Re: [Mailman-Developers] full anonymisation approach
![](https://secure.gravatar.com/avatar/fac1df10bf2267bb88ef12a01da82791.jpg?s=120&d=mm&r=g)
Stephen J. Turnbull writes:
It's really not possible. You have proposed an implementation but not addressed the requirements issue. In software engineering "correct design" means "the implementation will satisfy the requirements", but I still don't really understand what problem you are trying to solve, let alone whether it's really important.
Motivation of an anonymous list: Need of such a list is to focus on discussing facts and ideas irrespective of a person's identity. As identities always come in between personal views. For eg, social stigma may cause a person to accept an idea he/she dosen't feel for or may not express what he/she feels is right.
In my opinion the requirements are:
- Keeping list members anonymous from each other.
- Not allowing any way out for revealing any one's identity except if the user chooses to reveal it himself.
- List admin knows everything going in the backstage, like who is who.
I have tried to analyze some possible use cases here http://www.mail-archive.com/mailman-developers%40python.org/msg15189.html
Regards Rashi
![](https://secure.gravatar.com/avatar/334b870d5b26878a79b2dc4cfcc500bc.jpg?s=120&d=mm&r=g)
Rashi Karanpuria writes:
Motivation of an anonymous list: Need of such a list is to focus on discussing facts and ideas irrespective of a person's identity. As identities always come in between personal views. For eg, social stigma may cause a person to accept an idea he/she dosen't feel for or may not express what he/she feels is right.
Aside: I think if you actually use such a list you'll discover that most "shy" people are far more afraid of being flamed than they are of social stigma *outside* of the discussion thread. OTOH, it is likely to bring out the worst in trolls. IMHO YMMV etc.
Leaving my personal opinions out of it, let's see:
In my opinion the requirements are:
- Keeping list members anonymous from each other.
OK, that follows from the above motivation. However, do you want the list identity to be consistent within threads, or will you assign a new list identity to a poster address for each post? Do you want list identity to be consistent across threads, or will you assign a new list identity to a poster address each time they first post to a new thread? If identity is consistent within threads but not across threads, how do you determine that two posts are in different threads? For example, if post B points to post A in the References field, but the subject was changed in a significant way (ie, other than listrivia prefixes, Re: and friends, and whitespace changes that occasionally creep in due to mishandling of folding whitespace), are they in different threads?
- Not allowing any way out for revealing any one's identity except if the user chooses to reveal it himself.
That's not a reasonable requirement. It is often possible to deduce an author's identity from their choice of argument or writing style, or their employer's automatically added footer. You could argue that message content is a user choice, but I don't think many users would be happy with that point of view.
I think you want to limit yourself to stuff that MUAs and MTAs put in the headers, including but not limited to originator addresses, display names, and comments. Speaking of "originator", what do you propose to do about non-subscribers who are CC'd?
You might also propose removing standard signatures (ie, everything after a string matching "\n-- \n").
Removing other footers is much more chancy, but you could try to catch them.
- List admin knows everything going in the backstage, like who is who.
This is reasonable for many use cases (eg, a psychotherapy group, with the list admin being the therapist or her administrative staff -- they have access to billing etc, after all, they are implicitly trusted), but not for some very similar ones (eg, a 12-step support group: they have mentors, but not a special authority).
I have tried to analyze some possible use cases here http://www.mail-archive.com/mailman-developers%40python.org/msg15189.html
I'm not sure that requirements 1) - 3) are a good fit for several of your use cases, but these requirements seem like a reasonable start.
Regards,
participants (2)
-
Rashi Karanpuria
-
Stephen J. Turnbull