Re: ETags, was: Issues from Interop/Interim WG Meeting

Well, it really all depends on how compelling the use cases are that require
ETag support.  It's ok to make servers do things that may be hard for some
implementations if it delivers solid benefits for the client.  I heard some
pretty solid use cases at the Interop event.

In your filesystem example, as long as all of the access to the files can be
intercepted by the server software somehow, keeping around an ETag is not a
big problem as you can bump up a sequence number on each write.  A simple
filesystem based server may also have problems with supporting UUIDs as
well.

--Eric

----- Original Message -----
From: "Julian Reschke" <julian.reschke@gmx.de>
To: "Eric Sedlar" <eric.sedlar@oracle.com>; "Webdav WG"
<w3c-dist-auth@w3c.org>
Sent: Wednesday, September 18, 2002 1:33 AM
Subject: RE: ETags, was: Issues from Interop/Interim WG Meeting


> Eric,
>
> the problem in requiring things is that you may get what you thought you
> want, but then discover it was wrong to ask for it.
>
> If you declare a server that doesn't have entity tags non-compliant, you
> might get the server implementor to reconsider his/her decision. In some
> cases, there may be a very good technical reason not to supply entity
tags -
> for instance, if your store is a file system, and you don't have any
> additional place for metadata. Computing a reliable entity tag just from
the
> file's properties (timestamp, length...) is impossible, but the
implementor
> might attempt it anyway. In the end you'll get a server that *does* supply
> entity tags, but they only work most of the time, so they aren't reliable.
> That's worse than not having them at all.
>
> So, there are valid reasons for RFC2616 to make these things optional
(same
> for server timestamps). I'm happy with language that explains that entity
> tags are an important feature, but it should also make clear that if a
> server can't provide reliable entity tags, it shouldn't attempt to.
>
> Julian
>
> --
> <green/>bytes GmbH -- http://www.greenbytes.de -- tel:+492512807760
>
> > -----Original Message-----
> > From: w3c-dist-auth-request@w3.org
> > [mailto:w3c-dist-auth-request@w3.org]On Behalf Of Eric Sedlar
> > Sent: Wednesday, September 18, 2002 9:10 AM
> > To: Webdav WG
> > Subject: Re: ETags, was: Issues from Interop/Interim WG Meeting
> >
> >
> >
> > RFC2518bis wouldn't invalidate a class of servers if it includes
> > a new token
> > in the DAV: header to indicate support for RFC2518bis.  Clients
> > would still
> > have to deal with no-Etag servers to support RFC2518, but this might
> > accellerate implementation of Etags.
> >
> > --Eric
> >
> > ----- Original Message -----
> > From: "Clemm, Geoff" <gclemm@rational.com>
> > To: "Webdav WG" <w3c-dist-auth@w3c.org>
> > Sent: Tuesday, September 17, 2002 7:57 PM
> > Subject: RE: ETags, was: Issues from Interop/Interim WG Meeting
> >
> >
> > >
> > > I have no objection to such a warning (in fact, it sounds
> > > like a good idea to me).  But I agree with Julian
> > > that RFC2518bis should not invalidate a whole class of
> > > valid 2518 servers, even for a worthy cause such as ETag support.
> > >
> > > Cheers,
> > > Geoff
> > >
> > > -----Original Message-----
> > > From: Eric Sedlar [mailto:eric.sedlar@oracle.com]
> > > Sent: Tuesday, September 17, 2002 8:47 PM
> > > To: Clemm, Geoff; Webdav WG
> > > Subject: Re: ETags, was: Issues from Interop/Interim WG Meeting
> > >
> > >
> > > As long as you don't mind a client saying something to the effect of:
> > >
> > > "This server does not support the minimal level of functionality that
> > > <product> requires of a WebDAV server (ETags).  We strongly
> > discourage you
> > > from using this server, as you may lose work."
> > >
> > > when it points at your server, then go ahead and don't support ETags.
> > >
> > > --Eric
> > >
> > > ----- Original Message -----
> > > From: "Clemm, Geoff" <gclemm@rational.com>
> > > To: "Webdav WG" <w3c-dist-auth@w3c.org>
> > > Sent: Tuesday, September 17, 2002 6:50 AM
> > > Subject: RE: ETags, was: Issues from Interop/Interim WG Meeting
> > >
> > >
> > > >
> > > > I agree.
> > > >
> > > > -----Original Message-----
> > > > From: Julian Reschke [mailto:julian.reschke@gmx.de]
> > > > Sent: Tuesday, September 17, 2002 4:58 AM
> > > > To: Lisa Dusseault; Webdav WG
> > > > Subject: ETags, was: Issues from Interop/Interim WG Meeting
> > > >
> > > >
> > > >
> > > > > From: w3c-dist-auth-request@w3.org
> > > > > [mailto:w3c-dist-auth-request@w3.org]On Behalf Of Lisa Dusseault
> > > > > Sent: Sunday, September 15, 2002 8:14 PM
> > > > > To: Webdav WG
> > > > > Subject: Issues from Interop/Interim WG Meeting
> > > > >
> > > > > ...
> > > > > -  Be clear in spec that servers MUST do ETags. Explain how
> > necessary
> > > > > this is to solve the lost update problem.
> > > > > ..
> > > >
> > > > ETags are a good thing, correct. However, HTTP (RFC2616)
> > doesn't require
> > > > them, RFC2518 doesn't require them, and they '*aren't* required for
> > > > interoperability. So there's no way to require them in
> > RFC2518bis -- it
> > > > would break all servers that don't have them.
> > > >
> > > > Julian
> > > >
> > > > --
> > > > <green/>bytes GmbH -- http://www.greenbytes.de -- tel:+492512807760
> > > >
> > > >
> > >
> > >
> >
>
>

Received on Wednesday, 18 September 2002 14:12:48 UTC