Issue 1699:  Abstract Interface issue (write_Abstract/read_Abstract)(01) (obv-rtf)
Source:  (, )
Nature:  Uncategorized Issue
Severity: 
Summary: Summary: 1. There are no write_Abstract and read_Abstract methods on 
    DataInputCtream and DataInputStream.  This looks like an oversight
    to me.
 

Resolution: 
Revised Text: 
Actions taken:
July 20, 1998: received issue
July 30, 1998: closed issue
Discussion: 

End of Annotations:=====
Return-Path: <www>
Date: Sat, 18 Jul 1998 14:44:53 -0400
From: www
To: juergen@omg.org, web-incoming@omg.org
Subject: WWW Form output

Name: Frank Laub
Company: Inferno Graphics
Email: flaub@mindspring.com
Notification: Yes
Specification: Objects By Value
Section: 5.4, 5.5
Formal #: orbos/98-01-18
Version: 
Revision_Date: 2/10/98
Page: 5-44, 5-46 - 5-49
Nature: Revision
Severity: Significant

2.  Concerning the ValueBox, is this supposed to be a named entity?
If it is, then the ValueBoxDef needs to reflect this, with an
attribute for name, repository id, and version, and the
create_value_box() in the Container interface needs to have parameters
for this information.  If it isn't supposed to be, then the grammar
needs to be modified for the value_box_dcl production.  This
production has an <identifier> piece to it.  During parsing, the only
thing that can be done with that identifier is throw it away??  Well,
if it's unnamed, perhaps we could rework the grammar so that a
value_box_dcl appears as a template_type_spec from the original IDL
grammar?  Because of the simmilarites between an unnamed value box and
sequence type or a string type.  In this way only typedefs (aliases)
can reach a value box.  (i.e. typedef value ValueBoxName).  What do
you think about this?
 
3.  This is more of an opinion than anything else, but let me ask you
a question first.  Do you think that a value is a specialization of an
interface?  In other words, don't you think that a ValueDef could
easily inherit an InterfaceDef?  There are many similarites to a value
and an interface.  And the way I see it, a value IS A interface with
state.  Or, potentially could be.  On a behavioral basis, a value only
adds new methods to the InterfaceDef, namely for creating initializers
and value members.  For the implementor of the Interface Repository,
this could make their life easier.  The only problem I see with this
is the describe_interface and describe_value methods clashing.  This
really isn't a problem since outsiders could view a ValueDef as either
a normal interface, or a value.  Viewing it as an interface would give
them the describe_interface method.  Viewing it as a value would give
them the describe_value method.
submit: Submit Issue Report