[PET] Negative Results in Security and Privacy

Wright, Matthew mwright at cse.uta.edu
Mon May 30 21:42:37 BST 2011


I would argue that, at this stage of our field's development, this is not necessary.

One reason for this is that much of our work is more like engineering than science. If I have an idea to enhance a PET, I get credit only if the idea is a good one in some respects. If the idea is no better than existing ideas, the only reason to publish is if it's an obvious extension from existing ideas (e.g. k-anonymity to location privacy) and the study is comprehensive enough to rule out most approaches along the same lines (just because I couldn't make it work doesn't mean much).

For work like Sören's, which is science, we should have venues to publish in. But I think that such venues exist -- surprising results of this type seem to be reasonably well-received among the community as far as I know.

I used to think that reproducing results was an important activity, and it is in some contexts, but very few results are worth directly reproducing in the long view. Good ideas get mimicked in various ways and further refined, which exposes them to further testing and thinking (again, in the context of engineering). Even attacks get revisited.

-Matt

On May 30, 2011, at 12:44 PM, Sören Preibusch wrote:

> When looking for negative results, inter alia, our recent field experiment
> did reveal: consumers do not prefer privacy-friendly retailers even if
> everything else is the same
> <http://ideas.repec.org/p/iza/izadps/dp5017.html>, "Unwillingness to Pay for
> Privacy: A Field Experiment".
> 
> So, yes, such findings are relevant.
> 
> Sören
> 
> -----Original Message-----
> From: pet-bounces at lists.links.org [mailto:pet-bounces at lists.links.org] On
> Behalf Of Matthijs R. Koot
> Sent: 15 May 2011 17:16
> To: Discussion of privacy enhancing technologies
> Subject: [PET] Negative Results in Security and Privacy
> 
> Hi pet at lists.links.org,
> 
> Biomedicine has a Journal of Negative Results in Biomedicine:
> 
> http://www.jnrbm.com/
> 
> Could a "Journal of Negative Results in Security and Privacy" (and/or
> Surveillance) be viable? Topic for a panel discussion at PETS 2011?
> 
> Best regards,
> Matthijs R. Koot
> University of Amsterdam, NL
> _______________________________________________
> PET mailing list
> PET at lists.links.org
> http://lists.links.org/mailman/listinfo/pet
> 
> _______________________________________________
> PET mailing list
> PET at lists.links.org
> http://lists.links.org/mailman/listinfo/pet



More information about the PET mailing list