[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference 7.286::digital

Title:	The Digital way of working

Moderator:	QUARK::LIONELON

Created:	Fri Feb 14 1986
Last Modified:	Fri Jun 06 1997
Last Successful Update:	Fri Jun 06 1997
Number of topics:	5321
Total number of notes:	139771

3126.0. "What else can I do ?" by ULYSSE::FINKA () Wed Jun 01 1994 06:23

(extract from a memo sent to my collegues)
    
Guys,

I'm very pleased to discover the SQL syntax style is used in PINS Terminal
Emulator. I am struggling in specifying a language to access IMS TE. For
instance VAX SCAN has its own proprietary language which cannot be used at
run-time. SQL is the answer !

This comforts me that the goal situation in the area of interface specification
is a self describing dynamic API based on an extended SQL used in all cases to
access any service :

	- Data Access
	- Terminal Emulation
	- ... and Module Access

The description of a part-number can be implemented by a DA, an MA or a TE (or
whatever else). However the client request would be unique :
 "SELECT DESCRIPTION INTO MY_DESC FROM PARTS WHERE DESCRIPTION = RD32-A"

The Metadata would at run-time be used to select the appropriate implementation
and perform the execution.

In other words, I've been claiming for a while that the interface specification
to any service is simply an extended SQL. The metadata definitions are used at
run-time.

Because it is a standard universally known in the computer industry so ready to
be immediately adopted and used, because it is simple and closest to natural
language, the extended SQL gives you the right specification to decouple the
service invokation from its implementation.

Defining such general SQL standard in Digital could be an excellent opportunity
and leadership in computer solution integration. And benefits would be
immediate to Digital.

Guess what's going to happen. Digital by not taking this opportunity will
discover the benefits of the extended SQL from outside. Signs are already
coming from SAP R/3 for instance.

1/ Is there an authority in Digital to address such important issues ?

2/ Why are we sentenced to miss such an opportunity ?

Cheers,
Jean (frustrated too !)

T.R	Title	User	Personal Name	Date	Lines
3126.1	SAY WHAT?????	ANGLIN::SULLIVAN	Take this job and LOVE it	`Wed Jun 01 1994 10:37`	4
	RE: .0 What the heck did he say/What the H are you talking about????
3126.2	One more time...slowly!	PARVAX::SCHUSTAK	Join the AlphaGeneration!	`Wed Jun 01 1994 10:52`	1
	I have to echo the sentiments of .1
3126.3		NPSS::BRANAM	Steve, Network Product Support	`Wed Jun 01 1994 11:58`	14
	I think what he is he saying is that SQL is a good command language for specifying requests to many types of systems, not limited strictly to database queries, and that Digital should consider adopting it as a generalized command language interface. The code behind the scenes would implement whatever needed to be done to carry out the request, not just do database retrievals. Just my interpretation for the non- SQuiLlish, not sure I entirely agree. I do sympathize with his sentiment that there are too many dissimilar command intepreters running around. Just try going from NCP command syntax in DECnet Phase IV to NCL syntax in DECnet/OSI to do essentially the same job. A little banging of the head on the wall to reorient your thought processes is recommended. This could have been expressed in a more natural-language-like layman's terms, though...
3126.4	This may already exist	QUEK::MOY	Michael Moy, DEC SQL Engineering	`Wed Jun 01 1994 16:26`	9
	I believe that products in Database Systems can do what you want. We have DEC SQL which talks to Dispatch which can talk to any data source that we have written a gateway to, or that you or an oem would care to write a gateway to. We can talk to RMS, MUMPS (I believe), DBMS32, Oracle, Sybase and others. michael
3126.5	who can figure?	DYPSS1::DYSERT	Barry - Custom Software Development	`Wed Jun 01 1994 16:51`	11
	Re: Note 3126.3 by NPSS::BRANAM �I do sympathize with his sentiment �that there are too many dissimilar command intepreters running around. �Just try going from NCP command syntax in DECnet Phase IV to NCL syntax �in DECnet/OSI to do essentially the same job. A little banging of �the head on the wall to reorient your thought processes is recommended. And yet, the world clamors for Unix. BD�
3126.6		NPSS::BRANAM	Steve, Network Product Support	`Wed Jun 01 1994 17:48`	1
	RE .5 - yeah, but we know better than the world (nudge, nudge, wink, wink!)
3126.7	Topic write-locked	SMURF::BLINN	Dream of gypsy wagons.	`Wed Jun 01 1994 19:45`	7
	This note has no clear relevance to the charter of this conference. It might be appropriate for ASKENET. It is now write-locked. You can send suggestions to the topic author by mail. Tom