home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
ftp.cs.arizona.edu
/
ftp.cs.arizona.edu.tar
/
ftp.cs.arizona.edu
/
tsql
/
doc
/
tsql.mail
/
000065_UZTBB@CUNYVM.CUNY.EDU _Fri Apr 2 12:05:59 1993.msg
< prev
next >
Wrap
Internet Message Format
|
1996-01-31
|
3KB
Received: from Arizona.edu (Penny.Telcom.Arizona.EDU) by optima.cs.arizona.edu (5.65c/15) via SMTP
id AA15850; Fri, 2 Apr 1993 10:09:39 MST
Received: from CUNYVM.CUNY.EDU (MAILER@CUNYVMV2) by Arizona.edu (PMDF #2381 )
id <01GWJ5S0O2YO8ZEWWY@Arizona.edu>; Fri, 2 Apr 1993 10:08:24 MST
Received: from CUNYVM.CUNY.EDU (NJE origin UZTBB@CUNYVM) by CUNYVM.CUNY.EDU
(LMail V1.1d/1.7f) with RFC822 id 1595; Fri, 2 Apr 1993 12:06:44 -0500
Date: 02 Apr 1993 12:05:59 -0500 (EST)
From: UZTBB@CUNYVM.BITNET
Subject: benchmark draft
To: "C. Jensen" <csj@iesd.auc.dk>
Cc: tsql@cs.arizona.edu
Message-Id: <01GWJ5S0O2YQ8ZEWWY@Arizona.edu>
Content-Type: TEXT/PLAIN; CHARSET=US-ASCII
Content-Transfer-Encoding: 7BIT
ON THE TSQL BEMCHMARK DRAFT
On the expressive power....
_____________________________
Crist states that the aim of this draft is on user friendliness of
temporal query languages. I think the user friendliness and the
expressive power of temporal query languages are closely related.
We should not seperate them. In this respect we should include
all the possible queries in the bench mark. Perhaps, a subset of
these queries is more common in real life and this subset may be
a desirable one. At the earlier stage of the benchmark we place
our attention on this subset if thw TSLQ community feels that way.
Allowing nested queries seems to me more reasonable
____________________________
Using each relation once is restrictive
_________________________
Update and modifications should be included. Retrieval and
update operations form a minimal logical subset of the operations
of any query language. I do not have the same strong feelings
about the aggregates.
On the relation schemes
________________________
I think we should not bother whether benchmark relations are in 3NF.
These relations specify only the sematics of the real world it
describes. A particalar temporal data model would put it into its
proper normal form.
It is not clear to me whether skills is a temporal attribute or not.
If it is not I propose to make it temporal. To clearify, salary
coverts to a multi-valued attribute when time is considered.
Skills is a multi-valued attribute without time and it is also
multi-valued with time. There may be multi-valued attributes
without time as well, i.e eye colors ( if anybody is interested in them).
I would propose adding a BUDGET attribute to Mgr relation. It is a
temporal attribute and allows comparison between two temporal attributes
from two different relations.
I propse keeping both Gender and B-date in emp relation. Both are
constant but B-date allows comparison on time.