Thesis/Todo

Difference (from prior major revision) (no other diffs)

Changed: 1,47c1

- todo list extracted from outline
(each of these should generate a few paragraphs)

background

* find papers/chapters on advantages of p2p networks - prove load sharing, robustness (look in p2p book)

* more reading on semantic web - purpose, origins, evolution, current status - read through xml, xsl, xslt, xml schema docs/primers - what problems are these designed to solve

state of the art

* read up on openarchives.org - how does it work, how is it the same, how is it different - other meta-data search engines - are they all centralized? have they taken steps towards dist?

* read through edutella stuff again - are schema's limited? would edutella allow extensible apps? does it solve community discovery problem?

* jxtasearch - what level of structure is there in messages. can schema be shoehorned in there? how are searches specified/matches returned? can it be used to find xml objects?

* read up on chord/distributed file system stuff.

* read up on jabber - structured messaging - anything else?

* read up on bitzi - what are they doing there - trust, content-hashing - are schemas flexible? centralized.

* iptps papers - any relevant?

research question

* read about existing structured doc initiatives - cml, genetic, biodiversity - how are these being used right now to share information - case studies have they been used with p2p

* what p2p has been used in scientific communities? (smallworld paper from iptps?)

* new chapter? - "well-known problem of sharing and discovering design patterns" - find refs - existing approaches, analyze weaknesses, etc.

implementation

* proof of concept - run on top of jxta - independent of network layer?

* test running on jetty - create simple installation.

* indexing of metadata (only applies to centralized db?)

* local "metabase" - what is format for this repository - what is query interface - investigate xml query languages: xml query, xpath? try to get a copy of felber's xml filtering paper

* read up on UML. Generate some formalism for U-P2P. investigate tools.