Changes between Version 1 and Version 2 of DescriptionDiscovery

Show
Ignore:
Timestamp:
2008/02/12 10:09:42 (11 years ago)
Author:
markw
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • DescriptionDiscovery

    v1 v2  
     1= WEB SERVICES = 
     2 
    13== What metadata is required to adequately describe a service? == 
     4 
     5 * service version 
     6 * software title 
     7    * nature of algorithm (likely = myGrid Task ontology) 
     8 * software version 
     9 * bandwidth and/or # requests per minute 
     10 * example input 
     11 * example output and/or REGEXP to test output 
     12 * some description of error-handling capacity 
     13 * sync/async 
     14 * nature of underlying data 
     15   * organism 
     16   * biological nature of data 
     17     * DNA/RNA/Protein 
     18     * SAGE 
     19     * Microarray 
     20        * platform, etc. 
     21 * input parameters and purpose of each 
     22 * output parameters and purpose of each 
     23 * usage/license restrictions 
     24 * authentication (is needed or not?) 
     25 * usage statistics (as per service provider) 
     26 * usage statistics (as per third party commentary) 
     27 * protocol (moby, soap, GET, POST, etc.) 
     28 * mirrors? 
    229 
    330== What ontologies exist that could provide this metadata? == 
    431 
     32 * myGrid - provides many of the annotation information  
     33 * myGrid - provides an ontology for bioinformatics tasks 
     34 * Moby Object - an ontology of data-types 
     35 * Moby Service - similar to myGrid 
     36 
    537== Do we need additional ontologies?  Who should make them? == 
     38 
     39 * TBD 
    640 
    741== What would a service discovery request look like? == 
    842 
     43 * TBD 
     44 
     45 
     46[[BR]] 
     47[[BR]] 
     48[[BR]] 
     49[[BR]] 
     50 
     51= RECORD IDENTITY = 
     52 
     53== W3C HCLSWG has some recommendations == 
     54 
     55 * these are largely URL-based + various HTTP protocol redirects etc for discovery of metadata 
     56 
     57== Life Sciences Identifiers == 
     58 
     59 * LSIDs have formal spec for retrieval of both data and metadata 
     60   * LSIDs do not support "REST" access 
     61 
     62== DOI or other redirect == 
     63 
     64 * ??? 
     65 
     66== [http://bio2rdf.org Bio2RDF] project & Banff Manifesto == 
     67 
     68 * proposed URI format:  bm:namespace:id  
     69   * for example bm:pdb:122345 
     70 * Freebase database of data providers and their namespaces 
     71   * Freebase is editable by end-users 
     72   * could be automated v.v. lookup 
     73 * Somehow map bm:pdb:122345  to its equivalent URL 
     74 * can we convince NAR to do this, since it has a list of all databases?