| Building New SQL |
[Dec. 16th, 2013|09:19 pm] |
Abstract
In the following article we criticize SQL for not following the relational algebra; describe the
problem of the multitude of different relation representations in SQL; analyze in detail the
disadvantages of the notions “TABLE” and “FOREIGN KEY”; and propose a complex solution
which includes brand new query language and relatively small yet very significant alteration of the
data storage concept.
In the proposed language we will try to overcome the following apparent flaws of SQL:
– human language mimicking
-– which leads to inconsistent, completely inextensible, and needlessly complicated syntax
– mixing relational operations together and with the output
– imperative DDL
– impossibility of relations between relations
-– which leads to non-homogenous representation and undermines the relation abstraction itself,
lowering the level of programming
by abandoning the notion of table and by creating:
– frugal, extensible, functional-style syntax, utilizing “set-theory properties” of relations
– implicit and independent notions of relational operations and I/O operations
– declarative DDL
– single unified representation for relations capable of adopting another relation as a domain
To support the last and the most important novelty we propose certain storage means for relations
between relations.
http://file.bestmx.net/ee/articles/SQL-2.0/sql-2.0.pdf
http://file.bestmx.net/ee/articles/SQL-2.0/sql-2.0.odt |
|
|