LSD-2 Chair: Harald Alvestrand Chair: Roland Hedberg Minutes, Patrik Faltstrom Agenda was bashed by Harald, which also presented the scope of the discussuions. The goal is to have a recommendation to the IETF at 11.30 about what to do, if the IETF is going to do anything. Patrik Faltstrom and Roland Hedberg presented the TISDAG project in Sweden, which is running a proxy/referral service since about a year back. Harald Alvestrand presented the Katalogforum which is an initiative in Norway for setting up (in the first step) a directory of directories. Peter Gietz presented the Desire Indexing project which have a subproject which works with white pages information. During these three presentations, questions were asked about (among other things): - Tagged Index Objects + The ability to reconstruct records + Privacy issues + aggregation of tagged index objects; pro and cons - Size of the index - Crawling (fetching information from servers) + performance + load impact on servers + versus having servers provide a tagged index object themself It was said explicitely that we should/could not discuss business models. After discussing the issue a bit, we agreed that we probably have to walk (dangerously) close to the edge between technology architecture and business models. As long as we only talked about technology which enables design of different buissness models, we should be on the safe side. We discussed what a group like this could discuss, and came up with the following list: - The protocol may be less of a point than the service model; stop talking about protocols? - Real Life means that protocol variations and differences must be handled. - Do we need to document services? - Internationalization and localization? - Integration of White Pages with network (QoS) data? Not particularly popular....."Walk before running" - Do we need to assume a small number of "mediator" organizations? Or is this "open market"? - What is the relation to DNS (server location) - What is the cost of participating in multiple schemes? - Is it too early for this discussion (Too late?) - Is the IETF the right forum? (EMA Directory Track?) - Is there a market for a global interconnectivity? - What is the overall service architecture(s)? - "Price and politics are among the core issues" - Service models Summary - Description of service models for wide-scale directory + What need a service fills + How the service fills the need - Schema recommendations + Properties need to be part of entry + I18N, multiple-CN names needs trackling - Client properties needed to take advantage of service + Multi-field searches + Referral handling - Testing with referrals in Real Life (tm) needed - Appropriate area is O&M + No new standards + Profiling and usage Possible places to work - IRTF + Long term, "new field" issues - Existing procol WGs + LDAPEXT, Web groups... - New/restarted protocol WGs + FIND/Schema - EMA + Openness questioned