Timeline for How to be a co-author anonymously?
Current License: CC BY-SA 3.0
12 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Nov 1, 2015 at 10:06 | comment | added | David Richerby | This cryptographic approach is misguided. It is designed to prevent other people from falsely claiming to be the author, but it is being used for a subtly different purpose: to keep the actual author's identity secret. If the author's government gets suspicious ("Hey, somebody just published a widget theory paper making a big song and dance about anonymity because their country hates our great enemy.") they just have to hack the non-anonymous co-author's computer and find all the emails discussing the project. Don't forget the lesson of Snowden: government agencies bypass cryptosystems. | |
| Nov 1, 2015 at 8:56 | comment | added | kasperd | If you want to rely on the contents of the document remaining secret after a hash of the document has been published, then ensure the document itself has more entropy than the output hash. Since it is extremely difficult to estimate the exact entropy of the text itself, it is safer to assume it is zero and put some randomly generated data with known entropy inside the document. For a previous situation with similar needs, I have put a line of random non-alphanumerical ASCII characters at the top of a file and another one at the bottom. That adds 5 bits of entropy for each character. | |
| Nov 1, 2015 at 8:43 | comment | added | kasperd | The note you suggest to put in the paper seems a bit too specific to me. It would point out exactly which country the anonymous author is from, which might prompt officials in said country to start investigating. | |
| Oct 30, 2015 at 12:38 | comment | added | cnst | @DCTLib, but there's no need for that -- just use the same unique name in all papers; another alternative would be to just sha256 the name itself, but then it would be vulnerable to have the name guessed and independently confirmed by any really interested party. :-) | |
| Oct 30, 2015 at 12:15 | comment | added | DCTLib | @cnst The working relationship of the two authors may last longer than a single paper. By providing the same SHA hash on all papers, the readers would get a clue that the pseudonym is supposed to refer to the same author. | |
| Oct 30, 2015 at 8:10 | comment | added | cnst | I've improved upon this answer there: academia.stackexchange.com/a/57203/6026 | |
| Oct 30, 2015 at 7:23 | comment | added | cnst | Why bother with an explanation? Just provide a sha sum of a copy of the paper with proper credits. But keep in mind that finding collisions with binary files is easier than with text, so, it's best to make a checksum of a text version of the article. | |
| Oct 28, 2015 at 8:37 | comment | added | David Mulder | @FedericoPoloni Even if people might not be familiar with it, if it's about applying for a job I am pretty sure they could check with the IT department in no time if they have any doubts. | |
| Oct 27, 2015 at 22:30 | comment | added | Federico Poloni | If you wish to go the crypto way, get a signed statement from your co-author anyway. Some people, especially outside computer science, might not be familiar with this whole hash thing, and will probably insist in seeing a piece of paper. And get it now, just in case your co-author gets hit by a bus. | |
| Oct 27, 2015 at 20:50 | comment | added | David Mulder | Yeah, considered exactly this (with a string instead of document) as well after I had finished writing 90% of my answer, and then I was like 'in general the idea of people owning a private key is better understood then having a string with a public hash', but yeah, same concept in the end, just depends a bit on how much attention you wish to draw. | |
| Oct 27, 2015 at 10:39 | comment | added | Taladris | Related to the second point, do journals usually allow pseudonyms? | |
| Oct 27, 2015 at 9:43 | history | answered | DCTLib | CC BY-SA 3.0 |