Re: GR proposal: code of conduct

2014-02-26 Thread Wouter Verhelst
Hi,

Op maandag 24 februari 2014 08:47:57 schreef Alexander Wirt:
> Sorry for being late.

No worries -- we don't always have the time :)

> That morning I found the time to read the CoC in
> detail. In that mail I speak primary for myself and not all listmasters. But
> I collected some opinions from the others forehand, therefore I hope that
> what I write is in line with the other listmasters.

Thanks

> I am quite happy with the CoC as it is, I just have a few
> supplementary notes.
> 
> - the CoC, can only be an extension to our (lists.d.o) Coc [1], as there are
> missing the mail/list specific parts.

Hm. The whole point of this exercise was to replace that code of conduct with 
a more generic and up-to-date one, so if you feel that this isn't good enough, 
then that's a bug.

Can you be more specific about the bits that you think should not be removed 
from the current mailinglist coc?

> I am also not that happy with having
> several documents with the name 'Code of Conduct', maybe we can find a
> solution somehow.

Yes, that would seem to be obvious; I don't think we need several codes of 
conduct.

> - I always found the netiquette [2] a very useful source, maybe we can add a
> link to it to the document.

Good idea.

> - "The administrators will divulge any bans to all Debian Developers for
>   review". I know that this is the case for lists.d.o now, but I never saw
>   other anything from other services.

I have seen several such announcements from owner@bugs.d.o now, too.

>  Are _all_ other administrators of
>   'Debian communication forums' aware of that change? If we go that way, we
>   should probably move away from announcing them on -private and move to
>   something else. Like an mbox on master, or something else (and in my eyes
> - non-public).

I don't think it's necessary to move that.

While the code of conduct says that bans should be made public to Debian 
Developers, it does not say how, where, in what manner, or even if bans should 
be made public _only_ to Debian Developers (although we might be somewhat more 
explicit about that). This is intentional; I think review of bans is a good 
thing, and I do think we should have it, but I don't want a document like this 
to impose any workflow on anyone.

As such, personally I don't expect this to result in a major increase (other 
than has already happened) of such announcements to -private.

I could be mistaken, of course.

-- 
This end should point toward the ground if you want to go to space.

If it starts pointing toward space you are having a bad problem and you
will not go to space today.

  -- http://xkcd.com/1133/


signature.asc
Description: This is a digitally signed message part.


Re: GR proposal: code of conduct

2014-02-26 Thread Alexander Wirt
On Wed, 26 Feb 2014, Wouter Verhelst wrote:

Hi,

*snip*
> > - the CoC, can only be an extension to our (lists.d.o) Coc [1], as there are
> > missing the mail/list specific parts.
> 
> Hm. The whole point of this exercise was to replace that code of conduct with 
> a more generic and up-to-date one, so if you feel that this isn't good 
> enough, 
> then that's a bug.
> 
> Can you be more specific about the bits that you think should not be removed 
> from the current mailinglist coc?
Your goals are honorable, but I am not sure if this possible. Let me see:

I have some example that I don't want to lose, but most are for example not
suitable for IRC:

- Do not send spam; see the advertising policy below. (the  advertising
  policy is the interesting part)
- Send all of your e-mails in English. Only use other languages on mailing
  lists where that is explicitly allowed (e.g. French on debian-user-french).
- Make sure that you are using the proper list. In particular, don't send
  user-related questions to developer-related mailing lists.
- Wrap your lines at 80 characters or less for ordinary discussion. Lines
  longer than 80 characters are acceptable for computer-generated output (e.g.,
  ls -l).
- Do not send automated out-of-office or vacation messages.
- Do not send test messages to determine whether your mail client is working.
- Do not send subscription or unsubscription requests to the list address
  itself; use the respective -request address instead.
- Never send your messages in HTML; use plain text instead.
- Avoid sending large attachments.
- Do not quote messages that were sent to you by other people in private mail,
  unless agreed beforehand.
- When replying to messages on the mailing list, do not send a carbon copy (CC)
  to the original poster unless they explicitly request to be copied.
- If you want to complain to someone who sent you a carbon copy when you did
  not ask for it, do it privately.
- If you send messages to lists to which you are not subscribed, always note
  that fact in the body of your message.

This are a lot of points, and most of them don't fit to other mediums.

> > I am also not that happy with having
> > several documents with the name 'Code of Conduct', maybe we can find a
> > solution somehow.
> 
> Yes, that would seem to be obvious; I don't think we need several codes of 
> conduct.
I think that there are always medium specific rules that don't apply to other
medium. One classic point specific to IRC would be not to use an CTCP VERSION
to all clients.

*snip*

> >  Are _all_ other administrators of
> >   'Debian communication forums' aware of that change? If we go that way, we
> >   should probably move away from announcing them on -private and move to
> >   something else. Like an mbox on master, or something else (and in my eyes
> > - non-public).
> 
> I don't think it's necessary to move that.
> 
> While the code of conduct says that bans should be made public to Debian 
> Developers, it does not say how, where, in what manner, or even if bans 
> should 
> be made public _only_ to Debian Developers (although we might be somewhat 
> more 
> explicit about that). This is intentional; I think review of bans is a good 
> thing, and I do think we should have it, but I don't want a document like 
> this 
> to impose any workflow on anyone.
> 
> As such, personally I don't expect this to result in a major increase (other 
> than has already happened) of such announcements to -private.
OK


Alex


pgpb3zqlJ52jx.pgp
Description: PGP signature


Re: GR proposal: code of conduct

2014-02-26 Thread Wouter Verhelst
Op woensdag 26 februari 2014 15:25:25 schreef u:
> On Wed, 26 Feb 2014, Wouter Verhelst wrote:
> 
> Hi,
> 
> *snip*
> 
> > > - the CoC, can only be an extension to our (lists.d.o) Coc [1], as there
> > > are missing the mail/list specific parts.
> > 
> > Hm. The whole point of this exercise was to replace that code of conduct
> > with a more generic and up-to-date one, so if you feel that this isn't
> > good enough, then that's a bug.
> > 
> > Can you be more specific about the bits that you think should not be
> > removed from the current mailinglist coc?
> 
> Your goals are honorable, but I am not sure if this possible. Let me see:
> 
> I have some example that I don't want to lose, but most are for example not
> suitable for IRC:
> 
> - Do not send spam; see the advertising policy below. (the  advertising
>   policy is the interesting part)

Right, that one.

I'm not sure this belongs in a code of conduct, for the same reason that we 
shouldn't publish bans for trolls or spammers. A code of conduct should be 
about conduct, i.e., social behaviour, not about "don't be a pest".

That doesn't mean we should not have a "do not spam" policy, nor that we 
cannot publish such a policy; just that I don't think it should be part of a 
code of _conduct_.

In addition, personally I am not convinced that this part of the current code 
of conduct is very efficient in fighting spam, but then I am not in your 
shoes. Do you believe otherwise? If so, can you clarify?

> - Send all of your e-mails in English. Only use other languages on mailing
>   lists where that is explicitly allowed (e.g. French on
> debian-user-french).

A clause like

"Please use the appropriate language for the medium you are using. In Debian, 
this is usually English, but there are exceptions (e.g. use French on the 
debian-user-french mailinglist, or Dutch on the #debian-nl IRC channel)."

could work.

Having said that, I should note that my very first draft[1] did still contain 
this clause (or a similar one, at least); I'm not sure anymore why it was 
removed.

[1] https://lists.debian.org/debian-project/2013/05/msg00060.html

> - Make sure that you are using the proper list. In
> particular, don't send user-related questions to developer-related mailing
> lists.

"Some of our communication channels have topic-specific subdivisions; please 
use the appropriate one for your topic", possibly with an example?

> - Wrap your lines at 80 characters or less for ordinary discussion. Lines
>   longer than 80 characters are acceptable for computer-generated output
> (e.g., ls -l).
> - Do not send automated out-of-office or vacation messages.
> - Do not send test messages to determine whether your mail client is
> working.
> - Do not send subscription or unsubscription requests to the list
> address itself; use the respective -request address instead.
> - Never send your messages in HTML; use plain text instead.
> - Avoid sending large attachments.

While I agree that these are useful suggestions (and that therefore they 
probably should be retained), these sound more like technical guidelines; I 
don't think a code of _conduct_ should contain technical explanations on how 
to configure your mail client.

So I would suggest that for these things, we create something else (not a code 
of conduct) that is maintained by you, our listmasters. The (proposed) code of 
conduct could obviously refer to it from the "further reading" section, if 
that seems appropriate.

Does that make sense?

Additionally, the bits about "large attachments" and "HTML" sound like things 
that could more easily be done by a filter. If we don't want large 
attachments, we should make it technically impossible for people to send them 
(while making sure that those who try will get an informative bounce message).

> - Do not quote messages that were sent to you by other people in private
> mail, unless agreed beforehand.

I believe such a clause was originally part of the "Be open" item in my draft, 
but it got edited out. We could add it back, of course...

> - When replying to messages on the mailing list, do not send a carbon copy
> (CC) to the original poster unless they explicitly request to be copied.

Well, heh.

On that one, I think the current code of conduct is a mistake, because most 
mail clients make it very hard to do that.

Yes, some mail clients do have a "list reply" option, but some will only send 
the reply to the mailinglist on which the person replying received the mail in 
question; any cross-posted mailinglists will be dropped, which is not always 
the right thing to do.

Yes, one can edit the list of recipients and remove non-list recipients, but 
then those recipients who explicitly asked to be Cc'd somewhere up the thread 
will not receive those requested copies.

I think we should default to what tools make easy, not to the option which 
requires manual work.

I understand that this is the current policy, and if there is a strong feeling 
that we should re

Re: GR proposal: code of conduct

2014-02-26 Thread Paul Wise
On Wed, Feb 26, 2014 at 10:25 PM, Alexander Wirt wrote:

> - Do not send spam; see the advertising policy below. (the  advertising
>   policy is the interesting part)
> - Send all of your e-mails in English. Only use other languages on mailing
>   lists where that is explicitly allowed (e.g. French on debian-user-french).
> - Make sure that you are using the proper list. In particular, don't send
>   user-related questions to developer-related mailing lists.

Reworded to use generic terminology these would be the same on IRC.

> - Wrap your lines at 80 characters or less for ordinary discussion. Lines
>   longer than 80 characters are acceptable for computer-generated output 
> (e.g.,
>   ls -l).

IRC equivalent would be to use pastebins.

> - Do not send automated out-of-office or vacation messages.
> - Do not send test messages to determine whether your mail client is working.

Reworded to use generic terminology these would be the same on IRC.

> - Do not send subscription or unsubscription requests to the list address
>   itself; use the respective -request address instead.
> - Never send your messages in HTML; use plain text instead.

Seem to be lists-specific.

> - Avoid sending large attachments.

IRC equivalent would be to use pastebins.

> - Do not quote messages that were sent to you by other people in private mail,
>   unless agreed beforehand.

Reworded to use generic terminology this would be the same on IRC.

> - When replying to messages on the mailing list, do not send a carbon copy 
> (CC)
>   to the original poster unless they explicitly request to be copied.
> - If you want to complain to someone who sent you a carbon copy when you did
>   not ask for it, do it privately.
> - If you send messages to lists to which you are not subscribed, always note
>   that fact in the body of your message.

Seem to be lists-specific.

> I think that there are always medium specific rules that don't apply to other
> medium. One classic point specific to IRC would be not to use an CTCP VERSION
> to all clients.

Indeed, but many rules are also applicable to other forms of
communication if worded generically.

-- 
bye,
pabs

http://wiki.debian.org/PaulWise


-- 
To UNSUBSCRIBE, email to debian-vote-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: 
https://lists.debian.org/caktje6hktotyd4emukd1ssakvch9frhpwd+ntdm5teyudpp...@mail.gmail.com



Re: GR proposal: code of conduct

2014-02-26 Thread Paul Wise
On Wed, Feb 26, 2014 at 11:41 PM, Wouter Verhelst wrote:
> Op woensdag 26 februari 2014 15:25:25 schreef Alexander Wirt:
>> - When replying to messages on the mailing list, do not send a carbon copy
>> (CC) to the original poster unless they explicitly request to be copied.
>
> Well, heh.
...
> I think we should default to what tools make easy, not to the option which
> requires manual work.

The tools will never be able to do the right thing because the right
thing is highly context dependent. On Debian lists it is convention to
not CC unless there is some indicator that you should. On LKML not
CCing widely is discouraged. Each community has their own rules and
there are too many tools for all of them to ever be able to know about
these differences.

> I understand that this is the current policy, and if there is a strong feeling
> that we should retain it, I won't oppose keeping it. But my personal opinion
> is that it should go.

I'd prefer that it didn't change. If it did I'd have to figure out how
to use nore...@debian.org in From, unsubscribe from the lists or get a
completely new client and mail setup.

-- 
bye,
pabs

http://wiki.debian.org/PaulWise


-- 
To UNSUBSCRIBE, email to debian-vote-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: 
https://lists.debian.org/CAKTje6GpE=pTt5cuFEKV=vys+nbj8hifosrhhpfgbsgqkfs...@mail.gmail.com