Message-ID: <8D18C4F9CBA1D311900F00A0C990C97F67C8CD@neimail.networkengines.com>
From: Andrew Sporner <andy.sporner@networkengines.com>
To: "'db@aptant.com'" <db@aptant.com>, pups@minnie.cs.adfa.edu.au
Subject: RE: save everything and divisiveness
Date: Mon, 19 Jun 2000 10:30:39 -0400
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: owner-pups@minnie.cs.adfa.edu.au
Precedence: bulk

+ my $0.02 makes $1.00

>From my perspective I have watched this argument on this
list about purism and otherwise.

>From a practical sense, historical trueness makes sense
when we are considering changes to something.  That is 
to evaluate whether it was better before or after; with
the ultimate goal of coming up with a truly usefull sytem.
Otherwise O/S researchers would never be able to make
advancements because they would be repeating each others
mistakes.   But to take a lesson from history makes having
such an archive of old source important.

To get hung up on a particular release makes sense I guess
if you are a collector, such as one who collects vases
because that is an art form.  A vase from the Ming chinesse
period is worth more if it has not been modified (for instance
some later owner decides that there are not enough flowers
on the vase--so he adds some).   However with Systems software
this is not the case because it is not a tangible item such
as a processor such as a PDP-11 or PDP-8.  I know many people
that still run PDP-8's (I have one myself), but universally
ever user of the '8 is trying to make the software on it
run better and more efficiently.  

So I would not be one to castigate some pioneers of systems
software whoses names happened not to be K&R.  I am sure that
the both Kernigan and Richie both are marveled at what Unix
has become.  In fact I believe one of them went on to write
Plan-9 which is really off-the-wall compared to their earlier
work. 

Good software is inherrently in a steady process of evolution.
The only piece of software I have ever seen that never evolved
was the classic "Hello World" program that everybody learns to
write on their first lesson in programming.  

OK, That's it...



Andy Sporner


> 
> My $0.02:
> 
> 
> I once wondered whether the techniques of literary textual
> criticism could be used in order to determine whether a Linux,
> FreeBSD, groff -- whatever! -- is in any way derived from an
> earlier work. Textual criticism considers a work by examining
> several or all of the extant textual variations in an attempt
> to determine what the author originally wrote; it has been
> used to reconstruct the "original" texts of the ancient as
> well as some modern writers, such as James Joyce. It
> yields a tree of texts, in which the root is the "original,"
> and the sibling children of any node are the descendants of a
> common, perhaps hypothetical, text. I don't know much
> else about it, except that its results may depend on alot
> of knowledge and informed speculation. The textual critics
> work bottom-up to arrive at an original text; I am thinking
> of a top-down process, working from an original text, to show that a
> work lower in a tree is derived from the original. If such a
> technique were valid at all, its validity would only be improved
> with the availablity of many, many "texts." The techniques might be
> more useful where, for example, there were several V7 tapes 
> that people
> thought were original, but which, on inspection, turned out 
> to be different.
> In this situation, textual criticism might be used to 
> reconstruct a "true," V7
> release tape, and, in this situation, would be a bottom-up 
> application of the techniques.
> 
> In any event, I think that it is important to preserve alot of
> tapes, and to keep them separate with as much information as
> possible about their pedigree. If someone ever did use such
> a technique -- or any other technique -- to reconstruct a 
> "true" release,
> it is important that they document their work and not throw away the
> tapes that contributed to the "true" tape, because even more
> tapes may appear in the future which could lead to
> the reconstruction of an even truer tape.
>