Re: It's all about the DAV, baby! [Tim's WWW8 Keynote]

Tim Berners-Lee (
Wed, 12 May 1999 18:59:56 -0400

-----Original Message-----
From: Rohit Khare <>
To: <>
Cc: <>; <>; <>
Date: Wednesday, May 12, 1999 2:09 PM
Subject: It's all about the DAV, baby! [Tim's WWW8 Keynote]

>So I'm sitting in Tim's keynote -- which I'm sure is posted under
>, because it's the same talk from the LCS 35th
>Anniversary last month, right down to the title (!) -- and he brings

:-/ I always clone a talk and 50% of the time I leave the title... sigh
it is in a separate file....

>up the Jim[Whitehead]Cam dream again: a digital camera with a
>wireless Internet connection and the ability to immediately "marshal"
>a family album back to the Web -- authoring by devices, in other words

It isn't the difficulty of transmitting the pics - it is teh fact that the
virtual coffee table on which we marsh^H^H^H^Hassemble a photo
album. I don't have a picture album object with drag and drop even
though I have the Kodak free CD and eth Olumpus free CD etc....
Just makng a web page up is difficult enough. Everyone has their
own scripts. A lot of progs off free CDs change

> (I love that engineer! "marshal a family album", indeed, like an RPC...)
>The slides are at:

><DISCLAIMER STYLE="Friendly advice in an informal vein">
>"needs authentication, PUT, catch lost update, version management, etc..."
>We were waiting for commercial WP
>We said forget it, we started using Amaya B+E and we found that while
>HTTP/1.1 had a paragraph on lost update, it took a lot of times to
>adapt Jigsaw to do this. And we now use it. And so I don't squash
>people's updates...
>"Where are the papers [on authoring and collaboration and people
>working together] at this conference? are they going somewhere else?
>[Yes: SIGCHI, CSCW/ECSCW, Hypertext, Workflow].

Oh good.

> You know, we had a
>collaboration workshop a while ago, maybe we should have another [of
>the 1995 confab Karen MacArthur managed??] Why aren't people sending
>me cool little shareware programs doing this kind of collaboration?
>[How about Office 2000?]"

As cool shareware programs go, ...

>Tim, we need to have a chat. An Internet Proposed Standard *does*
>exist for this application, and it deserves better than handwaving
>about "oh, but we can do the critical part with just 1.1 and Jigsaw."
>And it does deserve better than ignoring it entirely in public.

Wrong. Dav does not do lost update protection - Dan just checked. HTTP 1.1

>I heard Henrik posted a little idea of how to avoid lost-update to
>the DAV-list recently -- without using DAV. Fair enough.

There is a paper due on "how actually to really use etags" as it wasn't
straightforward when combined with format negotiation.

>But silence is a statement itself -- raising the scepter of "wouldn't
>it be great if we could have collaborative authoring and make PUT
>work" is a little out-of-date

It was supposed to be a "that was one thing it took - and you would have
thought it would have been done but in fact it took some work."

>+ usual blast against sites which are all-POST rather than URI space
>with GETs. I presume he was talking about the
>warning that if you use a web accelerator/cache the site won't work.

The webmaster got back to me that another 5 sites do it. And Bankboston is
etrrible about not giving you a URI for bank account. Use the back button
and follow a link and they server respons ":you weren't supposed to be
htere - we
thought you were somewhere else so you can't go here. go back and try again
don't double click or use the back button)." Or words to that efect...

>+ recapping the 1994 Geneva conference definition of semantics. A
>dictionary is a reflexive document [-- there is no bottom turtle, as
>I'd say].

Exactly. Witgenstein? Gimme a definiive reference.

>+ "There were lots of hypertext systems before the web, but they all
>talked to one disk".
>+ In recapping 'evolvability', he asked for a show of hands of
>#Brisbane alums, and it wasn't close to half. (@@anyone have better
>+ the airplane part example - accounting only needs to understand the
><PRICE> tags -- this is really getting to be an ur-Archetype in every
>XML introductory talk. I wonder what the root of the meme is -- not
>just out of sheer egotism, to see if my ranting about purchase order
>sank in :-) -- but to understand 'convergent evolution' - why this
>one example resonates so broadly with the essential concept of XML.
>Or is it just a sign of capitalist globalisation -- SovietML would
>have been for identity cards with interoperable KGB and GRU tagsets?
>+ TimBL enumerated: distribution, endorsement, configuration,
>personal, privacy, library, payment methods, financial, weather, and
>about 13 more possible RDF schemas
>+ TimBL mocked the interactions of If-Match and If-Modified-Since,
>pleading for a simple data model (but not, presumably back in 1995 to
>defend Roy's condition fields :-)

I actually had a student recoding HTTP in sGML

>+ " you can't write a check for the "smallest amount which is the sum
>of two perfect cubes" [-- but why not chain DTDs and use MathML in
>eChecks :-) ]
>+ it's disappointing to me that only half-an-hour into it that he
>intended to use the equivalence of 'fire-engine red' and 'Mars red'
>as his semantic equivalence problem to keep referring back to.
>+ TimBL presumed that everyone knows every acronym, which speaks
>again to the two-caste WWWn attendance structure: half will still
>walk away scratching their heads saying "he's a Nelsonian rocket
>scientist" -- RDF might as well be 'Zipper Lists' for the magical
>power he ascribes them...

Agreed. RDF indoduced befor Semantic web.

>+ Complaints about Amaya's speed -- but it genuinely was in
>simultaneously in edit mode. Amaya2 has XHTML in too -- gotta check
>it out, but still, it's not an Open Source Triumph (TM) yet either --
>no public big-fix submissions

The screen rewrite speed was a shock - my thinkpad is in a weird mode.
Alan thought it was stuck in power-saving mode. I wonderd whether use of the
port had freaked the video drivers. It wasn't like that this am!

>+ Blurb for tutorial T2; "RDF was once independent of XML, but now
>they're definitely converging" [intertwingled]. Applications: DC
>(Dublin Core), RPMfind (linux module dependence), ACL, ... "many
>engines exist" [Once again, only acronyms on slides; expansions only
>in voice]
>+ "If you're cool, BTW, you call 'em schemata" :-)
>+ "we now have a joint IETF/W3C xml-sig group for signing XML and
>RDF" -- not so fast, Tim, not just yet. Schiller's mixed IETF/MIT
>loyalty aside, I think the effort still fractures along separate
>lines: 1) representing dsigs in XML format (replace PKCS-7, the same
>tilting at windmills we've hoped for a long time -- but XML hype may
>finally defeat ASN.1) and 2) canoicalize XML (for stable hash values
>across "equivalent" XML instances, regardless of whitespace, charset,

>+ Personal opinion on Patents (not as LCS or W3C): "not serving this
>industry as they were intended". The bar for "innovative" is far too
>low. Ethos of "whatever you can get away with". Too much trade in
>rumors, failure to work in Web time. "Dear community, please solve
>this problem" [Well, Tim, hint: the business people behind it aren't
>in Toronto. Perhaps some face time at PC Forum, Vortex, or Venture
>Market East might make a real impact.]

Maybe you could mention it when you are there?

>+ "We have changed the world, we can change it again, hey, we change
>it every six months!"
>+ "There are lots of interesting mechanisms for mechanizing social
>process" -- hint, communities and voting and recommendations are hot
>Internet stocks, way past 'research'. In general, it's an interesting
>question: why are so few of the Dow Jones Internet Index companies
>members of W3C? Why should/shouldn't W3C be on the hockey-stick
>growth path of the Internet industry?
>+ I asked the
>"there's lots of reasons to not do something, and if someone else's
>doing it, that's the best :-)"
>"management concerns... review comments on WebDav weren't passed back
>and forth in timely fashion; it wasn't a perfect liason, but it's
>fine in the end"
>"going forward is there a document planning what we/W3C needs to take
>action on to prevent incompatibility?" [implication that there's no
>worry on that front]
>Tim: "What does W3C need to do?"
>Rohit: "simply mention the existing Internet standards when you
>describe collaboration scenarios like this one; it's about promotion
>Tim: "Please listen to Rohit's endorsement of WebDAV" :-)
>[We make such a pair of chitchatters: We were equally roundabout and
>fast-talking and took three rounds of back and forth to get to our
>points :-)] [oh, and purely comic PS: since I was sitting in front of
>the speakers, they couldn't use the mike. Good thing I have no qualms
>loudly addressing a plenary hall in my speaking voice :-]
>+ integrate GPS into digicams: "search for all photos ever taken of
>this point; synthesize a VR map and show me". We should chat about
>some of the innovations of geospatiotemporal hypertext: the emergence
>of "places with phone numbers" (like the Piazza de san Marco),
>location-dependent name resolution (CNRP, anyone?).
>+ "Q: does [Internet2 bandwidth] herald another Web-like innovation?
>A: No. ... just faster and bigger." [He incorrectly conflated IPv6
>transition with Internet2 ]-- raising the FUD of "a totally new
>Interent 2, we'd have to at worst invent a new prefix, http2: but the
>web would still work.. you can still make links, you even send
>packets" -- "changes in Internet structure we leave to other
>people... something I'm not worried about"
>+ Web3D --what's up with this new consortium? I hear they're courting
>W3C in advance of standards a la WAP -- but what are they *doing* --
>and speaking of which, there isn't a shred of VRML visible at the

So which conference have they gone to?

>+ Mike Spreitzer asked a closing question on the diversity of data
>models: RDF isn't enough/ A: RDF is enough; it's purely low-level.
>Tim likened it to OO typing and the debate over inheritance and
>multiple inheritance, but RDF has a generic constraint mechanism. If
>you find the document hard to read, it's because the experts have
>gone down very deep and tried to describe every case, a big effort."
>[No, Tim, it's a flashing red siren warning of precisely what Mike
>raised -- the world won't "map everything to RDF" if they don't
>understand it! Ora and I and Ralph actually commiserated about the
>difficulty of our very different RDF tutorials -- not a great sign:
>that there's no educational silver bullet there]

If there is an educational silver bullet, it is listening instead of
talking. Has to be done 1:1.
I hope we have time to do a bunch of it. Who are the least understanding
and what is their
real data model?

>PS. the intersession music is far better than past WWWn's muzak:
>Brubeck and other fine jazz so far :-)