Is there a physical phenomenon that could be used to record digital information in such a way that it has the following properties:

  • it relies on some immutable physical law and does not rely on any third party authority, neither centralized nor distributed

  • it is associated with a physical absolute time stamp (the time stamp can be translated into physical clock time with some fixed and useful time resolution (e.g. no worse than x hours, or 1 day))

  • the time stamp is immutable (it cannot be changed later, without detection)

  • the time stamp is unique (for some fixed resolution, no two time stamps are identical)

  • any existing time stamp is uncopyable (it cannot be recreated at a later time, without detection)

  • any future time stamp is not pre-constructible (it cannot be forged ahead of time)

  • there is a method for checking time stamp seriality (times stamps s and t can be shown to be generated by a specific source of such time stamps, and can be reliably ordered based on some physical principle)

  • [optional] it is global (i.e., time stamps can be used across more than one instance of the time stamper)

  • [optional] it is feasible as a technology (i.e., does not involve taking images with a telescope or via high energy physics experiment)

  • [optional] resilient to certain classes of failures (e.g. inaccessibility of any third party infrastructure)

If this is not possible, I would be interested in learning why. If it is possible, I would be interested in examples of phenomena that could be leveraged. In either case, I would appreciate simple-language explanations, with document links for mathematical details.

Thank you in advance for spending any time on this, and apologies if you don't consider this question to be on topic.

Update 1: @AndersSandberg pointed me to an interested paper by Haber and Stornetta (https://www.anf.es/pdf/Haber_Stornetta.pdf). On cursory reading, I suspect that it violates some of the conditions I mentioned above. Nevertheless, it's an elegant (blockchainish) idea.

  • $\begingroup$ Looking at Haber, S., & Stornetta, W. S. (1990, August). How to time-stamp a digital document. In Conference on the Theory and Application of Cryptography (pp. 437-455). Springer, Berlin, Heidelberg. anf.es/pdf/Haber_Stornetta.pdf - the part your question relates to is getting a good physical hash, perhaps a good (single, public) physical source of unpredictable bits, and a good public repository of hashes/timestamps. $\endgroup$ – Anders Sandberg Aug 6 '17 at 9:24
  • $\begingroup$ @AndersSandberg thank you for your suggestion. That's a very interesting paper, and contains an elegant idea: to ensure validity via hashing, signature, and inclusion of hashes of previous documents in time stamping new ones. Cursory reading suggests that it still does rely on a clock for the absolute time stamp, meaning that if a group of people wanted to time stamp a set of documents to 1/1/2042, they could. My question needs this to be not possible. It is also not clear whether validation can occur if some of the records are missing or destroyed? Thank you for that excellent source. $\endgroup$ – Swapnil Bhatia Aug 7 '17 at 3:26
  • $\begingroup$ If you have a good source of unpredictable but public and documentable bits (say, a randomly fluctuating remote star) you can use it to date information. If the information includes the same bits as were generated at time T, then we know it cannot have been made before T. $\endgroup$ – Anders Sandberg Aug 7 '17 at 8:26
  • $\begingroup$ @AndersSandberg true. But a malicious group can also back date. These were indeed some of the considerations in my requiring third party-free physical mechanism. Nevertheless, your source and your ideas are intriguing and I am glad we are able to talk. Thank you! $\endgroup$ – Swapnil Bhatia Aug 8 '17 at 15:08

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.