RE: datatyping is not needed

Dale, welcome to the list.  I think that was your first email to the wG?

At 10:09 AM 7/21/98 PDT, Dale Lowry wrote:
>Let me take a stab at providing a use case for obscure dead properties.
>
>In a document management system you may have a rigid schema for a given
>repository. That schema may include live properties such as "Date Created"
>and dead properties such as "Comment." However, it may be possible to store
>other properties in the metadata. For example, "Date Published" could be a
>property that is stored with the document metadata by a custom application.
>This property is only known and manipulated by this external application. 
>
>In the above scenario you want the end user to know about, be able to update
> and be able to search the "Date Created" and "Client" properties but not the
> "Date Published" property. You certainly don't want "Date Published " to
> show up in the schema for that repository. 
>
>I would catagorize "Date Published" in the above example as an obscure
> dead property.

That's quite close to what I meant when I introduced the term, but not quite.

The differences are:

1) all famous dead properties are searchable by definition.  It is the
process of indexing them that makes them famous.  Recall that live
properties are those that the server enforces syntax or semantics of, and
dead are all others.  In order to index a property, a server has to know at
least a little about the syntax, but not really enough to make it live.  It
would be pointless to index a property that was not searchable.  (Minor
exception, they might not be searchable by *all* parties, but at least by
some.)

2) Obscure dead properties may also be searchable, although perhaps not
with the same speed as the famous.   Some servers may only advertise in
their schema the live and famous dead, but nothing in DASL prohibits them
from also advertising some or all the obscure dead.

3) famousness says nothing about whether a property is readable (with
PROPFIND) or updatable (via PROPPATCH).  At present, there is no access
control in WebDAV, so nothing (within either WebDAV nor DASL protocols)
defines whether a given property can be read or written by the client.  Any
such access control is accomplished by means outside the protocol.

Received on Tuesday, 21 July 1998 16:16:42 UTC