[LINK] RFID Tags - Pesky Privacy Advocates Delay Nirvana

[Chirgwin, Richard]

[Note: Graeme; I have copied you on this response to the column because if I
make an error of fact, please feel free to correct me!]

There are technical lucanae in the article which deserve correction, because
they impact the validity of the conclusions. I'm going to skip the
philosophical stuff at the start and cut to parts where my knowledge is
relevant...

[snip]

>   The battle is over the rapid proliferation of RFID (radio frequency 
> identification) tags for merchandising. RFID tags are tiny (about the 
> size of a grain of sand) microchips that can be attached to any item, 
> such as a tube of toothpaste or an item of clothing, and which can 
> transmit details of that item's location for up to 20 years.

RC: This is one of the oversimplifications which feeds both fear and
confusion. Some RFID tags are tiny; generally the passive retail tags. But
one example of packaging and capability does not sum up the entire class of
product.

RFID tags do >not< "transmit details of that item's location." They do not
*know* an item's location. All any RFID tag knows, of itself, is that it has
received a wake-up signal from a reader, and how it should respond.
Everything else is inference from the location of the receiver.

> 
>   Retailers around the world are implementing RFID technology as a 
> replacement for bar codes. 

RC: To date, successful deployments have been of largish active tags, for eg
in video stores, CD retailers etc, which are typically removed and recycled
at point of sale. Retailers around the world are *evaluating* passive tags,
but implementing is in the future.

In my opinion, you can't base a "don't worry" line on confused tense.

[snip]

>   RFID has many technological advantages over bar codes. A bar code 
> contains about 100 bytes of information, a RFID tag can contain 80 
> times more, allowing the tags to contain individual item numbers, 
> where a bar code is typically limited to the product code. They can 
> also be modified, allowing details of the sale to be recorded on them 
> at the time of purchase. Think about it.

RC: Once again, two forms of technology are combined ("Composition"
according to my "Twelve Errors of Logic" from the Cadillac Encyclopaedia).
Passive tags are very hard to rewrite, because they have no on-board power
supply and it's damn near impossible to supply enough electricity "over the
air" to rewrite stable logic (like Flash for eg).

Now, as to the move from 100 bytes of information to many Kb. This creates a
critical problem for IT systems; to turn two bytes (99) into four bytes
(2000) cost billions. The vendor side of IT knows RFID presents a similar
"upgrade opportunity" - they even talk about it in briefings. What is not
discussed is whether there is a net cost benefit.

> 
>   RFID tags do not depend on line of sight, and can be read through 
> walls or clothing. 

RC: Composition again. An RFID tag with a battery fits this bill. A passive
tag most certainly does *not*. It delivers a signal at pico- or femtowatt
levels. A long white paper I read a couple of months back on RFID reader
antenna design was quite clear: if you want a manageably-sized antenna for a
retail application, you must not expect a range of greater than about 6". 

The other alternative is to put a HUGE antenna on the tag itself. But to get
a transmit power from the tag of 1W, you would need antenna length in the
order of tens of metres (or more, I can't be bothered doing the math right
now).

> They can transmit their information to any RFID 
> reader, using power picked up from that reader, and operate in a much 
> greater range of environmental conditions than bar codes.

RC: "Any compatible reader" is not "any reader". Different applications are
already splitting into different standards, particularly wrt to frequencies
and modulation schemes (low power tags can only manage the very
interference-prone AM modulation; active tags can use more sophisticated
modulation). 
Again, assumptions which are true for active tags (range of environmental
conditions) are not true for passive tags (for eg, you can't use passive
tags in a warehouse of Campbell's Soup, for the obvious reason...).

[snip]

>   Australia's Coles Myer and Woolworths chains are trialling the 
> technology but in the US retail giant Walmart has mandated it. And 
> just a couple of weeks ago, Microsoft formed its own RFID council to 
> ensure its partners use RFID is such a way that it can be 
> incorporated into future Microsoft products.

RC: "Composition" again: passives are RFID tags, therefore all RFID tags are
passive tags, therefore every trial or mandate is for passive tags.

> 
>   Clearly, RFID is a technology whose time has come. 

RC: Not so clearly, IMO. Lessons which should have been learned out of
previous failures of delivery clearly have not been learned.

Even the standardisation effort is forking. For example, Walmart has said
"no thanks" to the object naming service standard (presumably because it
doesn't want to hand off its product query data to Verisign). 

To counter this piece of analysis, I offer my own:
- Passive, item-level retail tagging is, at the moment, an adventure
playground for engineers, not a serious business technology;

- Particular shortcomings include reliability, robustness, and a
standardisation effort which in many places wants to institutionalise
external dependency and vulnerability. A retailer which adopted passive RFID
at the current state of the technology is signing on for a very expensive
failure.

- Pro-RFID analysis imposes a false dichotemy on the debate - "use RFID or
do nothing". 
There is, of course, a third alternative: create new practises and processes
based on information which already exists and does not need new technology.
It will be cheaper to move from a "90% accurate" view to a "95% accurate"
view than to go from 95% to 100%; so the smart business will weigh the cost
of the last 5% against the cost of a company-wide RFID rollout, and base its
decisions on that rather than on vendor puffery.

Richard Chirgwin
> 
> 
> [Nice tag-line (so to speak), and a decent analysis.  Pity it adopts 
> the anti-social values of the pro-technology-industry lobby.  Public 
> interest advocates are viewed as political obstacles, rather than as 
> people drawing attention to the harm that inappropriate uses of 
> technology do to people, and presenting arguments that demand at 
> least as much attention as McNeely's ravings]
> 
> 
> -- 
> Roger Clarke              http://www.anu.edu.au/people/Roger.Clarke/
> 			            
> Xamax Consultancy Pty Ltd, 78 Sidaway St, Chapman ACT 2611 AUSTRALIA
>                  Tel: +61 2 6288 1472, and 6288 6916
> mailto:Roger.Clarke at xamax.com.au            http://www.xamax.com.au/
> 
> Visiting Professor in the eCommerce Program, University of Hong Kong
> Visiting Professor in the Baker Cyberspace Law & Policy 
> Centre, U.N.S.W
> Visiting Fellow in Computer Science, Australian National University
> _______________________________________________
> Link mailing list
> Link at mailman.anu.edu.au
> http://mailman.anu.edu.au/mailman/listinfo/link
>