Thursday, January 06, 2005

Sxip, Identity Commons, and digital identity If you've been struggling and juggling with dozens of IDs and passwords, you know user authentication is one of the major "pain points" online in this era of multiplicating lightweight services. Sxip is one of the promising avenues toward solving this problem with minimal hairpulling on either the user or developer end. Martin Terre Blanche explains the user side better than I could in his post "Single sign-on at last". Here are the first two paragraphs. I urge you to read the last few, too. "Stephen Downes today mentions "something big", which is exactly what I have been waiting for for quite a while now - a viable single sign-on system for internet sites and services. Sxip seems to be it - "a simple, secure and open platform for true digital identity. Sites that implement Sxip support are able to easily provide features like single sign-on and automatic form fill". This might sound like nothing special, but for me it feels like the final missing piece of the puzzle of online collaboration (and collaborative learning). The last few years have seen the emergence of wonderfully simple, intuitive and powerful collaboration standards (such as RSS) and tools (such as blogs, feed aggregators, collaborative bookmarking services, etc. etc), and collectively they provide a much richer and more flexible environment than any single system (such as one of those massively cumbersome learning management systems universities love to waste money on) ever could. The problem is that each service necessarily requires first laboriously signing up and then signing in again on a daily basis (if you can remember your id and password, that is). Now imagine what a pleasure it would be if services like blogger, bloglines, furl, feedburner, flickr and even commercial outfits like amazon could agree on a single identity authentication system." Marc Canter writes about the developer side of Sxip, which also appears to be relatively free of headaches: "Dave Winer pointed out that the average software developer is not in a position to support one of these complex identity systems out there, like Liberty Alliance or the WS-* stack. Dave called it the 6th Law of Identity - KISS. So BAM! It is just so obvious to me why Dick Hardt built Sxip they way that they did! It's exactly what Dave asked for! Sxip can easily be supported (with as little as some Javascript) and provides as thin of a layer as possible, while providing a DNS-like service for digital identity. You can dive into it deeper - building your own Sxip-compatible storage portals or just create a site that reads compatible Sxip profiles and meta-data. But the only money that's paid to Sxip is for maintaining a HomeSite or MemberSite - and it's roughly the same price as maintaining a digital certificate or domain name registered - and end-users pay nothing." Identity Commons is another digital identity proposal that looks sensible, and I just learned from this page that there are plans to interoperate with Sxip. While it's too early to tell who's going to crack this nut, here's to hoping 2005 will be the year of single sign-on. If you care to go in more depth on this topic, see Jon Udell's The semantic web, digital identity, and Internet governance, which has this gem quote from knowledge representation specialist John Sowa's essay The Law of Standards : "Whenever a major organization develops a new system as an official standard for X, the primary result is the widespread adoption of some simpler system as a de facto standard for X." Another quote from that piece, this time from master Udell himself, echoes a suspicion I've had as well: "when others assert facts about you -- as they increasingly will -- the tide could begin to turn. Individual acts of self-defense may ultimately combine to bootstrap the semantic Web." What do you think? []  links to this post    4:31:52 PM   Multidimensional categorization of music danah boyd posts about her disjointed playlisting habits, and the impact such habits have on the output of last.FM's collaborative filtering algorithm. One of the commenters writes, "Maybe it is just that the software is revealing the shockingly deep and secret connection between Johnny Cash and PsyTrance?" The connections most often drawn between musical pieces are based on stylistic resemblance, and standard taxonomies reflect that. But style is a surface characteristic of music; I often find there is a common spirit to songs with wildly diverging styles, that may not be as readily identified but is nevertheless there. I know people have commented that there's a kind of unity to the picks in my Webjay playlist, even though it's all over the map as far as genres go. The All Music Guide categorizes bands along an axis called "mood" (formerly "tone"). Examples include "rambunctious", "quirky", or "trippy"... I've found moods to be a terrific alternative for finding good patches of music. So yes, mood tags could indeed be useful on music community sites. Indeed, indiscover.net encourages categorization by mood, and offers mood-specific playlists to registered users (based on an algorithm developed by Daniel Lemire and others at my institute at the NRC). Generally speaking, as collaborative filtering becomes mature and widespread, I expect some of it will profitably evolve beyond one-dimensional overlap proximity measures, though this may require much more data and processing effort. What do you think? []  links to this post    12:59:12 PM

Copyleft 2006 Sebastien Paquet. Last update: 4/22/2006; 12:19:38 PM. This theme is based on the SoundWaves (blue) Manila theme.

January 2005
Sun	Mon	Tue	Wed	Thu	Fri	Sat
						1
2	3	4	5	6	7	8
9	10	11	12	13	14	15
16	17	18	19	20	21	22
23	24	25	26	27	28	29
30	31
Dec   Feb

Syndicated content: