Re: Hierarchical Relationship

<joe_celko_at_my-deja.com> wrote in message news:8atkfk$l14$1_at_nnrp1.deja.com...

> >> Same here, although I do wish he's get back to us about why Access
> is such a "terrible product"... <<
 

> The datatypes are wrong, the square brackets and quote marks are wrong,
> the BETWEEN predicate does not work, the subquery predicates are wrong,
> the DISTINCTROW is an abomination, the optimizer is pathetic, it is
> missing so much of SQL-92 that other desktops like SQL Anywhere have,
> etc. Someone had a list of about 150 differences in ACCESS versus SQL-
> 92; maybe ic an fidn it again and post it.

What about TOP? Do these objections apply only to DAO, the data access API used by Access until Access 2000, or do they apply also to ADO? ADO, while still buggy, is supposed to make all of the problems of DAO, RDO, ODBC, ODBCDirect, etc. etc., go away, since Microsoft's "Total Cost of Ownership" initiative means that many problems in existing products can't be fixed for fear of breaking the existing workarounds. I wonder how SQL-92-compliant ADO really is.

> While not part of the language, ACCESS programmers have a hell of a
> time learning to think in sets instead of procedures. I spend so much
> time trying to UN-learn them in classes. For example, ACCESS
> programmers write VB code instead of thinking about a pure SQL answer

ISTR your having similar concerns about stored procedures. Most any host language or SP environment can encourage one to give up too easily on finding a SQL solution, for that matter.

> to a problem. They put those stupid little prefixes on names (i.e.
> tblPersonnel, intQuantity, etc.) which are at best redundant because
> SQL knows what the schema objects are from syntax. That is why you can
> legally give a column and a table the same name (don't do that!).

Finally, someone who agrees with me on this! Isn't this why compilers have symbol tables?

> In the 1950's FORTRAN needed to be told the identifers that began with
> I thru N were integers and the rest were reals. In tyhe 1960's, BASIC
> needed to see a dollar sign to tell strings from reals (i.e. X and X$).
> Compilers were both weak and close to the machinery in those days.
> What this practice says about the mind set is that the programmer needs
> to see the physical storage at all time. Without a strong abstract
> reasoning ability learning advanced SQL is almost impossible.

The "Hungarian" prefix notation everyone seems so hot-and-bothered about seems to have arisen from the bowels of Microsoft in the 80's, well after the invention of the symbol table. Type prefixes are too much like comments in that a little code drift can make then horribly misleading!

--
Joe Foster <mailto:jfoster_at_ricochet.net>  Space Cooties! <http://www.xenu.net/>
WARNING: I cannot be held responsible for the above        They're   coming  to
because  my cats have  apparently  learned to type.        take me away, ha ha!