[19:39:40] --- michael has joined
[20:56:05] * michael can't stay awake for iiri.....
[21:33:57] --- ggm has joined
[21:34:27] <ggm> damn. I hope we get a scribe.
[21:35:32] --- tonyhansen has joined
[21:48:26] --- resnick has joined
[21:49:42] --- ludomp has joined
[21:52:41] --- Melinda has joined
[21:55:43] <ggm> do we have a scribe here? I'm in dnsop and wanted to hear the story..
[21:56:11] --- hartmans has joined
[21:56:42] <hartmans> Just now looking for scribe and just starting
[21:58:06] --- hardie has joined
[21:58:38] --- leslie has joined
[21:58:40] <tonyhansen> we're trying to find a jabber scribe -- looks like I'm it (sigh :-) )
[21:58:54] <hardie> Andy notes that Wang Liang and Guo Yiping were not able to come.
[21:58:59] <hardie> They sent PPT with sound
[21:59:00] <tonyhansen> sam hartman will also be taking notes
[21:59:01] <hartmans> I'm going to try and help scribe, but don't really know names
[21:59:13] <hartmans> Two primary presentrs could not make it here
[21:59:19] <hardie> Andy clarifies that is a BoF for a discussion, and not for decisions
[21:59:22] <hartmans> For discussion of technologies; no decisions that are coming out of this BOF
[21:59:36] <tonyhansen> (Wang and Guo had visa problems)
[22:00:05] <tonyhansen> overview of agenda -- jon klensin is up first
[22:01:35] --- michael has left: Disconnected
[22:02:03] <tonyhansen> history overview of search, discovery, retrieval
[22:02:14] --- SAH has joined
[22:02:33] <hartmans> Talking about where we've been in this arear previously.
[22:03:27] <hartmans> jon: We are not new to this area
[22:03:51] <hartmans> jon: A lot of work finding things on the Internet: X.500, LDAP, whois
[22:04:07] <tonyhansen> jon: FIND BOF, srvloc
[22:04:26] <hartmans> jon BOF called find which covered some specific areas although from query standpoint not from retrieval standpoint
[22:04:38] <hartmans> Also service location and name guessing on DNS
[22:05:00] <tonyhansen> various things have been tried "above DNS", but have gotten little traction in ietf
[22:05:26] <tonyhansen> many rats filling this rat hole
[22:05:26] --- resnick has left: Replaced by new connection
[22:05:26] --- resnick has joined
[22:06:05] <hartmans> Not much activity, even less progress; Jon is depressed.
[22:06:22] <hartmans> We look at queries downward rather than taxonomy issues and database/general support
[22:06:51] <tonyhansen> ietf does ui poorly
[22:06:56] <hartmans> We've not looked at retrieval models shared among query protocols.
[22:07:04] <tonyhansen> ui =. user interface
[22:07:31] <tonyhansen> summary: lots of questions, but not new issues
[22:07:38] <hartmans> Questions about expertese and focus; could we get traction to add value? No new fundamental issues here.
[22:07:55] <hartmans> questions? None.
[22:08:22] <leslie> Lisa Dusseault at the microphone
[22:08:47] <hartmans> LISA: Points to DASL (sp) a protocl that did SQL-like access to repositories
[22:09:08] <hartmans> ANdy: Could you find the dasl server?
[22:09:24] <hartmans> lisa: No, didn't cover service location
[22:09:31] <tonyhansen> dasl
[22:09:55] <hartmans> Kurt z: XML in LDAP new work
[22:10:03] <tonyhansen> kurt: other work going on wih xml-enabled directories
[22:10:13] <tonyhansen> (xed)
[22:10:35] <tonyhansen> on to a presentation: intro to problem space and dris
[22:11:37] <hartmans> And I won't be able to help scribing this at all I suspect. Presentation from authors of draft
[22:12:37] --- dcrocker has joined
[22:13:33] --- nsb has joined
[22:14:17] --- pgmillard has joined
[22:15:53] --- resnick has left: Replaced by new connection
[22:16:04] --- tonyhansen has left: Disconnected
[22:16:25] <hardie> note that this presumes a hierarchical organization, rather than one based on the structure of information; further this organization is country-based in the example.
[22:16:35] --- tonyhansen has joined
[22:17:25] --- hta has joined
[22:17:25] --- hta has left
[22:18:04] <hardie> now presentting "metadata harvest system"
[22:18:25] --- resnick has joined
[22:18:28] --- paf has joined
[22:18:30] <Melinda> There doesn't appear to be a mailing list archive at the URL given in the agenda - does anybody know where to
[22:18:33] <Melinda> find it?
[22:18:49] <hardie> now presenting "distributed search system", nb: still represented as a hierarchy, but with a search engine instead of a harvest-point at the midpoint.
[22:19:43] --- amarine has joined
[22:20:24] --- klensin-ietf has joined
[22:21:32] <hardie> This seems to be fundamentally contrary to the "web" nature of most of the resources they would want to describe--how would they connect data linked from their own system to Stanford's, for example? They're in two different collections according this presentation, even if there were a direct link on the web between the Stanford digital library project and their work.
[22:22:19] <tonyhansen> (lost my conn.) a version of the powerpoint slides appear to be available at http://www.lib.hust.edu.cn/dl-lib/English/IETF/document.htm
[22:22:33] <hardie> proposing a dris. hostname convention
[22:24:39] * paf not being in the room asks whether they have any reference to the work earlier done in the "Common Indexing Protocol", i.e. RFC 2651?
[22:25:17] <hartmans> paf - No previous mention of that.
[22:26:38] <tonyhansen> http://www.lib.hust.edu.cn/dl-lib/doc/Introducation%20to%20problem%20and%20%20DRIS%20.ppt is the powerpoint currently being looked at, now on page 31
[22:27:22] <paf> Yes, I started to download it, but it is *HUGE*...
[22:27:44] <hartmans> [not as scribe] So, has there been an explanation of why this problem is worth solving, or anything more than just an assertion of the problem?
[22:27:58] <tonyhansen> conclusion slide: DRIS will integrate all kinds of information resources on Internet. DRIS builds the information retrieval infrastructure of Internet, but not the final search engine. Basic idea of DRIS: Search should be the internal function of Internet, everyone should have his own search engine. Basic principle of DRIS: in appropriate scope (three levels domain), applying appropriate information retrieval system, DRIS will build an appropriate information management frame for the whole Internet. You will get what you really want from DRIS: fresh, precise, comprehensive information.
[22:28:33] <hartmans> back to comments
[22:29:34] <tonyhansen> larry masinter @mike: is this a fantasy? wishful thinking?
[22:30:30] <tonyhansen> jon k: they do have a prototype, but can it scale?
[22:30:32] <hartmans> Jon: The presenters have this up working and tested, he believes.
[22:30:42] <hartmans> Larry: But scaling to the Internet . . .
[22:30:47] --- alexeymelnikov has joined
[22:31:20] <hartmans> Larry: Problem with metadata: scalability of namesapace. What I think as an author might be different than what you think as a searcher/author.
[22:31:48] <hartmans> This happens in small controlled requirements. Trying to get them to use the same set of metadata is a deployment problem. Doing that over the Internet is hard.
[22:32:01] <tonyhansen> leslie daigle: refers to FIND bof
[22:32:27] <tonyhansen> has sympathy for this proposal, sees this as an updated version of that work
[22:32:51] <tonyhansen> but it didn't work 5-6 years ago
[22:33:03] <tonyhansen> not convinced it couldn't ever work
[22:33:13] <tonyhansen> but there are issues
[22:33:23] <tonyhansen> as raised by larry
[22:33:30] <tonyhansen> jon talking
[22:33:42] <tonyhansen> this work may scale better
[22:34:01] <tonyhansen> jon asked to elaborate on how so
[22:34:34] <amarine> april, not in room, asks "which jon"?
[22:34:43] <hartmans> [I'm very confused by what is going on here]
[22:34:44] <tonyhansen> jon klensin: this is looking at query interface rather than retrieval interfaces
[22:34:50] <amarine> thanks
[22:35:51] <tonyhansen> jon: by doing so, this work may scale better
[22:36:12] <tonyhansen> ted hardie @mike
[22:36:30] --- leslie has left: Replaced by new connection
[22:36:33] <hartmans> Ted: Scope? If there is a scalaing problem --can it scale to the Internet. Answer is no. But is that the right question?
[22:36:37] <tonyhansen> ted: if there's a scaling problem, but should we be asking about it?
[22:37:07] <hartmans> What about smaller scale search engines. For example output of Harvest used to create topic-specific searches at NASA.
[22:37:10] --- leslie has joined
[22:37:33] <tonyhansen> (that's what ted is asking)
[22:38:13] <tonyhansen> nasa created topic-specific search engines, plus meta-data connectios between them
[22:38:21] <Melinda> I'm not in the room and I'm missing much of what's going on (but *many* thanks to the scribes), but if the scope is narrowed to the query interface the scaling problem kind of goes away. However, if the scope is narrowed to the query interface then data representation becomes an issue and I'm pretty sure that's something the IETF is ill-equipped to tackle.
[22:39:12] <hartmans> This proposes to create a hierarchy independent of the data/information.
[22:39:56] <tonyhansen> what about consortiums using internet as collaborative network?
[22:40:23] <hartmans> Actually what Ted's saying seems to be that NASA failed explicitly because they couldn't get the interconnections worked out.
[22:40:50] <tonyhansen> leslie @mike
[22:41:17] <tonyhansen> leslie was refering earlier to "finding stuff bof", not "cip" work
[22:41:24] <hartmans> finding stuff bof different from FIND working group
[22:42:17] <tonyhansen> (Melinda: I mentioned your comment on the mike)
[22:42:24] <leslie> For the record (and the minutes) -- findstuf bof
[22:42:26] <leslie> http://www.ietf.org/proceedings/98aug/98aug-38.htm
[22:42:32] <tonyhansen> on to the next powerpoint
[22:42:45] <tonyhansen> http://www.lib.hust.edu.cn/dl-lib/doc/Introducation%20to%20problem%20and%20%20DRIS%20.ppt
[22:43:30] <tonyhansen> oops: http://www.lib.hust.edu.cn/dl-lib/doc/Internet%20information%20retrieval%20infrastructure%20and%20digital%20library.ppt
[22:47:20] <ggm> thx for ptrs to slides.
[22:47:59] --- hta has joined
[22:49:08] --- leslie has left: Disconnected
[22:49:49] --- jhutz has joined
[22:51:33] --- jhutz has left: Logged out
[22:51:33] --- jhutz has joined
[22:51:33] --- jhutz has left: Logged out
[22:52:21] --- leslie has joined
[22:55:08] --- hta has left: Disconnected
[22:55:35] --- pgmillard has left
[22:57:31] --- Melinda has left
[22:58:32] <ggm> do we have a scribe? we're werrrry quiett ..... we're hunting waaaaaabits
[22:58:46] <leslie> It's really hard to scribe these presos.
[22:59:10] <hartmans> ESpecially for these presentations you'll be better off reading the ppt slides
[22:59:13] <amarine> is "presos" the new bling bling?
[22:59:14] <tonyhansen> summary slide: For search engine company: DRIS will transform the basic commercial mode of current search engine,providing better service for its users but not advertisement. For common users : Get what they want from their own search engine.
[22:59:26] <hardie> No, presos is not the new bling bling.
[22:59:27] <tonyhansen> slides are done
[22:59:30] <ggm> if thats aussie, it would be presso's
[22:59:39] <ggm> is bling bling the new bling?
[22:59:45] <tonyhansen> ask for comments -- none
[22:59:51] <tonyhansen> on to presentation on draft
[23:00:21] --- jhutz has joined
[23:00:32] <tonyhansen> http://www.lib.hust.edu.cn/dl-lib/doc/DRIS%20protocol(korea)(1).ppt
[23:00:45] <hartmans> [I still don't understand the basic problem that unites all this. Admittedly It's been hard for me to follow the presentations]
[23:01:51] <hartmans> [For example, having read the draft, how does this really relate to the DL presentation?]
[23:03:24] <tonyhansen> (questions on the sound quality of ppt presentation)
[23:04:48] <hartmans> Skipping presentation
[23:05:32] <tonyhansen> leslie @mike
[23:05:42] <tonyhansen> sam: how does this all fit together
[23:05:52] <tonyhansen> leslie: 2 levels of comments
[23:06:10] <hartmans> Leslie: There are other technologios that have much more of a shakedown.
[23:06:37] <hartmans> IF we pursue this the protocol itself would have to be modified. No one has implemented this at scale if they put that many operators into the query languge.
[23:06:49] <hartmans> Other set of questions back to Sam's point. What is the purpose of this.
[23:07:05] <hartmans> For this room: what is the dirivng function for something that scales to the Internet.
[23:07:41] <tonyhansen> larry @mike
[23:08:00] <tonyhansen> larry: did anyone go to conferences on digital libraries a few years ago?
[23:08:11] <hartmans> Larry: There used to be a series of conferences about digital libraries. I thought I walked into a sesnion of a coference from one of these.
[23:12:00] <tonyhansen> james seng @Mike
[23:12:34] <tonyhansen> james: draft belongs somewhere else, not @ ietf
[23:13:27] <hartmans> Because it uses XML etc is not a good reason for it not to be here.
[23:13:42] <hartmans> The way that it proposes to do things etc. Sounds much like semantic web.
[23:14:08] --- amarine has left
[23:15:04] <hartmans> Chair: who in this room would be interested in working on this in the IETF. [no one]
[23:15:05] <tonyhansen> ? for floor: who in the room wants to work on this in the ietf? no one raised their hands
[23:15:41] <tonyhansen> ipv6 & iiri is out of scope for this bof, so no comment
[23:15:45] <tonyhansen> adjourning
[23:15:50] --- hartmans has left
[23:15:56] --- resnick has left
[23:16:08] --- tonyhansen has left
[23:16:38] --- hardie has left
[23:17:28] --- SAH has left
[23:17:31] --- klensin-ietf has left
[23:17:46] --- jhutz has left: Logged out
[23:17:47] --- jhutz has joined
[23:17:47] --- jhutz has left: Logged out
[23:18:18] --- ludomp has left
[23:23:25] --- ggm has left
[23:24:16] --- nsb has left: Disconnected.
[23:25:04] --- dcrocker has left: Disconnected
[23:25:18] --- leslie has left
[23:28:25] --- mccreary has joined
[23:29:51] --- paf has left
[23:36:22] --- alexeymelnikov has left: Disconnected