Next message: Geoffrey M. Clemm: "Re: Naive question"
To: ietf-dav-versioning@w3.org
Message-ID: <OFEC21658E.CC4B3C75-ON85256958.008360D0@raleigh.ibm.com>
From: "Jim Amsden/Raleigh/IBM" <marjorie@us.ibm.com>
Date: Tue, 12 Sep 2000 20:16:49 -0400
Subject: Re: Naive question
See <jra> tags. But the primary point is that clients will expect the
version selector to redirect to the target for contents whether the target
is the default (specified by the DAV:target property of the version
selector), or explicitly specified in the Target-Selector header. It is
reasonable that PROPFIND should behave the same way, and except for live
properties, the spec says it does. I'm only suggesting that we remove this
special case.
The list of cases you describe below indicate that we are getting close to
creating too many special cases in the protocol. Version selectors have
semantics like bind/redirect references, but are not bindings or redirect
reference resources. So they end up introducing a lot of special cases that
make the protocol more complicated. Some of these will be unavoidable. I'm
just looking for ways to reduce them.
"Geoffrey M. Clemm"
<geoffrey.clemm@rational To: ietf-dav-versioning@w3.org
.com> cc:
Sent by: Subject: Re: Naive question
ietf-dav-versioning-requ
est@w3.org
09/12/2000 05:42 PM
From: "Jim Amsden/Raleigh/IBM" <jamsden@us.ibm.com>
I think there are some problems with what you describe below. Any
operation on a version selector should be redirected to the target
version including PROPFIND/PROPPATCH regardless if a Target-Selector
header was specified or not.
Why?
<jra>
If a method operates on the target of a version selector, it should operate
on that target regardless of how the target is specified, explicitly in the
Target-Selector or implicitly in the DAV:target property of the version
selector.
</jra>
Some of the reasons that operations on a version selector are not
just redirected to the current target version include:
- a MOVE of a version selector is different from a MOVE of the
target version
- a DELETE of a version selector is different from a DELETE of the
target version
<jra>
This is not described in the spec. MOVE and DELETE introduced the most
difficult issues in the binding protocol. I hope we're not re-introducing
all those problems with version selectors.
</jra>
- a LOCK of a version selector is different from a LOCK of the
target version (the lock properties appear on the version selector,
not on the version)
<jra>
Then LOCK is ambiguous. I don't like the protocol using very different
semantics on the same target URL based on control couples in headers.
Headers should parameterize the operation, not change its meaning entirely.
As with the direct and redirect references arguments, if a resource needs
to be a resource sometimes, and a binding to another resource some other
times, we need to expose clear protocol for distinguishing them. This is
not the case for version selectors because sometimes the absence of a
Target-Selector causes the method to operate on the version selector, and
sometimes it doesn't. Maybe we need a header to say the method operates on
the version selector, not the target as in
Target-Selector:version-selector. Then MOVE or DELETE will (usually) fail
if the target is a version not the version selector. Then no special
cases. This is ugly too, but at least it could be consistent.
</jra>
- the members of a collection version selector are version selectors
and versionable resources, while the members of a collection version
are version histories (this is discussed at length in several earlier
threads in this mailing list).
- for efficiency (and other reasons), many servers will want to have
certain live properties of a version selector to be distinct from
those of the target version, so that the target version does not have
to be queried/updated whenever the version selector is (and vica
versa).
<jra>
This depends entirely on the properties the client asks for, not the
semantics of the protocol itself.
</jra>
The Target-Selector header is only there to
override the DAV:target of the version selector. We don't want different
semantics depending on how you specified you version.
I disagree. The semantics of a version selector are very different from
those of a version (see above). The Target-Selector header is there to
give you a convenient way to access a version resource, *NOT* as a
replacement or alternative for a version selector.
<jra>
The difference is not between a version selector and a version, but the
target of a version selector specified by its DAV:target property or
through the Target-Selector header. These both specify a version.
</jra>
Version selectors
are special resources, not bindings (too bad, but they aren't), so we
can define the semantics of PROPFIND/PROPPATCH on a version selector to
return both its properties and the properties of its target version.
Or we can define that the dead properties of a version selector are
identical to those of the current DAV:target version. Achieves the
desired result, without breaking the necessary distinction between
a version selector and a version.
<jra>
No problem with the dead properties, its the live properties that are at
issue. We should avoid defining the protocol so that clients have to use
different marshaling for some properties than others.
</jra>
There should be no overlap, so no properties will be hidden.
I think this is consistent with the definition of target in the core
versioning section.
Consistency with the definition of target is not the issue -- correct
version selector and version semantics is.
Cheers,
Geoff
From: Tim_Ellison@uk.ibm.com
How do you refer to a version selector rather than the version it
selects?
(i.e. to PROPFIND/PROPPATCH it's properties)
A version selector has a URL which is different from the URL
of any particular version. When you do a PROPFIND/PROPPATCH on a
version selector URL, you operate on the properties of the version
selector. When you do a PROPFIND/PROPPATCH on a version URL,
you operate on the properties of the version.
Note though that all the dead properties of a version selector
correspond (i.e. have the same value) as those of the version that
is that target of that version selector.
Further note that if you use a Target-Selector header with a
PROPFIND/PROPPATCH request, you operate on the properties of
the selected version, and *not* on the properties of the
version selector.
From: "Jim Amsden/Raleigh/IBM" <jamsden@us.ibm.com>
You can't. PROPFIND returns the properties of both.
PROPFIND returns the properties of whatever resource you
applied it to. In particular, the live properties of a version
selector can be different from those of its target version.
A while back, we modeled version selectors as a "redirector"
that redirected methods to the version, but that was a
good while ago, before we looked carefully at modeling
versioned collections.
Note that we can
create version selectors with VERSION-CONTROL, but DELETE doesn't
actually say you can delete them.
You can use DELETE to delete a version selector.
DELETE of a version is undefined.
That is correct, but DELETE of a version selector is defined.
DELETE on a version selector should just delete the version selector,
but then we have a special case where the version selector is
accessed
as a resource itself.
A version selector is always accessed as a resource itself.
A Target-Selector header can be used to redirect a request
from a version selector to the specified version, but without
a Target-Selector header, a method applied to a version selector
URL is applied to the version selector resource.
Remember the BIND/DELETE/UNBIND arguments? Geoff?
Yes, we used to need this kind of argument when we modeled
version selectors as redirectors, but we don't need them
any more, now that we don't model them that way anymore.
Cheers,
Geoff
--=_alternative 0072E64385256958_=
Content-Type: text/html; charset="us-ascii"
<br><font size=2 face="sans-serif">Geoff,</font>
<br><font size=2 face="sans-serif">I think there are some problems with
what you describe below. Any operation on a version selector should be
redirected to the target version including PROPFIND/PROPPATCH regardless if
a Target-Selector header was specified or not. The Target-Selector header
is only there to override the DAV:target of the version selector. We don't
want different semantics depending on how you specified you version.
Version selectors are special resources, not bindings (too bad, but they
aren't), so we can define the semantics of PROPFIND/PROPPATCH on a version
selector to return both its properties and the properties of its target
version. There should be no overlap, so no properties will be
hidden.</font>
<br>
<br><font size=2 face="sans-serif">I think this is consistent with the
definition of target in the core versioning section.</font>
<br>
<br>
<br>
<br><font size=2><tt> From: Tim_Ellison@uk.ibm.com<br>
</tt></font>
<br><font size=2><tt> How do you refer to a version selector
rather than the version it selects?<br>
(i.e. to PROPFIND/PROPPATCH it's properties)<br>
</tt></font>
<br>
<br><font size=2><tt>A version selector has a URL which is different
from the URL<br>
of any particular version. When you do a PROPFIND/PROPPATCH on
a<br>
version selector URL, you operate on the properties of the version<br>
selector. When you do a PROPFIND/PROPPATCH on a version URL,<br>
you operate on the properties of the version.<br>
</tt></font>
<br><font size=2><tt>Note though that all the dead properties of a
version selector<br>
correspond (i.e. have the same value) as those of the version that<br>
is that target of that version selector.<br>
</tt></font>
<br><font size=2><tt>Further note that if you use a Target-Selector
header with a<br>
PROPFIND/PROPPATCH request, you operate on the properties of<br>
the selected version, and *not* on the properties of the<br>
version selector.<br>
</tt></font>
<br>
<br><font size=2><tt> From: "Jim
Amsden/Raleigh/IBM" <jamsden@us.ibm.com><br>
</tt></font>
<br><font size=2><tt> You can't. PROPFIND returns the
properties of both.<br>
</tt></font>
<br><font size=2><tt>PROPFIND returns the properties of whatever
resource you<br>
applied it to. In particular, the live properties of a version<br>
selector can be different from those of its target version.<br>
A while back, we modeled version selectors as a
"redirector"<br>
that redirected methods to the version, but that was a<br>
good while ago, before we looked carefully at modeling<br>
versioned collections.<br>
</tt></font>
<br><font size=2><tt> Note that we can<br>
create version selectors with VERSION-CONTROL, but DELETE
doesn't<br>
actually say you can delete them.<br>
</tt></font>
<br><font size=2><tt>You can use DELETE to delete a version
selector.<br>
</tt></font>
<br><font size=2><tt> DELETE of a version is undefined.<br>
</tt></font>
<br><font size=2><tt>That is correct, but DELETE of a version selector
is defined.<br>
</tt></font>
<br><font size=2><tt> DELETE on a version selector should
just delete the version selector,<br>
but then we have a special case where the version selector
is accessed<br>
as a resource itself.<br>
</tt></font>
<br><font size=2><tt>A version selector is always accessed as a resource
itself.<br>
A Target-Selector header can be used to redirect a request<br>
from a version selector to the specified version, but without<br>
a Target-Selector header, a method applied to a version selector<br>
URL is applied to the version selector resource.<br>
</tt></font>
<br><font size=2><tt> Remember the BIND/DELETE/UNBIND
arguments? Geoff?<br>
</tt></font>
<br><font size=2><tt>Yes, we used to need this kind of argument when we
modeled<br>
version selectors as redirectors, but we don't need them<br>
any more, now that we don't model them that way anymore.<br>
</tt></font>
<br><font size=2><tt>Cheers,<br>
Geoff<br>
</tt></font>
<br>
<br>
<br>
<br>
--=_alternative 0072E64385256958_=--