On 28/03/2012 02:10 AM, Graydon Hoare wrote:
> Have I called you names so far? I'm sorry if so. Trying to remain civil.
Asa covered this. Thank you for apologizing; sorry for misunderstanding you!
> As I said before, I think non-exhaustiveness of textual policy is not a
> sufficient excuse for avoiding its construction or substantially
> curtailing obvious and predictable specifics it might address.
You gave some arguments against case-by-case judging at the end, I'll respond
there (my post was, upon rereading in the clear light of morning, somewhat
repetitive. Sorry!).
> I don't
> understand why [our CoC] would be different from other policies.
I'm unclear what you mean here, however: which other policies? Those of other
communities, or other policies we have? I presume the latter, in which case:
because it purely and only concerns human interactions and resulting emotions
(and ones orthogonal to our mission at that!), which I believe is much harder to
capture in rules and guidelines than say, data retention and privacy policies.
> Given that the root of this thread was Mitchell's decision to task Deb
> with writing (surveying, accumulating, editing) some text, I ... am
> confused about which part (if any) you're addressing. Do you want to
> have nothing written? Or merely cut it short?
The root of this part of the thread, however, was your idea to put more specific
terms in the CoC which Deb had drafted earlier. In particular, you actually
disagreed with Mitchell and said:
> I think
> that discussing and establishing some _specific_ behavioral norms is
> exactly the task at hand.
So I do not want to have nothing written, or even 'merely to cut it short'. I
would want it written without reference to specifics you mentioned, like
discrimination, harassment and so on. I'd rather cast the net wide and have the
committee be judicious in cases where it is invoked. I will elaborate at the
bottom, so we don't spread this argument across the entire message. :-)
>> Instead, I think we should leave whatever group of people (that ends up
>> deciding when problems arise) to make their own decisions.
>
> Really? You'd prefer we establish a group of people and give them,
> presumably, some kind of conflict-resolving authority, but have them
> exercise that with _no_ written guidelines? Or no specific ones, only
> vague ones?
Yes, only vague ones. Again, see the bottom. :-)
>> I would happily trust them with making those decisions;
>> much more than I would trust us to pre-emptively formulate an entire
>> para-legal process detailing what is and isn't OK.
>
> Ok. Nothing we do or say here has force of law -- really at worst it's
> probably just requests for apologies and/or requests to leave certain
> forums -- but I (personally) prefer some clarity-in-writing about
> standards. Written documents have a bit of dispassionate (not to mention
> time-saving) force in conflict resolution. It's why we have laws in the
> first place, not just judges.
This is an interesting remark, particularly as the recent debate saw calls for
resignations and banning/removing people from forums, rather than requests to
leave or separate content (those are distinct, at least in the way you
formulated it here; if they are not, I think your description of 'request to
leave' is a bit too euphemistic and in that sense a tad unfair).
The current document has no text about consequences. Out of interest, would you
like to specify those as well?
>> This list is not exhaustive, as I warned above.
>
> I believe it's not too long a list (7 qualifications?) to look at the
> elements individually. You can throw a few more in and we won't drown in
> complexity.
This is exactly what I said I was afraid of earlier in my message: you're
responding to the specifics, and then saying: look, it's not too bad!
The list being non-exhaustive, I don't think it is feasible to make it so.
To use a mathematical analogy: it's as if I said you couldn't enumerate all
primes, and you said "I don't believe that, and will only discuss individual
primes which you believe are missing or wrong". Now I gave you some primes,
notwithstanding that there are many more. Sadly, one can prove there is an
infinite number of primes, but because there are no logical connections between
different facets of human behaviour, I can't prove that the number of things to
take into account is 'large' (or 'too large'). :-(
> I don't think most of them should apply -- which medical
> issues have a bearing on harassment? -- but you can discuss that.
> They're worth considering if you are concerned that they should really
> influence policy.
Stress, insomnia, etc. I think there is often overlap between medical and
psychological issues, and attempted to be inclusive. This goes in the
'mitigating factor' box, as discussed elsewhere.
>> really see how you would intend to codify something like this.
>
> We've been looking at a number of documents that do just that. Indeed,
> you link into the middle of wikipedia's document-cluster on just this
> matter. Do you not understand how those bodies-of-policy work?
I do (at least, I think so!), I just disagree that ours should work the same
way. Wikipedia is a different type of community than we are, even if there are
also many similarities. By nature they are open to (semi-)anonymous
contributions, people can have multiple identities, ties between contributors
are much more loose, there is less credit, there is almost no employment and
therefore less of a mix of paid/unpaid contributions, etc. etc. So the stakes of
contributors are lower, and without more stringent policies, it is easier to
engage in destructive behaviour as people have less to lose.
>> As you asked for examples, here's a very concrete (albeit completely
>> fictitious) one:
>
> Great! Let me answer equally concretely on how I think such a policy --
> using the guidelines Deb has posted and the refinements I've proposed --
> might apply. In the spirit of creative fiction. Naturally if you
> disagree with any of the suggestions in anything Deb or I or anyone else
> might have written on this, you'll disagree with the snap judgments I'm
> making in response:
Just some notes here:
>> Does it
>> matter if it was about male or female circumcision (about which Western
>> cultures, as far as I can tell, hold different opinions (something which
>> I don't want to debate right now, fwiw!)).
>
> No, and I don't see how it could. In this putative blog post you didn't
> say anything about male circumcision, so nobody who identifies with the
> "cause" or culture of male circumcision has any words-of-yours to feel
> hurt by.
Ah, the brackets around 'female' were meant to annotate that it could be about
either (hypothetically!). The outcome may or may not be different.
>> Maybe I offended people of a
>> particular faith (even if I didn't mention religion). Does that make it
>> worse?
>
> "Offense" is a straw-man word that nobody has actually proposed we
> provide policy restraints on. I ask that you stop using it.
>
OK. The reason I used 'offense' is that from my perspective, cause for invoking
the CoC is found by the aggrieved, not the perpetrator, and this seemed like a
reasonable catch-all phrase. Put another way, I do not believe of anyone in this
community that they would purposefully harass, intimidate, ..., or discriminate
against others in it. Robert Kaiser also made this point: I trust the people in
this community. Without evidence to the contrary, I will presume hurt to be
accidental. That doesn't make it OK, nor does it mean that we can't discipline
people who do it repeatedly or disregard warnings about it, but it should effect
(sic!) how we deal with it.
> If you insulted, intimidated, discriminated-against (in the public
> proselytizing sense) that religion, then sure. That'd be problematic in
> the same way it's problematic to do so against citizens of a country.
Ah, but religion often has (or is argued to have) a stake in other parts of
life. As just one example from the US, the recent controversy about Catholic
hospitals and offering insurance policies that cover contraceptives.
(NB: let's not discuss about whether or not discrimination against Catholics did
happen there, but it's clear that some people felt this way, even though the
bill itself obviously didn't single out Catholic institutions specifically)
> And I think there's an additional dimension that's worth
> considering, that you left out: _what did you do in response_.
> Complaints are seldom about single actions, but sequences or patterns of
> action-response-reaction. If you react in good faith to their complaint
> ("oh sorry, I didn't realize this was part of your national identity,
> I'll re-tag that as personal and not involve the mozilla community with
> it") it might well represent a suitable resolution, and the problem
> would pass quickly. The hurtful reactions are more along the lines of
> denial of hurt, changing the topic to first amendment rights, or telling
> the hurt person to grow a thicker skin. Those reactions are often worse
> than the initial acts, and are what wind up escalating.
I think this is an extremely important point. It also hints at something else
which I think is important, namely that the CoC should be a kind of 'last
resort'. It would be preferable if people resolved their issues privately and/or
immediately, in the same forum, on a personal level, without invoking a third
party to mediate. In some cases that isn't possible, or the situation gets worse
(as you described), and then we can allow for mediation.
Regarding responding to a complaint: I also think we should allow room for
people to say "I did not (intend to) say what you're reading into it". This
happened in this thread, funnily enough: I and Asa thought you were calling Gerv
a troll, and you weren't. It should be OK to say what you said, even without the
apology: you didn't intend to insult anyone, you clarified what you meant, it is
clear in hindsight that this was the case, and that's it. It's honorable that
you apologized for it, I would hopefully do the same, but that's arguably a
personality thing. There are other people in our community who would argue that
it's not their fault that they were misinterpreted, and they have a point. I'd
rather not 'enforce' that they apologize, if they can plausibly point out that
there was miscommunication. But, again, yes, we should consider responses to
attempts to solve the matter on a personal level.
>> Again, I would prefer situations like this to be resolved on a
>> case-by-case basis. What do you think we gain by trying to pre-emptively
>> construct a moral framework of what is and isn't OK? What kind of leeway
>> will whoever gets to apply such a framework have in doing so?
>
> I don't think any of the yes/no/maybe responses I gave in the creative
> fiction exercise above are difficult to derive from the bits of
> proposed-policy we've been discussing so far. Nor from the bits of
> policy other communities use to set such standards. If they are, please
> point out where you see ambiguity.
I think people could disagree, given an actual blogpost and responses to
criticism, about whether such a post was discriminatory. If I had adjusted my
post and removed the advocacy / links to charity, but linked to the UN
describing the problem (which also campaigns against this, I believe? Writing
this without internet connectivity, so I can't check), it would (depending on
word choices etc.) be (even) more borderline.
More generally, I think the snap judgments you made are not difficult to derive
without your proposed amendments. However, I also think that people could derive
otherwise (with or without your amendments). Some people in the community will
agree with you, some won't. This certainly showed in the previous incident with
Gerv's blogpost. There was disagreement, sometimes very vehement disagreement. I
do not think a CoC *by itself* should be designed to resolve this disagreement.
I fully expect to need a group of real people to look at problems and decide
what to do. Perhaps this is my Dutch background, see
http://en.wikipedia.org/wiki/Poldermodel .
> As for leeway [for the adjudicators of the CoC]... well, that's part of the topic at hand. I'm arguing
> for narrower leeway (more specificity) than the thing Deb initially
> posted. You seem to be arguing for wider?
Yes.
>> I am just too cautious. Let me attempt to be more blunt: I think
>> including these terms is a mistake because they are inherently vague and
>> unclear, and would prefer a CoC without them. Here are some arguments in
>> link form:
>>
>>
http://www.cs.tut.fi/~jkorpela/wiio.html
>>
http://en.wikipedia.org/wiki/Wikipedia:HA#NOT
>
> The former is good humor, but if you believe it in earnest I don't know
> why you're even participating in this discussion. By its measure, all
> text is hopeless.
All good humor has a basis in reality, otherwise it wouldn't be funny.
Of course I do not literally believe all communication is pointless. I do
however believe that all (human) communication is lossy (see also Shannon's
model of communication), and that as a consequence, claims that communication is
open to one interpretation only are misguided; an example was present in this
thread (re: trolling). I think the CoC should respect this.
> The latter is a surprising link; I think ... it supports my view? It's
> the middle of a page going to some length to parametrize the definition
> of harassment, to preempt disagreement in case-by-case reasoning. I
> think it's quite helpful for it to be highlighting both what-is and
> what-is-not harassment in their eyes. It makes the line clear by
> examples and criteria on both sides.
The link actually specifically points to what is *not* harassment, which gives
some examples of policy abuses, which illustrates a few of the dangers of trying
to specify this kind of thing. Regardless of whether you attempt to clarify the
definition in the document, apparently you need to separately warn against the
abuses, and even then they still happen (or the section would have been removed,
presumably).
> Not the morality of all things in all contexts. Merely the morality of a
> specific list of behaviors, when enacted in specific contexts (places
> and times) that this community and organization owns and moderates.
> Morals in the sense of "community norms".
The context is often not 'just Mozilla', though, because of how open we are.
Blog posts which are syndicated to PMO aren't *just* read by PMO readers, and
when reading through PMO, people may read content with a different hat than just
"Oh, a post from some dude(tte) in the Mozilla community". Similarly, the
context from which people come is much more diverse, too. See eg. the recent
post from Gary Kovacs (IIRC?) who had to ask what webkit was, and the other post
by someone (I forget who) who had spent a lot of time at Microsoft and now had
to learn how different Mozilla was. All of this is part of a diversity which is
awesome about Mozilla, and I don't think that the context and so on is
constrained enough to provide all necessary and sufficient qualifiers to
separate OK behaviour from non-OK behaviour in all cases.
>> Again, I'm in favour of solving this by a social process - *on a case by
>> case basis* . This simplifies the decision as it will pertain to real
>> issues and real people with a real context, rather than hypotheticals.
>
> I'm opposed to "case-by-case" for a few reasons:
>
> - It gives no insight into what future cases might be. For people who
> are wavering on whether to stay or go -- whether they feel welcome
> or excluded -- it leaves the issue up in the air. For people who are
> wavering over the choice of whether to tag a post 'personal' or
> 'mozilla', it gives no guidelines about how to make that choice.
The first point surprises me. The original CoC which you argued was not explicit
enough gives explicit acknowledgement (in all proposed reformulations from Gerv
and others, too!) about how we welcome (or honour, respect, etc.) very diverse
individuals. People disagreed about exactly how to do this, but it seems clear
that clarifying our openness is a main goal. I don't dispute that; I do dispute
we need your proposed specificity in terms of harassment, discrimination etc.,
to do so.
For the second, I would expect (the result of) 'judgments' of such a group based
on a CoC to be somewhat public. This should offer some idea. More importantly,
again, I trust our community. Given the massive number of posts to planet, to
the newsgroups, mailing lists, conversations on IRC, conference calls, meetings
and so on, the number of incidents is actually surprisingly low.
In the reverse sense, having such guidelines beforehand also has a chilling
effect. Given the relative success of operating without a CoC altogether (by
which I do not mean to marginalize incidents, which are serious and should be
considered), I do not believe that a more proactive approach is warranted.
Instead, I think it is fine for the CoC to have a more reactive, "last resort"
(where the first one should, where possible, be direct engagement with the
problem), usecase.
> - It is costly. Written documents, once established, can be read and
> re-read and re-re-read at everyone's leisure without bothering the
> same few people to resolve all conflicts. Two people can look at the
> rules and peacefully reach an agreement without escalating. Rules
> scale better than judges.
This is true iff the rules help people reach an agreement without escalating,
and there is an overwhelming number of problems. I believe neither to be the
case: that is, I think that people are mostly able to reach agreements already,
and that where they are not, the CoC by itself would not help them - they would
simply end up with a second conflict about their interpretation of the CoC
(whether something qualifies as harassment, whether it was intentional, whether
it was 'a joke' and so on). In fact, they would most likely already end up with
these debates without a CoC, as one of the participants would make claims using
similar words, and the other would deny them (in the case where they don't
manage to resolve the issue, anyway). We in fact saw some of this in the recent
blogpost issue too (I don't want to rehash that problem or take a position on it
at all, but it seems clear there was disagreement among various people about
whether by itself, it was discriminatory and/or various other things).
> - It is less prone to favoritism and other personal failings. Humans
> asked to "adjudicate" disputes almost always come down on the side
> of their friends. I'm doing it here! I'm friends with Tim and this
> influences my willingness to support his claims. This sort of human
> factor is something textual policies help counteract. Text is more
> stable and dispassionate.
>
> (There are numerous other human judgment-making flaws that come out
> in such situations; the discriminatory biases of the people
> involved, the time of day, level of hunger, etc. etc.)
Right. And all those human flaws also affect the original problem. And they
affect us now as we are drafting the CoC, as you pointed out (and for full
disclosure, to put it euphemistically, I cannot claim to be unbiased in my
evaluation of Gerv, even if we don't always agree :-) ).
I do not believe shifting these problems into the writing of this text helps.
Rather, the problems are muddled further by hypothetical rather than real
issues, recent history, and the limited and unbalanced selection of participants.
Instead, I would argue that the group of people doing adjudication should be
balanced, to counter 'favoritism and other personal failings'. Pick people with
various characters, backgrounds, functions within the community, and experience
in this community (ie people who came here recently vs. people who are basically
part of the furniture ;-) ). Make it 5 or 7 people. It's not going to be
perfect, but that's something I expect we'll have to live with regardless.
Regarding your point that non-exhaustiveness shouldn't prevent specificity: I
see why you feel that way (I think), but I disagree. If the list is
non-exhaustive and omits factors, reasons or behaviours which should play a role
in the decision making process, it unfairly biases people against such things,
irrespective of their legitimacy ("that's not in the CoC!").
About those vague guidelines: yes, I think human interpretation and evaluation
is useful, necessary and even sufficient, in combination with a very basic
aspirational document. If we strive to respect other people, and it's clear that
(in an intentional way, I would say) no respect was shown, then there should be
consequences (TBD). Harassment, insults, discrimination and so on all violate
the 'be excellent to each other' and 'respect each other' (however we formulate
it) dicta.
Hope that clarifies.
Cheers,
Gijs