🚀 go-pugleaf

toms@fcs260c2.ncifcrf.gov (Tom Schneider) <C8Byt1.Axp@ncifcrf.gov> :
| 
| >   1) erasure is not a source of  energy dissipation
| >      (see landaure, IBM J. Res. Dev, vol 6, (1961), 183)
| 
| Let's consider the coin again since it is SO simple we can't make mistakes
| (ha!).  Consider a coin sitting on a table.  It can store 1 bit of

	[ hypothetical manipulation of a coin, see Tom's article ]

| So what are these folks saying?  Erasure - what's that?  It means to PRIME a
| device and then to DISSIPATE away the energy.  Note that erasure is identical
| to the operation of storing information.  They have been fooling with your
| brain telling you that erasure was something special!
| 
| FURTHERMORE, from the example above, it is clear that the storage portion of
| the erasure requires dissipation.


I think this touches on a semantic ambiguity:  What is erasure?  When I
read Tom's posting, I initially disagreed with the description of
"Erasure - what's that?"  To me, erasure is the described priming
operation.  The "operation of storing information" that subsequently
dissipates energy isn't part of the erasure as such, it's a practical
step based on the fact that rotating coins and dynamic RAM cells can
neither one remain in the ambiguous state for very long (only about as
long as Schroedinger's cat can hold its breath :-).

But consider a sheet of paper, on which I write a zero or a one (or an
A, T, G, C, or whatever) to record some item of interest.  If I erase
this item, I "prime" the space on the paper by removing the previous
mark.  But I *don't* write a zero or any other placeholder value; I just
leave the paper blank until I want to store some new information.  This
cat can hold its breath as long as it wants to.

But this is the ambiguity.  On a sheet of paper I really do erase the
old information.  In a computer memory (or disk, etc.) I don't literally
erase information.  Instead, I *overwrite* it with some constant.  What
Tom describes above is not the dictionary definition of erasure --- but
it is the mechanism used by many or most storage systems to remove
previous information.  So maybe I accept Tom's points, or maybe not.  It
depends on the storage system, and what I intend to do to erase it.

I won't attempt a thermodynamic view of this except to comment that
erasing a mark from a sheet of paper can dissipate even more energy than
writing it did, depending on the pencil and how badly the eraser's been
chewed.

-bob,mon.

(disclaimer:  My tangents are my own.)