Date: 10 Oct 93 12:05 +0500
From: "P. Desjardins" <Desjardins@citi.doc.ca>
To: <uri@bunyip.com>
Message-Id: <1220*Desjardins@citi.doc.ca>
Subject: UR*, wrappers,...
The "U" (uniform/universal... and eventually ubiquitous?) naturaly implies some
level of genericity ("granfathering") - which usually ends up being "our best
shot" level because we often have only an incomplete/imprecise picture of the
problem domain. Getting this "best shot" to be as close as possible
to ultimate genericity, depends on how imcomplete/imprecise our statement of
the underlying requirements is, since it will drive the next step which is
establishing a coherent taxonomy of problem domain objects.
The "R" (resource... and hopefully resoluble?) ranks pretty high, at least in my
universe of discourse, on the genericity scale in terms of the problem domain
objects it potentially stands for.
The "*" (whatever... "star light, star bright,...") looks to me as a top layer
classification (taxonomy) of the problem domain objects. Each of these classes
seems destined to satisfy a specific set of requirements which ultimately
biases the syntax of the textual representations used as the specifiers of
relevant problem domain objects.
In the context of the above, am I correct in understanding that
the current debate on wrappers is in fact trying to get agreement on the notion
of a default (* = L) class of objects?
-Pierre
---
Pierre Desjardins | Desjardins@citi.doc.CA
CITI | S=Desjardins; P=CCRITmhs; A=telecom.canada; C=ca
1575 Chomedey blvd |
Laval, Quebec | +1 (514) 973 5820 (voice)
Canada H7V 2X2 | +1 (514) 973 5757 (facsimile)
---