NetNews Usenet Archive 1992 #19/NN_1992_19.iso/spool/comp/lang/rexx/841

Internet Message Format | 1992-08-31 | 3.4 KB

Path: sparky!uunet!gatech!bloom-beacon!eru.mt.luth.se!lunic!sunic!ugle.unit.no!ugle!anders
From: [e-mail] (Anders Christensen)
Newsgroups: comp.lang.rexx
Subject: Re: Blanks, REXX, and portability...
Message-ID: <[e-mail]>
Date: 31 Aug 92 13:13:35 GMT
References: <[e-mail]>
Sender: [e-mail] (NetNews Administrator)
Organization: University of Trondheim, Norway
Lines: 67
In-Reply-To: Dave Gomberg's message of Thu, 27 Aug 1992 22:32:04 PDT

In article <[e-mail]> Dave Gomberg <[e-mail]> writes:
> On Fri, 28 Aug 1992 02:13:58 LCL Anders Christensen said:
> >In article <[e-mail]> Dave Gomberg
> ><[e-mail]> writes:
> >>       If there is no consensus among users as to which characters are
> >>       equivalent to blanks, then only the blank character should be so
> >>       implemented."
> >
> >I am not really sure how to interpret this. As long as "users",
> >"consensus" and "be implemented" (implemented where? and by whom?) is
> >rather unspecified, I think that sentence should be left out.
>
> A standard mainly tells implementers what to implement, and to a lesser
> extent, users what kind of code to write.   If users don't know certain
> funny characters are spaces, the implementers in that environment should
> not implement them that way.  Seems perfectly straightforward to me.

No. It is not straightforward.

1) It implicitly assumes that the only problems with whitespace is
   whether to have one or many. Example: Suppose half the users thinks
   that Space and Tab are whitespace, and the rest thinks that Space
   Tab and Formfeed are whitespace. Then, since there are no consensus
   among users about whitespace, inplementations can only use Space as
   whitespace.  Even though all users agree that Tab is a whitespace.

2) The term 'users' are unspecified. Is it _all_ the users, or is it
   the users within one operating system? What about variants of
   operating systems? What if all users in Europe agrees on one thing,
   and all users in America agrees on another?

   Does 'users' refer to all the users within one community, or just
   the Rexx users? Maybe just Rexx programmers?

3) The term "consensus" is rather vague. Is it more than 50%? More
   than 75%? Is it enough that _one_ person disagrees? Does
   experienced users count as much as beginners?

Given these uncertainties, an implementor can choose to give it an
liberal interpretation. Then, he may implement whatever he wants,
justifying by something like:

  "We asked three of our biggest customers, and after we explained the
   situation, they agreed with our perception of 'whitespace'"

Or even just:

  "Our interpreter targets the users who support our perception of
   'whitespace'"

Now, do you see what I meant by 'unspecified'?

A better statement might be something like

  "Exactly which characters are considered whitespace is
   implementation-dependent, but implementors are strongly recommended
   to define as whitespace only those characters commonly used as
   whitespace for the operating system in question."

Or something like that .... The reason why I care is simply that I
understood your notes as a suggestion for the ANSI standard, and IMHO
such a vague statement should not appear in a standard.

-anders

PS: It's sad to hear that postings must look like the front page of
    "The National Inquirer" (sp?) to be read.