potiuk commented on PR #2:
URL: https://github.com/apache/comdev/pull/2#issuecomment-4318843481

   > My point is that PII may appear in private lists, and should not be 
distributed beyond them.
   
   > Unless you can somehow ensure that any messages with PII are somehow 
excluded from the analysis, I don't see how that is not a violation of Privacy.
   
   Not excluded from analysis - but in some cases (like security@ mailing list) 
- there is no PII to care about. 
   
   I spend about half an hour discussing it with Christian yesterday, I 
understand that you might have that opinion - but not's that Christian said - 
when it comes to `security@` and similar mailing list. I really wanted to 
understand how it works and why things are defined like that - rather than base 
it on my own - or other people who have no authority or any role related to 
priavacy - necessary limited and narrow knowledge.
   
   The point of Christian (not mine - so you would have to discuss it with him 
if you disagree) in our discussion was that security@ are a "work" focused 
mailing lists - where sole purpose of those is to work on "disclosing security 
features" 
   
   Those kind of communications (especially when they are opt-in - by the fact 
of deciding to send the report - do not have the same PII expectations as 
regular "discussions". 
   
   But you can talk about it to Christian - feel absolutely free. He is going 
to work on a new policies - he promised to share it with me when they are 
drafted so that I can have a saying, I am sure if you talk to him he will also 
share it with you - and quite for sure there will be ASF-wide discussion about 
it 
   
   As  side comment. I am extremely happy about it - this was a fantastic news 
for the ASF  that the whole heated privacy discussion will lead to a policy 
that will be well defined and not prone to  sometimes not grounded in true 
knowledge about it - statements that are "definitive" and "obvious" (a side 
comment I find quite interesting how one can think they know better about 
something than real experts - that never ceases to amase me). 
   
   Also - just to clarify it even further - in the security team we are already 
planning - once those policies are clalrified - to make it even more explicit - 
including letting know the reporters that their reports WILL be made public 
eventually  after fixing the issues 
https://github.com/apache/security-site/pull/54  - this is absolutely necessary 
move now in the world where we MUST share information about both - identities 
of the reporters and content of their reports - to protect from AI generate 
flood of the issues. This has been already legal (as Christian explained) - but 
we want to even make it explicit. 
   
   I even told Christian about Ponymail - and he asked me whether we can 
protect access selectively for that purpose and it's then when the idea where 
ponymail MCP donated to ASF could be a model example where this is **actually 
done** and he loved that idea. So my PR to Rich's MCP was a direct result of me 
agreeing with Christian that we should do it. 
   
   Again - you can talk to Christian and check all that I told you - if you 
have any doubts if what I am writing above is wrong - it's not my 
interpretation this time - it's what Christian is going to propose in the new 
policies.
   
   
   My point is that PII may appear in private lists, and should not be 
distributed beyond them.
   
   Unless you can somehow ensure that any messages with PII are somehow 
excluded from the analysis, I don't see how that is not a violation of Privacy.
   
   > It would help to link to it.
   
   As Rich helpfully stated above - the PR of mine is already merged and 
present in this PR (see the last commit). I am also going to add a fixup to do 
"privacy by default" - i.e. make sure that all private lists are excluded by 
default - currently I excluded only foundation-wide, but I think better 
approach will be to exclude all the private lists **AND** security - and you 
will be able to opt-in to enable particular security lists as opt-in when you 
will configure the Ponymail MCP locally.
   
   There is a bit of heuristics to it for now - because not all private lists 
in the ASF have `private` in the name. But maybe there is a way ? Maybe your 
energy and knowledge @sebbASF could be used to help us to find out an "Exact" 
way of knowing that. Is there an LDAP or other place we can get the information 
whether the list is private or no? 
   
   How should we do it @sebbASF ? Can you help us ?
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to