Stream: zulip

Topic: zulip-archive.rust-lang.org robots.txt?


view this post on Zulip Laurențiu (Mar 24 2021 at 13:24):

I know Zulip is relatively public, but zulip-archive.rust-lang.org is indexed by search engines, which does make it easier to dig through if you're doxxing someone. I think it would be reasonable to disable indexing.

view this post on Zulip Laurențiu (Mar 24 2021 at 13:26):

I mean, if I google for site:zulip-archive.rust-lang.org in a private window, the third result is a message I posted (might be different for others)

view this post on Zulip Mario Carneiro (Mar 25 2021 at 03:47):

Isn't the whole point of an archive so that it can be indexed by search engines? The doxxing concern doesn't make any sense to me, this is a public chat

view this post on Zulip Laurențiu (Mar 25 2021 at 06:55):

I thought the point of it was to be able to show conversations to people who don't want to log in

view this post on Zulip Laurențiu (Mar 25 2021 at 06:56):

The archive isn't a chat, Zulip is, and on Zulip you can delete your account, change your name, edit messages etc. You can't do that on the archive and it's not even covered by the privacy policy on rust-lang.org. When you register, you won't even know it exists.

view this post on Zulip Mario Carneiro (Mar 25 2021 at 07:34):

I don't think you can edit or delete even your own messages after a short time frame. I don't think deleting your account deletes your messages either. So I don't see a big difference between comments made here vs e.g. github issues

view this post on Zulip Mario Carneiro (Mar 25 2021 at 07:35):

AFAIK if you edit messages here they will eventually end up edited on the archive too - it's just a mirror for Zulip

view this post on Zulip Laurențiu (Mar 25 2021 at 07:35):

You can certainly delete comments on GitHub, and if you delete your account it shows up as "ghost" or something like that.

view this post on Zulip Laurențiu (Mar 25 2021 at 07:36):

Mario Carneiro said:

AFAIK if you edit messages here they will eventually end up edited on the archive too - it's just a mirror for Zulip

It doesn't seem to: https://zulip-archive.rust-lang.org/122653zulip/07243zuliparchiverustlangorgrobotstxt.html (I changed my display name yesterday)

view this post on Zulip Mario Carneiro (Mar 25 2021 at 07:37):

On the lean zulip (that I also use) there is no time limit on editing messages like there is here, so I guess that's something the admins set up

view this post on Zulip Mario Carneiro (Mar 25 2021 at 07:39):

The rust zulip archive is something run by rust-lang so I guess you could file an issue

view this post on Zulip Laurențiu (Mar 25 2021 at 07:39):

I sent an email to privacy@rust-lang.org yesterday, after reading through https://github.com/rust-lang/crates.io/issues/955

view this post on Zulip Mario Carneiro (Mar 25 2021 at 07:40):

But zulip search is kind of... not great, and that was one of the primary motivating reasons for setting up the lean zulip archive, so especially since this zulip instance gets so much traffic I think having an external searchable archive is a big value add

view this post on Zulip Laurențiu (Mar 25 2021 at 07:41):

Not respecting the GDPR is also kind of not great.

view this post on Zulip Mario Carneiro (Mar 25 2021 at 07:41):

Can you explain more precisely what you mean by that?

view this post on Zulip Mario Carneiro (Mar 25 2021 at 07:41):

Would it be sufficient if admins respond to individual requests to remove sensitive information?

view this post on Zulip Laurențiu (Mar 25 2021 at 07:42):

Mario Carneiro said:

Would it be sufficient if admins respond to individual requests to remove sensitive information?

Yes, AFAIK, if it's also documented in the privacy policy.

view this post on Zulip Laurențiu (Mar 25 2021 at 07:42):

But an request might also say "here is my GitHub username, please remove all my messages"

view this post on Zulip Mario Carneiro (Mar 25 2021 at 07:43):

As I understand it one of the motivations for preventing folks for modifying old messages is so that you can't remove evidence of abuse or something

view this post on Zulip Laurențiu (Mar 25 2021 at 07:44):

https://gdpr-info.eu/art-17-gdpr/

view this post on Zulip Mario Carneiro (Mar 25 2021 at 07:44):

But making all of the archive not searchable seems like way overkill for that

view this post on Zulip Laurențiu (Mar 25 2021 at 07:45):

That's why I agreed that removing individual messages would be sufficient

view this post on Zulip Mario Carneiro (Mar 25 2021 at 07:45):

Can you remove the messages on zulip? It might suffice to make sure that such deletions are mirrored correctly

view this post on Zulip Mario Carneiro (Mar 25 2021 at 07:46):

I guess admins can remove messages with no time limit

view this post on Zulip Laurențiu (Mar 25 2021 at 07:48):

My name change wasn't mirrored, so I'm not sure.

view this post on Zulip Laurențiu (Mar 25 2021 at 07:50):

https://github.com/rust-lang/zulip_archive/blob/master/archive.py

view this post on Zulip Mario Carneiro (Mar 25 2021 at 07:53):

I guess that's a bug in the incremental update; if the json cache stored user IDs and a mapping to names, instead of storing the already-rendered user names, then it would reflect the name change

view this post on Zulip Mario Carneiro (Mar 25 2021 at 07:54):

There is a full update option, so perhaps an admin should just clear and rebuild the cache, although it might be expensive to rebuild everything

view this post on Zulip Laurențiu (Mar 25 2021 at 07:54):

TBH, I'd like a "don't archive any of my messages" option, but I don't see that being implemented, so the only alternative would be to stop using Zulip. Not because I have a problem with Zulip itself, but because I don't like the public archive.

view this post on Zulip Mario Carneiro (Mar 25 2021 at 07:55):

You could just DM everyone

view this post on Zulip Laurențiu (Mar 25 2021 at 07:56):

That still wouldn't help with the older messages in the archive.

view this post on Zulip Mario Carneiro (Mar 25 2021 at 07:59):

removing all your messages from zulip and then rebuilding the archive would achieve that, although there isn't an easy way to build this into the incremental update unless it specifically recognizes such an operation, so it would have to be rebuilt


Last updated: Oct 11 2021 at 22:34 UTC