2009Aug20

From Filtered Push Wiki
Jump to: navigation, search

Agenda


User: Bob Morris | User: Zhimin Wang | User: Maureen Kelly | User: Paul J. Morris | User: James Macklin | User:
"User:" cannot be used as a page name in this wiki.
| User:
"User:" cannot be used as a page name in this wiki.
| User:
"User:" cannot be used as a page name in this wiki.
| User:
"User:" cannot be used as a page name in this wiki.
| User:
"User:" cannot be used as a page name in this wiki.
| User:
"User:" cannot be used as a page name in this wiki.



  • Donna progress: Sorry I can't make it today. Please see 2009Aug13 for an example of Template: Event. I discussed with Bob that what we really want is a new "skin" for Filtered Push, not templates yet. Once we have the new overall css design, then we can start using templates in areas where it makes sense, e.g., meetings, events. We are going to talk next week about how I can access and edit the php file(s) in the /skins folder of the Mediawiki directory, and then we will need to edit the Filtered Push pages to point to the new php page. No point using my test page example yet, since by using skins I can make it look even better!
  • Attribute dependencies in attributes of a (collection) record, which are important for determining duplicates?
  • AGU American Geophysical Union December 14 SFO; deadline Sept 3
Paul reports there are interesting possibilities.
e.g. http://www.paleoportal.org/ data services in DiGIR. Paul will draft an abstract.


  • Donna's current sample:
On 2009Aug20, a meeting will be held at HUH . We will discuss updates and progress of the filtered push project.


  • New FP work
    • Fuzzy match --Zhimin has primitive FM working; No present indexing. Can slow down as data increases
Z found some old C++ algorithms for M-Tree indexing. ("Metric Space Indexing"). Maureen asks what is different from standard GIS indexing ("R-Trees", traditionally). After many dimwitted questions from Bob with well crafted answers from the rest, he suggested that maybe the real benefit of M-Trees is that they are meaningful when the metric space has no natural arithmetic(?), or at least metric on a base space which is providing Euclidean metric on the feature space.


    • IPT
Zhimin has schema for annotation.
    • Tasks for David Lowery
      • Make a demo deployment with several nodes; maybe later on 3VMs on a single laptop for realism.
David and Maureen to meet in the first week of September before meeting of Thursday Sep 3 to discuss what the task is.
    • Bob to work on sensor data reasoning as an FPCQC analysis engine? SMW?

Consensus: SMW more important first

No Progress
    • What is the necessary to represent sensor data (a few weeks from now?)

Semantic global and local KR?

No Progress
Oops. Bob forgot to do.

General scenario:

user has a record for matching. System first makes fuzzy match on Cname Cnum only. Caches returns locally if not already there. Also, system offers list of attributes in entropy order for user to ask for match list based on attributes to be chosen by user.

Other accomplishments this week:

James and Bob did some budgeting. Bob and Rich Ragin agreed on the correct bills Harvard should pay. Bob will prepare UMB extension budget.

2009Aug27

Agenda

  • Admin
    • annual report
    • UMB extension budget
  • Technical
    • SMW; finish UMB extension budget
    • IPT
    • AGU
    • MTree report from Zhimin