Based in Sydney, Australia, Paula le dieu can sometimes be found writing here but can more often be found at

etcon:2003:day4:david weinberger

what groups will well anyway...[joke]

learning from experience is the worst possible way to learn - clay shirky

why is that we refuse to learn from our experiences in groups? *health warning* this presentation

group = a set of folks that know one another and have an awareness that they are in a group

community = a group that cares about one another more than they have to

the central premise is that groups are really important the secondary premise is that the web is really bad at facilitating my participation and management and even understanding of my set of groups. decentralisation serves the purpose of building the connections necessary for forming groups but that same decentralisation means that the net is poor at managing those groups.

(reads law)

example: friendster [laughter cos screen isn't working so david is painting a picture in words...screen fades in and he is pointing at exactly the right place on screen]

friendster's limitations - is that you only have a binary relationship - either friend or not room for associate, new peer, etc - the interests page is really a marketing are marketing yourself and therefore being highly selective about your choices that you list. - also it assumes that there is only one "me" and does not cater for the multiple modes of my existance.

the current environments do not encapsulate the wildly complex nature of self description and relationships.

david wary of environments that seek to service all types of relationship building and management.

explicit and implicit - most of the tools requires that we make things explict...he posits that in doing this you lose the context required and the ambiguity of implicit (ambiguity adds richness in social contexts - i think that was the position).

what comes first - the group or the constitution that governs the group?...david argues that you cannot establish a constitution initially as the relationships ie. the group has to drive and direct the constitution. also building constitutions requires explicit statements (i think he is using explicitness in the context of actually stating things) which we don't like as it is an act of violence ?? - i think he means that we are uncomfortable defining instrinsics...implied.

knowledge management - the big problem with km as social software is that it ignored the value of ambiguity - really trying to understand what exactly he means by this statement...??

km devised to help try and find the diamond information in the rough..problem with information is that it takes on diamond value to someone while having coal properties to someone else. it ignores the social value in context. you therefore end up with systems that are self delusionary.

it was also devised to capture individual knowledge bases ie. a person's knowledge. it also failed cos most knowledge bases ie. people don;t have the skills necessary to make that knowledge explicit.

if ambuigity (ahh...i think he means that when something remains in its context it remains ambiguous).

good social software allows structure to be developed while understanding that this is inherently an unnatural/uncomfortable experience for groups.

the same point being made that real needs/real context drive groups

1. there will be tons of services for group use 2. the coming age of dID maybe we will think less about managing groups and more about managing our various selves in the public of the web. 3. we are at the beginning of this whole social software - why is it that we are still only at the beginning...why a buzz now - maybe, the reason that we are now buzzed about ss is that we are coming out of our infatuation with binary thinking - maybe we are moving/ready to embrace the ambiguity inherent in "webs" and social networks.

questions/comments: metadata is crap - context is hugely important - pn. this is really interesting but surely context is just another metadata variable...these folks are having fab conversations about metadata....implicit and explicit - david argues that explicit metadata is crap..reputation models floating in my head...not sure why.

there is a buzz about the netscan presentation yesterday

we need to liberate groups/individuals from db...

finishes by saying that the future of ubiquitous group transactions is the use of something like threadsML!

etcon:2003:day4:geoff cohen