The world rejoiced as mascarm@mascari.com (Mike Mascari) wrote:
> It's a very provocative read. At a minimum, one can learn what to
> avoid with SQL. The language looks neat on paper. Perhaps one day
> someone will provide an open source implementation. One could envision
> a "D" project along the same lines as the same sort of project that
> added SQL to Postgres...
I think you summed it up nicely. The "manifesto" is a provocative, if
painful, read. It is very useful at pointing out "pointy edges" of
SQL that might be wise to avoid.
I'm not thrilled with the language; I think they have made a mistake
in trying to make it too abbreviation-oriented. They keep operator
names short, to a fault.
As you say, the most likely way for a "D" to emerge in a popular way
would be by someone adding the language to an existing database
system.
There is a project out on SourceForge for a "D implementation," called
"Duro." It takes the opposite approach; the operators are all defined
as C functions, so you write all your code in C. It uses a data store
built atop Berkeley DB.
I think an implementor would be better off using an SQL database
underneath, and using their code layer in between to accomplish the
"divorce" from the aspects of SQL that they disapprove of. Sort of
like MaVerick, a Pick implementation in Java that uses a DBMS such as
PostgreSQL as the underlying data store.
You do a "proof of concept" by building something that translates D
requests to SQL requests. And _then_ get a project going to put a "D
parser" in as an alternative to the SQL parser. (Yes, that
oversimplifies matters. Tough...)
--
let name="cbbrowne" and tld="ntlug.org" in name ^ "@" ^ tld;;
http://www3.sympatico.ca/cbbrowne/rdbms.html
Rules of the Evil Overlord #81. "If I am fighting with the hero atop a
moving platform, have disarmed him, and am about to finish him off and
he glances behind me and drops flat, I too will drop flat instead of
quizzically turning around to find out what he saw."
<http://www.eviloverlord.com/>