Saturday, December 11, 2004

Questions to be answered by the introspector

Todays entry will start by defining the questions to be answered and give you a structure to the problem of the introspector rdf data.

Here are the questions to be answered :
  1. How can you describe code so that is can be generated?
    1. How can you transform from the high level structures that you want to think in into the structure of the code that the compiler uses
    2. How can you transform from the low level structures that the compiler stores the code in into the structure of the code that you want to read?
  2. What is the relationship between the ontology and the structure of the code?
    1. What is the structure of the rdf data?
    2. How can a rdf data set be converted into an ontology?
    3. How can you create code out of a ontology?
    4. How can you create the ontology out of the rdf data?
    5. How does the structure of the gcc affect the structure of the rdf data?
    6. How does the structure of the input source code affect the structure of the rdf data?
      1. What if the input source code to the introspector is the dumper code?
    7. How does the structure of the dumper code affect the structure of the rdf data?
  3. What is the relationship between the ontology and a query?
    1. How can an ontology be converted into a set of queries that checks a set of rdf data for meeting that ontology?
    2. How can a query be optimized by consuting an ontology?
  4. What is the relationship between an index and an ontology?
    1. What are the types of queries needed?
    2. What are the cardinalities of the data sets?
    3. What are histograms of values in a property?
    4. What are primary keys of class?
  5. How can a rdf data set have statistics extracted from it that feeds the indexing, ontology, query optimization aspects?
    1. How can we store this statistical data?
    2. How can we query this statistical data?
    3. What is the relationship beween an ontology and the statistical data extracted from a data set?
    4. Does this have to do with with the bayesian classification
      1. see also http://ebiquity.umbc.edu/v2.1/project/html/id/59/