Due to the nature of the project, consistency, urgency, accuracy and efficiency are top priorities to the whole project, therefore, essential improvement on data format, query efficiency –semantics, redundancy control, and efficiency of dictionary as well as the query language nature/ guidelines  are necessary to ensure the data dictionary’s proper functionality in the overall EA structure.

1. Format: Since data dictionary is a semantic as well as a definition of data elements, it can evolve into ontology. As it was mentioned in the beginning that it is a web-based solution, it might be a good idea to use universally compatible standards such as XML as web service schema for data reuse ease and domain integration conveniences.

2. Query Efficiency: Semantics well defined. As mentioned above, the accuracy of the data in this particular project is essential to the timeliness of decision making and implementation processes. Semantic integrity is the way to prevent errors in data during data run-time.

   A semantic integrity constraint specifies what states and state transitions are allowed in the database to reflect the real situation.  Three general constraints are as follows:

   1. Row constraints

   2. Table constraints

   3. Inter-table constraints - Consider rows from at least two tables, such as foreign key constraint.

      In addition to the general constraints, when evaluating from a DB states point of view, state oriented constraints are:

      1. State constraints

      2. State transition constraints

      3. State sequence

   Below is a comparison of integrity supported by different common commercial RDBS:

   

3. Redundancy Control: Due to the fact that the web portal is the exchange center of all information flow, reduce data redundancies and inconsistencies are crucial.  It is said that in a distributed system dictionary proper cataloging techniques can efficiently reduce redundancy issues.

   In our project, for a set of distributed database systems, automated procedures which performs functions in the distributed database dictionary can be proposed to implement on the distributed data base management system or part of network operating systems or ad hoc to ensure information consistency and reduce redundancy.

4. Efficiency of Dictionary: For the dynamic nature of the DBs and data involved in the project,  and the required data might belong to either single or several different DBs, the necessity to have both global and local dictionary under one network dictionary for distributed dictionary on centralized systems.  ( shown in the picture below)

1. About BrioQuery:

• Two types of table: Fact table and dimension table, can’t join fact tables to fact tables; join Dimension tables to Fact table(s)
• Use same name keys to join Dimension tables to a Fact table, i.e.  join Cost Collector Key on the Dimension table to Cost Collector key on the Fact table.
• Different Fact Tables can reuse the same Dimension tables, i.e.  if user is familiar with the GL ACCOUNT REPORT table,  this table can be reused to create many different stars.
• Only one join is allowed between any two tables
• In a query, all tables have to be joined in order to process.