RE: Property Life/Death

It is up to the server to decide what to do when it is asked to take a live
property and copy it dead. It could do things like put in an element which
says "This is a dead property" or change it into a string or copy its
current value or whatever it feels like. This is a property by property
issue.

As for your discomfort regarding a live property getting copied dead, you
are on the money. For example, as you pointed out, DAV defines a bunch of
properties as "live." However if you don't do an OPTIONS call and get back a
DAV: header with a "1" in it then you can't assume any of the DAV:*
properties are alive. Discovery is REQUIRED before assuming a property is
alive. Just having its name on the server is clearly not enough.

A reasonable extension to PROPFIND would be an argument which stated "Only
retrieve this value if it is alive." However I expect something like that to
come out of the schema definition/discovery work.

		Yaron

> -----Original Message-----
> From: Babich, Alan [mailto:ABabich@filenet.com]
> Sent: Wednesday, July 22, 1998 1:25 PM
> To: Yaron Goland; w3c-dist-auth@w3.org
> Subject: RE: Property Life/Death
> 
> 
> > If the client copies the resource to a server which doesn't 
> > know about the
> > system date property then the property will be dead. That is, the
> > destination server will not uphold the expectation that the 
> > retrieved value
> > will equal the current system date. Rather the server will 
> act like a
> > notepad, recording the value of the property when it was 
> > copied and not
> > changing it after that point.
> 
> OK. So I misunderstood dead properties, because
> I only concentrated on section 3.1.
> 
> A datetime property, originally live (a),
> could possibly be copied to another collection and
> become dead, but its datatype would apparently have 
> to be preserved, because you could retrieve 
> the value of the dead copy of the property. Right?
> But maybe not. Maybe string-izing it and saving it
> as a string is sufficient to act like a notepad?
> 
> If datatype must be preserved on a copy, dead 
> properties are not necessarily strings -- 
> apparently they can theoretically be practically 
> any datatype. Right?
> 
> Since the dead property's name is exactly the same, 
> but its semantics changed (the dead copy of the
> current date no longer tracks the date), I am 
> somewhat uncomfortable. It doesn't seem like it
> is exactly the same property. If the datatype
> can change on a dead copy (say, from a datetime
> to a string), then I'm even more uncomfortable.
> 
> Can live (b) properties (e.g., Age In Years) become 
> dead on a copy? If the datatype has to be preserved 
> on a copy, then I would say no. If not, then yes?
> 
> Alan Babich
>

Received on Wednesday, 22 July 1998 16:37:43 UTC