[AccessD] Database Patent

Dear List:
 
A client who is a patent and trademark attorney is interested in opinions on
a patent - basically it's a way to update databases on-the-fly by
disconnecting the data from the data definition. 
 
You can download the patent by going to this web page: 

http://www.pat2pdf.org/
 

and entering the patent number 7003524.

I guess I'm looking for opinions on whether this is a solution to a problem
that exists and whether there is something of value in this approach. 

I know there's some deep database design talent on this list so I'm thinking
this is the best place to start the research.

Here's a teaser.  From the patent:

TECHNICAL FIELD

The invention relates generally to databases and, more particularly, to the
underlying schematic structure of databases.

BACKGROUND

Large databases running in publicly accessible environments are notorious
for their inability to accommodate change. In today's world of massive
access to large databases via the Internet, it is increasingly common to
encounter messages to the effect of "database is down for maintenance"
instead of the actual data requested.

In a conventional database environment, when a change needs to be made to
the schematic structure of a database, the data in the database must be
extracted from the database in the old structure and re-written to the
database in the new structure. If new data were to be inserted into the
database while said changes were being effected, it could cause
unpredictable effects to the database. Such effects could include corruption
of pre-existing data, misapplication of database changes, misalignment of
data relative to internal data boundaries, or any number of problems that
could render the database effectively incoherent. Such results are untenable
in most live database deployments.

Conventionally, the most common solution to the problem of updating during
changes to the schematic structure is to simply disallow it.

Conventional databases also require a tight bind between the data type and
the data storage. Users require that the data they request be presented in a
manner consistent with the expected usage of the data. For example, a date
may be stored in the database as a string of decimal digits (e.g.,
20010303), but to present the data to the user in its raw form would be
unacceptable. A conventional computer user requires that it be presented in
a manner consistent with its usage (e.g., Saturday, 3, Mar. 2001. In order
for the date to be presented in a manner consistent with its usage, the
database must carry type-related information along with each unit of data.

Binding between data and type is conventionally accomplished by organizing
the data into metaphorical rows and columns. Rows of data are divided into
pre-defined columns, where each column represents a particular data type
and/or use of the data. Such data/type binding allows a computer program to
make assumptions and inferences about the data appropriate to its type.
Additional rows of data may be readily added to a database. However, if a
new column is desired in a database, then the database must typically be
made unavailable for a period of time so that data can be converted into the
new format. Modifications to pre-existing programs would have to be made,
along with the requisite testing and debugging necessary to validate any new
code.

It is also worth noting that in conventional databases there tends to be
redundancy in the storage of data. For example, cities, states, zip codes,
and telephone area codes may be repeated among a number of rows of data.
Such redundancy results in inefficient use of memory.

In light of the foregoing, it is apparent that there is a need for a system
and method for modifying the schematic structure of a database without
making the database unavailable for the entry of new data. Preferably, such
a system and method would, among other things, also minimize redundancy of
data in a database.

SUMMARY

The present invention, accordingly, provides a database in which an
application program accesses data stored in a records data set (RDS) and a
categories data set (CDS). The RDS contains at least one RDS entry having a
data field configured for representing data. The CDS contains at least one
CDS entry configured for being associated with at least one RDS entry and
having at least one data type field sufficient to describe the type of data
contained by the at least one associated RDS entry.

The invention thus provides a database whereby data is isolated from its
definition, thus avoiding the problems associated with the rows-and-columns
model enumerated above, and permits the underlying schematic form of the
database to be changed based on modifications to the CDS, without requiring
any actual changes to the schematic structure of the RDS or the underlying
database. By virtue of this same isolation it is also possible to reduce
data redundancy with varying levels of granularity and create an
object-oriented data model.


MTIA

Rocky