FOAF additions

Feb 28, 2004 10:59

Alright, I'm putting together updates to the FOAF information for the next version. I can't make any promises as to when it will go in, live, or even be passed onto a developer - this is a very rough draft of changes that I'm working on making.

Let me know if there's anything I missed.

Additions:

vocabularies, PPD, nodeID, location, userpic, keywords )

Leave a comment

mcfnord March 1 2004, 09:23:27 UTC
I'm confused about foaf. Is it operational?

Take a look at:

http://www.livejournal.com/userinfo.bml?user=fool_in_spirit

compare with

www.livejournal.com/users/fool_in_spirit/data/foaf

Where is _labyrinth? aardvarki? Should they be listed in foaf? They seem to be normal users.

Reply

crschmidt March 1 2004, 10:26:41 UTC
fool_in_spirit has probably created a FOAF-knows friends group to only list the users that s/he actually considers him/herself to have a FOAF knows relationship with. You're right in saying that all the journals look like normal journals, but there are a bunch of personal journals in that user's friends list that aren't in the FOAF file.

Anyone who wants to can easily create a list of who they know, simply by following the instructions in a post in this community.

Reply

mcfnord March 1 2004, 10:31:11 UTC
ok, i do remember some discussion. i'm going to bank on the belief that 99% of people will never do this. thanks for your reply.

Reply

crschmidt March 1 2004, 10:41:21 UTC
I'm not quite sure why you're banking on people not doing exactly what the Spec is designed for.

FOAF isn't designed to include a friends list: It's designed to include a list of people that you think you know. Since we don't have that kind of information, we use the Filter option instead. I think it's a good way of doing things.

What information do you expect to get out of this format? If you want a friends list, you're going to have to wait until we actually offer it, I think - the FOAF information is not designed to fill that role.

Reply

mcfnord March 1 2004, 10:53:53 UTC
Forgive me if I have offended. You've built a flexable architecture. I used filters and they work great. This list maps well to that existing architecture. But would you agree that 99% of people will not modify this list? At least in 2004, which is probably the timeframe of my work.

I am looking for who-i-read information. I don't like scraping it. I am testing v2 of the gadget at ljmindmap.com. I ran it scraping for a while and wondered if it got me auto-banned for like 30 minutes because livejournal.com died to my I.P. alone for that period. I emailed Brad and Evan about getting the dataset on cd-rom, but that interface also failed, so to speak. So I'm wishing and hoping an answer appears. Basically your work is an answer. If people change their foaf list, then their output is not as pretty. This and loss of self-described Location data are the only drawbacks to the best data aquisition strategy I've found. Welcome your thoughts.

Reply

crschmidt March 1 2004, 11:05:27 UTC
Sorry, I was a bit harsh in my last response - I didn't mean to be.

The network actually works better in the case that you're looking for "Who I read" information than if you're not. Although communities are obviously excluded, people who people list explicitly in their FOAF-knows filter are actually the people that map more closely to the person in question. For example, I have a number of people on my friends list that are only there because they wanted to read my friends only entries. I exclude them from my FOAF-knows filter - I don't read their entries, and I don't "know" them.

So, I think the information that you are getting, if people have created a filter like this, is truly more accurate than other information you might get otherwise.

I could be wrong, but it seems to me to be a pretty darn decent mechanism for determining who I really am related to in any kind of significant way, ya know?

Reply

mcfnord March 1 2004, 16:17:24 UTC
I hear you. To some degree, my project is just entertainment, pretty pictures, and it's valuable to have a larger dataset than a smaller one. Connections imply something more than a lack of connection can, even if the precise meaning is unclear. Foaf is as close as I have come to this data in a clean way. I'm going to stick with it for now. I would simply adore location information, as I use it for hovertips. As in:

http://ljmindmap.com/brad.htm

I tried yours but the complexity killed it! Still much to do.

Thanks

John

Reply


Leave a comment

Up