Re: SQL Puzzle - too many dimensions
On 13.05.2011 12:55, Lawrence D'Oliveiro wrote:
In message<dddba9e2-7e37-4fa4-
bcb2-2e575deda114@w21g2000yqm.googlegroups.com>, Robert Klemme wrote:
On 10 Mai, 03:50, Lawrence D'Oliveiro<l...@geek-
central.gen.new_zealand> wrote:
In message<92r02mF24...@mid.individual.net>, Robert Klemme wrote:
That's called an "inline view".
The only reason there's a special term for it is because SQL doesn'=
t
treat tables/views as first-class objects.
There are no objects in SQL. SQL is neither procedural nor object
oriented.
I didn't say it was =93object-oriented=94. In Computer Science, we us=
e the term
=93first-class object=94 to refer to what you can do with objects, not =
object-
orientation<http://en.wikipedia.org/wiki/First_class_object>.
That article speaks about programming languages. Since SQL isn't a
programming languages strictly speaking the definition does not apply -
or at least it has to be applied with a large grain of salt.
For example, since SQL does not have variables, you can never store
anything there - hence there can be no first class objects at all in SQL.=
You would have to update the term "FCO" for application in a declarative =
language as SQL before we can even discuss about whether there are FCOs
in SQL or not - let alone in which ways SQL could be improved by making
tables and views FCOs.
It is a declarative language implementing relational
algebra.
If it DID implement proper relational algebra as it is mathematically
defined, we wouldn't have this problem.
I won't debate whether SQL does or does not implement relational algebra =
properly because obviously it does implement a reasonable subset of it
(plus a few things more) and the term "probably" leaves too much room
for personal taste and interpretation.
Remember, the whole discussion started when you said that the term
"inline view" is only necessary because tables and views were no FCO in
SQL and you claimed that this deficiency could be remedied by bringing
more orthogonality into the language. Up to now neither I nor
apparently others (see Arved's recent posting) have been informed what
it is that you find deficient and want to change about SQL.
Where's the beef? As long as you don't deliver and tell us which
particular deficiencies you see in SQL and how they can be remedied by
adding orthogonality, I cannot take your talk about deficiencies serious.=
Cheers
robert
PS: The term "inline view" is necessary simply to distinguish a view
defined in a query from a view defined in the data dictionary where it
is given a name. That's plain reasonable, and it happens all the time
with language (our cars have four "wheels" and a "spare wheel").
Nothing deficient or weird about that. We give different things
different names all the time so we can distinguish them.
--
remember.guy do |as, often| as.you_can - without end
http://blog.rubybestpractices.com/