RFID's - Book Tagging, etc.

[Markus Kuhn]

"Peter Tomlinson" wrote on 2002-05-04 07:54 UTC:
> This is a classic as far as the ignorance of reporters is concerned: RFID
> tags don't keep a record of activity at all, so the library book will not be
> keeping a record of every borrower. If you think about it, present day
> library SYSTEMS can already keep a record of every borrower of every book.

Perhaps also a classic as far as naivity of marketroids about
engineering realities is concerned.

Adding non-volatile writable memory to a semiconductor device adds quite
significantly to its production cost, because this requires additional
fab processing steps. Writing requires far more current than reading,
resulting in a requirement for bigger antennas and field strengths as
well as more complex circuitry. It will not be done without a good
business reason. I can't see a viable RFID retail application where the
chip needs to store more than a static unique serial number plus error
correction code (written into the chip at the fab with a laser cutter
that opens metal fuses). Any additional dynamic information associated
with the tag can far more cheaply/conveniently/reliably be stored in
databases indexed under that number, just as it is done already today
with serial number barcodes, where it is useful.

There are some proposals under way to extend the EAN/UPC retail product
barcode with a unique serial number section. This would allow for
example to recall product batches with faults by loading the affected
serial number ranges into the till databases and help in optimizing
storage, as for example best-before dates could be inferred from the
serial number if the databases along the supply chain are linkes up well
enough. Law enforcement people could more easily match serial-numbered
products found at crime-scenes to shop CCTV footage of the buyer, if
distribution channels retain sales records with accurate timestamps for
long enough, which they might do for marketing reasons. It is for many
product types of course more expensive to add an individual serial bar
code rather than to print the EAN/UPC as part of the normal package
label printing, so I toubt whether it will become widely deployed. Where
the EAN/UPC is printed today individually, it is usually just on a paper
tag (e.g. with clothes) that the customer discard anyway after purchase.

I hear occasionally stories about RFID tags proposed for being woven
into clothes to stay there after the purchase, or even into banknotes. I
haven't seen yet a technology that makes that realistically possible,
because every tag needs to be reliably bonded to an antenna, an at least
a few centimeters large conductive loop or dipole. These tend to be
rather fragile and require a rigid structure to keep them from breaking.

Pete Chown wrote on 2002-05-04 10:32 UTC:
> It wouldn't take many people to put their books in the microwave before
> this system became unworkable...

It's surprising how well the ESD protection diodes of modern
semiconductors can handle a few minutes in an 800 W microwave oven.

Markus

-- 
Markus G. Kuhn, Computer Laboratory, University of Cambridge, UK
Email: mkuhn at acm.org,  WWW: <http://www.cl.cam.ac.uk/~mgk25/>