gistfile1.md

PS

While I was in the process of editing this article my good friend and co-founder Max Ogden told me he had just been asked to take down http://archive.couchdb.org, an IRC log and search utility he wrote for the IRC and Mailing List messages of the CouchDB project. [strike]It's unclear which esoteric Apache rule this tool violates, other than being useful and not being hosting at Apache, but[strike] he was told it's a "privacy violation" which is odd considering it only logs public IRC and Mailing List messages.

[Update] It has been clarified that the request to take down archive.couchdb.org was made by an ASF committer, possibly with the support of other #couchdb users, but is not a violation of ASF rules and the request was made on their own behalf and not the ASF. The only policy that has been metioned is the freenode guidlines which caution against logging but do not ban it outright.

I am happy to clarify that the person who requested this was me. The log archive that Max put up went back about a year. Anyone who chatted in the channel during that time was unaware that it was being logged, and so it is a violation of Freenode's guidelines, and indeed common decency, to publish these logs. As the official Group Contact for the channel, it is my duty to uphold the channel to the network's policies. I am personally responsible for it.

While the channel is public, it is not published. It's the difference between doing a podcast, and chatting with your friends at the pub. Both are fine, and both are essentially public, but they come with different assumptions about privacy. My request was that we remove the logs until we, as a group of IRC users, can decide what we want. If we decide we want to log the channel, we can put the site back up again, but without any retroactive data. We must also put a permanent notice in the channel topic to warn users that the channel is publicly logged.

Please note, again, that the IRC channel is run by me, personally. In turn, I delegate some of that authority to committers, using channel flags. While it is obviously the de-facto CouchDB channel, it is not run by Apache, and they have no say over how we operate it. Apache's only concern is that we do not make important project decisions on IRC where there is no visibility, and no permanent record. The mailing lists have to be used for that.

uhm, first of all, I don't agree with the idea that we start censoring public information (if anybody can join your channel, then that is a public channel), but put that aside.
Your pub example is a bit flawed: at a pub you usually have your own table, where you can adjust your volume, that others can't hear it.
I as a stranger can't just walk in, and sit to your table, and start listening to your conversation.
If I would, then probably you would ask me to leave, or you would leave, or you wouldn't discuss sensitive topics.

On a channel, everybody can "hear" everybody else, and having your channel open to join means that it is publically available to anybody.
I find it hard to believe that there are still people out there, who think that one can censor the internet. If you put something up, there is a good chance, that you can never remove it anymore.
We have newsgroup archives, mailing list archives now, it is natural that we also have irc archives, (and webcrawlers which cache and re-publish content, and web.archive.org, etc.).

AFAIK group contacts aren't "stakeholders" of the channel, but as the name suggests, they are contacts, if you want to poke freenode for support, your groups contact should take care of it, and when Freenode wants to contact you, guess what, they will poke the group contact.

So I can't see how should you be responsible for other people mistakes regarding your channel. So even if accept that logging is against Freenodes policy(which technically isn't, at it is only mentioned in the guidelines, and the following the guidelines is strongly adviced but not mandatory by the policy) I don't think that you could be personally responsible in any way,

Uhm, the irc channel is run by Freenode, not by you, I'm guessing that you ment that you are the channel operator/group contact, which totally makes sense, because the Group Contact should be somebody from the high-up from the group itself, so the an Apache representative shouldn't qualify for that.

btw. I would like to point out, that there are also existing archives for your channel out there:
http://irclogger.com/.couchdb/2010-11-27

Thank you for your comments.

I don't agree with the idea that we start censoring public information

This isn't censorship. If I send you a private email and ask you not to publish it, that is not censorship. If I prevent you from publishing your own emails, that is censorship. I am speaking on behalf of the channel when I ask other people to remove private IRC logs. We could publish the old logs, but we would need to get every person who's spoke in the channel in the last year to consent to them, or else redact those lines.

if anybody can join your channel, then that is a public channel

Agreed. But just because it is a public channel does not mean that it is publicly archived. Whether or not a channel is "on the record" or "off the record" is another matter entirely. Freenode advises that channels notify users that a channel has public logs. We have never done that, so it is unreasonable to assume that every person who spoke in the channel is okay with the logs being made public.

I as a stranger can't just walk in, and sit to your table, and start listening to your conversation.

Yes, I guess my example is flawed.

In real life, you can be in a public situation, and a nosey person can come to sit next to you and start recording your conversation, or taking notes. It's flawed because in real life, this behaviour is very obvious, and allows you to compensate for it.

On IRC, where any one of the hundreds of nicks could be a logging bot, you have to rely on network convention and trust. In our case, we have never advertised that we archive the logs and publish them, and so doing so at this point would break that bond of trust.

having your channel open to join means that it is publically available to anybody

Yes, but that doesn't mean that everything you say there has to be permanently on the public record.

I find it hard to believe that there are still people out there, who think that one can censor the internet

This isn't censorship, as I explained above.

We have newsgroup archives, mailing list archives now, it is natural that we also have irc archives, (and webcrawlers which cache and re-publish content, and web.archive.org, etc.).

We also have private mailing lists, opt-outs for Google's cache, opt-outs for the Internet Archive, and unlogged IRC channels.

So I can't see how should you be responsible for other people mistakes regarding your channel.

I'm not, but it's my channel, so I'm going to do my best to enforce our policies.

So even if accept that logging is against Freenodes policy

I never said that.

it is only mentioned in the guidelines, and the following the guidelines is strongly adviced but not mandatory by the policy

I am following them, because I think they make a lot of sense. I never said I had to, against my better judgement.

I don't think that you could be personally responsible in any way

I am not responsible in the sense that Freenode are going to get angry with me. I am responsible in the sense that it is my channel, and I want to run it a certain way. I was stressing this part to distance the policy, and my actions, from the Apache Software Foundation. In Mikeal's original post, he suggests that the request was made because the public archive violated some mysterious Apache policy, which is false.

Uhm, the irc channel is run by Freenode, not by you

Technicalities.

an Apache representative shouldn't qualify for that

Can you clarify?

I would like to point out, that there are also existing archives for your channel out there

I have emailed Chris Schneider with a take-down request.

"This isn't censorship. If I send you a private email and ask you not to publish it, that is not censorship."
But thats not the case here, because sending an email or private message would imply that the message is only intended to the explicitly stated recipients, imo irc messages sent to a public channel are under the "general publication", which would make them public domain:
"such a dissemination of the work of art itself among the public, as to justify the belief that it took place with the intention of rendering such work common property"

Of course there can be exceptions, if you copyright your speech before giving the public speech:
http://en.wikipedia.org/wiki/I_Have_a_Dream#Copyright_dispute

"This isn't censorship, as I explained above."
It all depends on the legal status of the irc messages, and it seems that we are disagreeing here.

"On IRC, where any one of the hundreds of nicks could be a logging bot, you have to rely on network convention and trust. In our case, we have never advertised that we archive the logs and publish them, and so doing so at this point would break that bond of trust."
I'm not sure that many people expect that nobody will log the conversations, and you can't guarantee that either (and as I mentioned, I don't think that you have the legal ground to force that).

"We also have private mailing lists, opt-outs for Google's cache, opt-outs for the Internet Archive, and unlogged IRC channels."
Depending on what you mean by "private" and "unlogged" it can mean that those aren't intended to the public, so they are legaly/technically covered.

"I'm not, but it's my channel, so I'm going to do my best to enforce our policies."
Ok, I don't mind if you set up your policies(even if those couldn't be enforced), but from your first message, it seemed that you are saying that those are the policies imposed by Freenode.

"I never said that."
.
"and so it is a violation of Freenode's guidelines, and indeed common decency, to publish these logs. As the official Group Contact for the channel, it is my duty to uphold the channel to the network's policies. I am personally responsible for it."

I think that pretty much implies that those are policies(violations of the guidelines vs upholding the network's policies).

"I am following them, because I think they make a lot of sense. I never said I had to, against my better judgement."
I think that it is a nice guideline, I just wanted to emphasize that it is only a guideline.

"I am not responsible in the sense that Freenode are going to get angry with me. I am responsible in the sense that it is my channel, and I want to run it a certain way. I was stressing this part to distance the policy, and my actions, from the Apache Software Foundation."
Uhm, I think that it isn't yours, but the property of the group, you are just the representative of the group for contacting Freenode.
http://freenode.net/policy.shtml#channelownership
"Channels on freenode are owned and operated by the groups which register them. No minimum level of activity or moderation is expected or required of channel owners. The network exists to further on-topic uses, as explained in this policy, and channels or groups may be removed from the network for activity which is considered to be off-topic."
http://freenode.net/group_registration.shtml // See the description of the Primary Contacts.

Of course you could have your own personal channel, but that couldn't be called #couchdb imo:
http://freenode.net/policy.shtml#channelnaming

"Technicalities."
Yeah, that can be nitpicking if the subject is irrelevant to the discussion, I just wanted to make sure that we are on the same page about that.

"Can you clarify?"
Same as above:
http://freenode.net/group_registration.shtml // See the description of the Primary Contacts.
http://freenode.net/policy.shtml#channelnaming
So an Apache representative couldn't register the group by itself, but if the project elect it, or for example he/she also happens to be the lead for the project that would qualify.
But I'm just reading the website here, I don't have any first-hand experience with the Freenode group registration workflow.

Most of your comments seem to be nitpicking vocabulary, which doesn't seem important.

I can, however, say with certainty that this has nothing to do with copyright. I don't think we'd need to get those people to agree to the logs being published because they have copyright claims as to what was said. Though, I expect they might. I think we'd need to get those people to agree to the logs being published because that is the polite thing to do, given their original expectations, and the current circumstances. This isn't a legal issue, and it certainly isn't a technological issue. It's a social issue. And this is why any discussion about the law, or about the technical aspects, misses the point.

"privacy violation", "While the channel is public, it is not published. It's the difference between doing a podcast, and chatting with your friends at the pub. Both are fine, and both are essentially public, but they come with different assumptions about privacy." "it is my duty to uphold the channel channel to the network's policies"

we just discussed the potentially privacy violation and the policies of freenode and your duties as a group contact, so I think those are relevant/important.

I'm happy that we are clarified that it is only a social issue, hope that you guys manage to fix that.

We're working on it.