Notes on planet

1. Aanalysis

1.1. I concentrated on planetrdf, but I expect that all planets look
     similar.

1.2. Technically, planet is a 'blog of blogs', showing the most recent submissions
     from members plus links to their archives (individual blogs)

1.3. There is a membership and only members can post. However, all can
     read the posting. I could not find any commenting/responding capabilities.
     I suupose that the memer should demonstarte certain level of professional 
     engagement.

1.4. The planet is organised around a single subject, usually research or technical
     challenge. It is assumed that contributions should stick to the area.

1.5. There is a certain amount of social posting, personal life sneaks into postings.

2. Purpose of the trust model

2.1. It is desirable to rate contributors regarding their professional
     prowess so that valuable contributors gets recognition and priority

2.2. The quality of submissions can only be judeged by other professionals.
     There is no solid method to automate the process. Actually, other
     members are in the best position to assess contributions.

2.3. Being human, members are prone to mistakes, pride, willingness to
     retribute (and fear of retribution), forgiveness etc.

3. Now, this is the idealistic vision 'how it should work' not how it actually
   works - kind of the proposal of trust model.

3.1. It is the community, not the producer-consumer model. Everybody is supposed
     to contribute and to benefit from it. Visitors can have a look but should
     not interfere with trust model.

3.2. We have three contexts here, each context with potentially different notion
     of good behaviour (trust)
	(a) professional contribution
		factual contribution, appropriate style, novelty, relevance
	(b) social contribution
		t.b.d.
	(c) reviewer's contribution
		timely, matter of fact, non-personal
     I suggest that we stick with (a) and (c) only, for simplicity

3.3. All submissions are identifiable - to login names or similar. There is
     a value in keeping those login names persistent. Registration process
     is in place - potentially identities are verifiable.
 
3.4. All members share the same understandng of what constitutes the
     desired contribution (at least in the (a) and (c) contexts)

3.5. Group of users:
	visitors - read-only, potentially with no trust ratings
	members  - can see the trust rating, can send posting
	contributor - member that sends postings, can send reviews
	reviewer - contributor that sends reviews, can send comments

3.6. The process of trust building:
 	1. contributors submit postings, identifiable 
	2. postings are available to all members/visitors
	3. postings are sorted/qualified by trust level (for members?)
	4. trust is associated with the entity/person
	5. only contributors can post the review - rating of 'good posting'
	6. review cannot be rebutted by the original poster (no flame wars)
	7. review can be commented (qualified) by anyone who reviewed something
	   except for the author of the actual posting 
	8. the review is also identifiable - preventing from defamation
	9. trust in the poster is the function of currently available set of reviews
	10. trust in the reviewer is the function of currently available set of comments
	11. the perception of trust is shared by all members
	12. the impact of the review depends on the trust in reviewer
	13. trust should be ideally re-evaluated every time (I guess it is not possible)
	- there is a new posting
	- there is a new review
	- there is a new comment on the reviewer
	14. all impacts diminish in time