Re: hamsterdb Transactional Storage (thanks to all of you)

From: Clifford Heath <no.spam_at_please.net>
Date: Thu, 01 Oct 2009 19:47:15 +1000
Message-ID: <FU_wm.18440$Bl2.841_at_newsfe14.iad>



compdb_at_hotmail.com wrote:
> On Sep 30, 1:02 am, Clifford Heath <no.s..._at_please.net> wrote:

>> If we had an interface with proper nested relation (rel-valued attrs)
>> support, a lot of the push to drive code into the DBMS would vanish,
>> and that'd be wonderful. But for that, a new query language has to
>> emerge.

> Date and Darwen have proper nested relations.

D is not a language; it's a set of requirements for a language. Tutorial D is only offered as an example of what a D language might look like.

I maintain there needs to be a language, then an implementation, then the implementation must be deployable. This last requirement probably means that it needs to run on an SQL backend provided by all the major RDBMs. The implementations must be multi-sourced to be adopted; enterprises won't trust their data to a single vendor any more. Finally there needs to be universities cranking out graduates able to use it, and all the end-user tools that are needed for enterprises to build, deploy and maintain applications.

We're still decades away at the current rate of progress. Maybe centuries... and it seems to be getting further way :-(.

>> If we had [X] a lot of the push to drive [Y] into the DBMS would vanish,
>> and that'd be wonderful.

> X need only be "the relational model being used". > Y is "NOT X" (see "A Call to Arms" for examples).

Yes, arguably; but for it to be used, there must be solid, deployable and well-supported implementations available from more than one vendor.

-- 
Clifford Heath, Data Constellation, http://dataconstellation.com
Agile Information Management and Design
Received on Thu Oct 01 2009 - 04:47:15 CDT

Original text of this message