would you like a new axis? time::
From: Emmanuel Bégué <emmanuel.begue@xxxxxxxxx>
Date: Sun, 14 Mar 2010 13:20:56 +0100
This idea really seems to have hit a nerve (judging from the twitter wall).
But if one projects the "time" axis on a space axis, then the existing
xpath axis can do the work: past === preceding-sibling.

What is not clear to me (but I haven't read the paper yet) is whether
the time axis will be found in the same xml document or in another
(previous / ancient) version of the same document...?
If we're looking at document sets, how do we manage those sets (how
can we tell if a document is a member of a given set, and how is the
set ordered)?
If we're looking at one big document that stores different versions of
the same nodes, then we need mapping rules...?

What is also unclear is whether historization (is that a word?) can
happen at the node level; if a node changes then surely the tree
containing this node has changed as well; also, the tree can change
without changing any node content (reordering of the nodes): is the
position of the node in the tree part of the node identity...?

Regards,
EB



On Sun, Mar 14, 2010 at 11:36 AM, Dave Pawson <davep@xxxxxxxxxxxxx> wrote:
> On 14/03/10 10:33, Justin Johansson wrote:
>>
>> Dave Pawson wrote:
>>>
>>> Suggestion from Oracle (Dana Florescu) presentation at #xmlprague.
>>>
>>> Idea being to ask for a specific version or nodeset existant
>>> at a specific time in the past!
>>>
>>>
>> Great idea, though I propose it be a forward axis!
>
>
> http://www.dbis.ethz.ch/research/publications/timemachinexml.pdf
>
> we also introduce the following axes:  rst::,
> earlier::, past::, past-or-current::, current::, later::, future::,
> future-or-current::, last::, all-times:: and local::. For exam-
> ple, later::* navigates to the next version of the node time-
> line to which the current node belongs or the empty sequence
> if it does not exist.
>
>
> Which is all relative to a node which has a position on
> the timeline, which may be in the past, to give you a 'forward'/future
axis!
>
> Some neat ideas.
>
>
>
> regards
>
> --
> Dave Pawson
> XSLT XSL-FO FAQ.
> http://www.dpawson.co.uk

Current Thread