From lechner@cs.uml.edu  Sun Oct 29 01:59:00 2006
Subject: Re: [UML_CS522_f2006_OOA] Re: Schema file for asgnt3
To: UML_CS522_f2006_OOA@yahoogroups.com
Cc: 06f522
RJRef: $PH/06f522/asgnt3/asgnt3_msdat_ns_rjl061029,txt

Thank you Nitin. Without loking at CK and CA in metaschema
I only looked at the last 3 CK specs (CK# 10,11,12)
The data you entered there DO make sense.

I'd like to know 2 things:
(1) By what processdid you define the rows of tables Ck and CA?
	(just a  guess: manual editing :-) 
(Others should appreciate that this took a lot of work to get right.)

(2) Your file is named asgnt3.msdat;  below it is called asgnt3.dat
	Which do you mean?

One consideration: TT-rows  6 and 7 declare tables CK and CA.
That is new metadata from metaschemaCK.sch and certainly 
belongs in metaschemaCK.msdat along with metaschema and SIS 
table declarations.

I would probably put the Ck and CA table row CONTENT into the same
.dat file as asgnt3/lechner/metatest3/SISrev061011.dat 
(the sample database of SIS rows
against which the queries will run and call keyExists).
However I can't claim that it matters [or not].

The distinction between .msdat and .dat file types is not clear but fuzzy,
apart from the firm need to make rows of .msdata read-only (const):
All code generation depends on the data model (DB format) not changing.

CK and CA are in the middle - it depends on how volatile they are.
One extreme pre-defines and can pre-compile all legal queries. 
The other extreme interprets each query dynamically and might infer 
its CK-->CA content on the fly. SQL can have it either way.

OTOH, .dat files are SIS DB content and are variable (although with 
differing degrees of volatility), The CQ query value list is certainly
variable and arguably need NOT be viewed as SIS database content.

To my way of thinking, formalizing the CQ test instance format 
in the schema and pr_loading each CQ value list makes sense 
because chgen -log also generates logging and replay capability.  
That suggests 'canning' the test procedure 
with the replay option as a possible test driver.

R Lechner

PS: I still expect inputs by TUes as to what each person 
believes he/she is tasked to do as a contribution to asgnt3.??

--------------------
> To: UML_CS522_f2006_OOA@yahoogroups.com, nitin.sonawane@gmail.com
> From: "Nitin Sonawane" <nitin.sonawane@gmail.com>
> Sender: UML_CS522_f2006_OOA@yahoogroups.com
> Subject: [UML_CS522_f2006_OOA] Re: Schema file for asgnt3
> 
> Hello,
> 
> I've uploaded a version of data file for asgnt3 called asgnt3.dat. This file
> has all the various candidate keys and candidate key attributes populated.
> The URL is
> http://www.cs.uml.edu/~nsonawan/asgnt3/asgnt3.msdat
> Also, in the asgnt3 directory, is my starting point for the test driver
> (which was used to generate the above .msdat file). The schema file used is
> in the directory
> 
> More to come..
> 
> Thanks,
> Nitin.
> 
> 
> On 10/28/06, Nitin Sonawane <nitin.sonawane@gmail.com> wrote:
> >
> > Hello again,
> >
[cut]


