18 Jun
2009
18 Jun
'09
11:50 a.m.
On 18 Jun 2009 at 11:56, blstuart at bellsouth.net
wrote:
For me the question is whether the change is irreversible.
I'm actually not very concerned with what is shown to the
tourist. I want to make sure that when a question arises
that can only be answer by access to the artifact itself,
we haven't thrown away the possibility of finding the
answer. If the choice is that it never runs again but we
preserve the possbility of historical research versus we
get to see it running but we permanently lose historical
information, I'll generally prefer the former.
I'm certainly no archeologist, but one of the things that
I've noticed in documentaries is that something has changed
in the practice of archeology when compared to say 100
years ago. Now if something cannot be done without
some degree of assurance that the change will not be
harmful (e.g. cleaning, repairing, etc), then it's just not
done. Instead the piece is kept until some future time
when the techniques and technology allow it to be
cleaned or repaired without any damage to the history.
If an old drum can be replace by a flash device as a
module that connects to the same connectors and matches
the drum electrically, then I don't have much of a problem
as long as a) the change is documented in such a way
that it is clear how to restore the machine to its original
condition, and b) none of what is removed is discarded.
But if someone makes an irreversible modification
to a
machine of which only a handful still exist, then that
substantially increases the chances that at some point,
no examples will exist in their original condition. Then
significant historical information will be lost.
For the really classic machines, the issue may be one of "how do we
get this thing to operate at all?" versus "well, it's a nice museum
piece--too bad it won't ever work again".
Replacing an antique memory device (Williams tube, thin-film, delay
line or even core) with modern memory may be the limit of what can be
done at reasonable expense. If it gets the system running, one can
always leave the old device in place for show. I think bitsavers' archiving tdocumentation is
probably more, or at
least as important as preserving the actual iron.
I absolutely agree with that.
I've learned more
about some old machines by perusing the archives than I ever would
have playing with the real thing. One aspect that the documentation
can convey that the hardware doesn't is "why we did it this way".
i.e. the thought process behind the implementation--and to me, that's
far more important.
I agree that the intellect behind the design is more interesting
than the realization of the design. But all of us who are
engineers and programmers know all too well how documentation
and artifact often don't match. If a part of the documented
design doesn't make sense and doesn't appear that it would
work, then referring to an actual machine that was once known
to work (though may not be operational now) will provide
information you can only guess at otherwise.
Playing with old hardware without documentation is
sometimes like
looking at object code without source. It's a grand puzzle and you
can figure out the logic, but you lose the sense of the human behind
the design. A well-commented piece of source code is worth its
weight in gold for getting the big picture and the mindset of the
programmer. The same goes for hardware technical manuals.
I agree to some extent, but I'll also say that neither the
comments nor the technical manual is required to get into
the mindset of the designer. Sometimes just looking at
what chips were chosen and how they are laid out on the
board can give some insight into how the designer thought.
Most of the code to come out of the Bell Labs systems group
is particularly sparse in comments. But reading that code,
looking at the variable names chosen, looking at how
functionality was divided among files, etc gives a better
picture of how someone like Ken Thompson thinks than
looking a the commented code from a commercial software
house.
BLS