We know that with or without data content analysis of actual email messages, lots of information can be inferred through various forms of metadata collection.  Given this reality the question becomes, what can be done?

One strategy might be to consider the adoption of a time delayed email system.  The reason why the use of such a mechanism to allow someone the ability to write an email, and then have it sent off at a specified (or randomly generated unspecified) date is useful for multiple reasons.

If a program could be coded in a way which could delay the actual transmission of data in such a manner that the original time of creation was adequately masked, it could hamper nefarious metadata collection of time-stamping and possibly geo-locating of user information.

For example:

A person walks into a computer café at 10am and sets an email to send at around 4am the following morning (via cloud or machine).  The data gets sent at 4am when the building is empty.  No CCTV photos/cameras are able to document who sent the message because the computer café is unoccupied at 4am.  The software is designed in such a fashion that it is nearly impossible to unmask the original time when the message was instructed to send, thus hiding the metadata associated with time-stamping, and thus hiding the true identity of the computer user at the café at 10am the previous day.

Any and all feed back regarding this idea is welcome.

--
Tuan Nghia DUONG

Élève-ingénieur en Informatique et Réseaux

ESISAR, Valence