[It took me so long to write this that responses from Russ and Steve L.
have since come through. I find myself in concord with both messages.]
At 2023-08-21T19:02:27+0000, Andrew M.A. Cater wrote:
> On Mon, Aug 21, 2023 at 05:32:22PM +0000, Jeremy Stanley wrote:
> > On 2023-08-21 20:16:22 +0300 (+0300), Dmitry Baryshkov wrote:
> > [...]
> > > According to Debian's CoC we use non-offensive ways to communicate
> > > within the project, so that everybody is welcome to speak and
> > > contribute. But we should not censor software. If there is a
> > > misogynistic comment in GNU HURD sources, should we censor it out?
> >
> > For that matter, if Debian was going to get into book burning over
> > racist, homophobic and misogynistic writing, all those packaged
> > versions of religious texts would presumably be the first things
> > tossed onto the pyre.
>
> OK. With respect to Branden, Sam, Rodrigo
His given name is Roberto, FWIW.
> Sanchez and Wouter - this isn't *just* a free speech matter and Debian
> isn't particularly censoring content.
I find your choice of adverbs revealing.
Okay, it _is_ a free speech issue, commingled with one or more other
things. I can entertain that perspective.
The other one's weirder. Debian isn't _particularly_ censoring
content--it's doing it...generally? A blanket ban of everything in
the fortunes-off package, including all future contents of the package
no matter how curated, and all such content incorporated into some
_other_ fortunes package would certainly meet the criterion of
"general". And would implicate the free speech element more strongly, I
would think.
> That being said: In some sense, the Code of Conduct governs how we
> behave with respect to the outside world
I regret to contradict you, but its stated scope is a poor fit with
your characterization.
"The Debian Project, the producers of the Debian system, have adopted a
code of conduct for participants to its mailinglists, IRC channels and
other modes of communication within the project."[1]
Last three words. Within. The. Project.
I wouldn't say that a participant in Debian should feel thereby
authorized to go out and be a jerkass[2] to people who have no direct
connection with our project, but if you're going to wield the CoC as an
instrument of control, I think it prudent to attend closely to its
stated terms and domain of application.
> and definitely colours how we appear there to Debian outsiders.
Sure. The CoC colors the impression the project makes to non-members.
> We have a Code of Conduct and folk expect us to follow it.
Being aware of it and attending to what it actually says don't seem to
be strongly coupled.
I'm going to rearrange your sentences a little bit here.
> In this instance:
>
> fortune, fortunes-off and so on: it's a GAME. Fortune as a *thing*
> existed before the BSDs but it became widely adopted with Unix v6 and
> then BSD.
In Unix V6 and every edition of the manual back to V1, section 6 of the
manual was known as "user-maintained programs". It had games, sure, but
also "amusements" like azel(6), which "predict[ed], in convenient form,
the apparent trajectories of Earth satellites whose orbital elements are
given in the argument files." V6 had azel but V7 apparently lost it,
possibly due to the unfortunate and untimely death of its credited
author, Joseph Ossanna, who wrote other programs much closer to my
expertise.[3] This concept of "user-maintained programs" is one I
propose to apply to the present situation. (I had thought this was
obvious from the course of action I proposed for myself in my earlier
message,[4] but people were perhaps stunned insensate before reading
that far.)
> It's not a core package.
Since when is _this_ a formal Debian concept?
> Fortunes-off is a leaf package of a small package.
Where is this terminology coming from? What's a "leaf package"? We
already have the terms "binary package" and "source package". And how
much does package size matter? According to dpkg on my machine:
Installed-Size: 1874
How many standard deviations below the mean is this? How many does it
need to be to be "small"?
If these questions aren't worth answering, your point wasn't worth
making.
> FreeBSD - our "upstream" apparently abandoned all fortunes apart from
> those relating to system administration in 2017 - because of
> complaints about Hitler quotes.
Why would taking out the Hitler quotes, or reviewing them for
acceptability, have been undesirable alternative responses? It seems
likely to me that FreeBSD discarded the package because they just didn't
want to deal with it. And that's fine, but it should not be mistaken
for a serious-minded content review.
> Ubuntu - our "downstream" has abandoned fortunes-off as incompatible
> with their Code of Conduct (which is strikingly similar to ours).
I read the relevant ticket.[5] The complaint was as follows.
"I've discovered that currently, as of Ubuntu 22.10, Ubuntu distributes
a package called "fortunes-off" that is full of homophobia, Hitler
quotes, virulent misogyny, racism, and more"
When we went around the block on this issue last year, the first 5
"offensive" fortunes I asked for, weren't. One I did find to be of low
quality: it wasn't homophobic, misogynistic, racist, or a quotation of
Adolf Hitler.[6]
I admit, however, that I can't deny that it fell into the category of
"more". But anyone familiar with elementary set theory appreciates that
the original complaint managed to incorporate the entire universe of
discourse.
(As an exercise, with minimal editing, recast the Ubuntu user's
complaint to manifest Russell's Paradox.)
Here is the entirety of Steve Langasek's reason for expelling the
package from Ubuntu.
'I am not interested in reading through this garbage to determine which
of the "offensive" entries fall just shy of some invisible line to be
worthy of continued inclusion. I have reviewed the contents of the
database sufficiently to establish the truth of the above charge and
will spend no further time parsing it. Effective immediately I am
removing the fortunes-off binary package from fortune-mod in Ubuntu.'
In a research paper this is like skipping from the Introduction to the
Conclusion with no Method or Findings. And that's fine for Ubuntu, I
guess. Possibly even necessary: while their Code of Conduct may
resemble ours (I didn't look), they have a SABDFL whose domain of
authority is anything he wants it to be. I understand someone taking a
cautious approach because it's not worth it to them to risk defending
their decisions to someone who signs their paycheck over a matter that
they consider a waste of their time in the first place. Not enough
hours in the day for that mess.
But Debian is different. We don't have a SABDFL, but a Constitution.
I concede that many Ubuntu users (and Red Hat ones, and others) consider
this a serious defect in our organization.
Nevertheless my understanding is that the legacy of the Debian
Constitution and its democratic procedures, largely handed down to us by
Ian Jackson from Mount Cambridge after he led us out of bondage under
the Pharaoh Perens, is one that Debian developers, maintainers, and
users continue to value. Those who don't, tend to go elsewhere.
> We had complaints in November
Well, one, anyway, to which the response was, I submit, an overreaction,
applying some kind of intellectual one-drop rule.[7] If there's some
pro-fascist BS in the package I'd be more than happy to rip it out or
couple it with a savage takedown by Noam Chomsky or other personage.
> and then a reminder in this thread.
Well, yes, but the package was already removed without replacement from
unstable months before the trixie release, so it would not be fair to
suggest that nothing was done about the perceived problem.
The problem, or question, now, is whether the package will ever be able
to come back, and on what terms.
> The US has guaranteed freedom of speech within the US:
As someone who remembers the September 11th attacks and the dark period
after it, with White House Press Secretary warning the journalists in
attendance and thereby the entire body politic to "watch what they
say"[8], I'd caution you against over-interpreting this principle.
> other jurisdictions specifically have provisions against Nazism, Nazi
> symbols, Nazi quotes in public. [France/Germany/Austria and others,
> particularly in Europe].
Indeed. And as I mentioned last year, there are laws against
lèse-majesté in Thailand that are sternly enforced. This fact _should_
be an arrow in the quiver of the opponents of the package, but they
never make this point. It's worth trying to infer why. Experience
shows that intimate knowledge of fortunes-off's contents is seldom in
evidence (from anyone).
But no one has, as yet, in either of these threads as far as I can
recall, identified more than one objectionable fortune. The one I'm
aware of is a quote by Anita Bryant. In isolation it is ugly. I think
I would either remove it, or pair it with another that comes up when one
runs "fortune -o -m Anita".
"As a mother, I know that homosexuals cannot biologically
reproduce children; therefore, they must recruit our children."
[Anita Bryant, 1977]
"If God dislikes gay so much, how come he picked Michaelangelo,
a known homosexual, to paint the Sistine Chapel ceiling while
assigning Anita to go on TV and push orange juice?"
[Greg R. Broderick]
Why can we not cope with this and further cases via the Debian Bug
Tracking System? The material can be ROT13ed if necessary, or by
preference.
> More cogently: where are we going to get our fortunes from - where's
> the canonical source now that FreeBSD has gone?
As an aside on style, one comes across better by letting one's audience
decide on the relative cogency of one's points, rather than populating
the scoreboard yourself.
To your point, who says we need a canonical source? Debian has plenty
of native packages.
> Who is going to take responsibility for checking quotes and
> translations in all languages and dealing with requests for additions
> and deletions?
Since when has this been a requirement for any package?
> [Each language should have the full quota of quotes where feasible -
> compare the Debian installer or the wiki - no language should be
> inferior as far as this is possible]
Sounds like a "nice to have", not a criterion for retention in the
distribution. I perceive no particular reason for this principle to
apply to fortune cookies. More than some material Debian distributes,
the fortunes have linguistic and cultural context.
This point seems to me like it is grasping for a rationale to me. Your
most cogent points, if you will, should lead your case.
> If it is the package maintainer, is this an appropriate burden for a
> package on which others may judge the project as a whole, rightly or
> wrongly?
The seems like a second-order makeweight objection; as if fearing to
lose the argument on the basis of the package's (English) content,
you're protesting that people will think ill of Debian because the
project wasn't diligent about translating materials you'd rather didn't
exist in the distribution in the first place.
> Whose freedom to select quotes trumps all other opinions?
Nobody "trumps" all other views. In my earlier lengthy reply this
month, I anticipated a disputation process that would work like this.
Here it is in detail.
1. Someone files a bug against the package, citing the individual
fortune(s) to which they object. There's no reason to require that
they be quoted; whatever unambiguously identifies the entries in
question would suffice. (Observing that the command's '-m' flags
work as union instead of intersection operators, I find myself
wishing that fortune(6) more closely resembled lookbib(1).) Some of
sort of volume- and rate-limiting is necessary; filing a report for
every fortune in the (former) fortunes-off package would be abusive
of project resources, not simply of me, who is used to it.
2. The package maintainer (me, I reckon) reviews the items and updates,
discards, or retains each--per my personal taste, which is informed
by affection for the Debian Project and a desire to see it endure,
since I have no better metric I can apply as a volunteer on
unscheduled time.
3. The reporter is either satisfied or not. If not, I reckon the forum
of appeal is the CoC committee.
4. The CoC deliberates.
5. The CoC either directs me to dispose of some item(s) or does not.
6. If they do direct me to dispose of one or more items, I decide
whether I can live with that decision. If I can, I adopt it and
upload a new version. If I can't, I guess my experiment in
maintaining the package concludes, and I orphan it.
7. Independently of the foregoing, I can at any point be referred to
the CoC committee for writing too many lengthy emails, or otherwise
for being a nuisance. I may be expelled, and thus the package
becomes orphaned and removed by default--a stern warning is sent to
any would-be adopters thereby. Use of the passive voice here is
deliberate. This sort of operation has no face. (If this dark
musing is incomprehensible to the reader, be glad you weren't around
20-25 years ago when our project had a cabal. TINC.)
> Branden - if you introduce a new "fortunes-nsfw", this is a new
> package which will obsolete all previous ones and will need to go
> through NEW?
I'm not au courant with incoming upload queue processing. So I suppose
it would, and it is therefore an excellent avenue for bottling the
package up indefinitely, if it's a decision that needs to be made with
no one person's fingerprints on it.
> if you really want the Project to continue with this package / these
> packages, may I suggest a straightforward series of small changes?
I proposed my own and got zero feedback.[4] I'm not sure what to make
of that. ("tl;dr" seems likely.)
> * Make the fortunes package a reader for fortune-format files.
That would be the current "fortune-mod" package, wouldn't it?
> * Add a doc package detailing how to create the valid format of files
> that fortune as a program will read. How to form a fortune from
> arbitrary text.
Good heavens, I don't think we need a separate -doc package for that.
The file format is already documented in the man page strfile(1).
Glancing over it, I get the feeling that the file format is not
described as rigorously as I would like. I fear that anyone who follows
upstream groff commits is tossing a bag of popcorn in the microwave now.
(Argh, and the strfile command throws anonymous diagnostics too...this
_really is_ old-school Unix.)
> * Debian as a whole stops shipping fortune formatted files
As opposed to Debian in part? I don't understand what your adverbial
phrase is communicating.
> and lets users compose or download/translate their own fortune
> databases.
Aren't we already doing this? Has anyone suggested that we stop?
I think a proposal for a future course of action can, for economy,
generally be understood to maintain the status quo in areas where it
clearly doesn't disrupt it.
> There's no censorship of files/thought/speech
I can't tell if this is an item in your proposed plan, or a declaration
that the statement is a necessarily true description of it. Either way,
it is a judgment call, not something that is measurable in the same
manner as your earlier points. It is not for you to decree that what
you're suggesting cannot be censorship because you've _said_ it isn't.
> Each user is free to create their own fortunes to suit how they feel
This seems redundant with your earlier point ("let users compose"),
which itself was a description of the status quo.
Offering something a person already has, as an inducement to accept your
removal of something else they possess, and characterizing this as a
fair exchange is generally frowned upon in negotiations and commerce.
Alternatively, if you have authority, or sufficient influence, to
prevent the return of the contents of the former fortunes-off package to
unstable (or even experimental), even with curation, then your proffer
is indeed not a deceptive one. But it does suggest that my fear last
year that Jonathan Dowland's upload was intended as (or to facilitate) a
fait accompli was well-founded.
> The Debian Project as a whole does not have to take a position on the
> content of any file, though noting that the removal of the prior
> fortunes files follows established practice by other distributions
> when encountering these problematic files with no sources, poor
> attribution or other issues.
"...no sources, poor attribution, or other issues." I can't see what
this point is intended to illuminate. To paraphrase, "this thing is, in
a way, like everything else". I don't know how meaningful analysis is
supposed to proceed from such an inchoate foundation.
I once joked that I had an easier time communicating with ontological
monists than Cartesian dualists, but this statement strains that
observation to the breaking point.
At 2023-08-21T19:05:47+0000, Jeremy Stanley wrote:
> Removing a package from the archive purely on the grounds that it
> contains objectionable text, if such is the reason for not
> distributing it, is making a value judgement of that text. The
> concerns that have been raised so far for objecting to the content
> of the package in question are applicable to quite a number of other
> packages in Debian as well. Hyperbolic perhaps, but it doesn't seem
> that far separated as analogies go.
I keep trying to make the point that if people would just quote the
specific darned fortunes that they have a problem with, we could focus
this discussion immensely.
> Maybe book banning rather than book burning is more familiar to
> modern audiences? In the country where I reside, libraries are
> pressured not to carry books that vocal members of the community
> find offensive for whatever reason, and those libraries often cave
> to the pressure because it's easier than explaining to
> pitchfork-carrying mobs that not every book in the library is going
> to be to their tastes.
>
> Removing a package from the archive because there's nobody
> interested in maintaining it (not merely expressing an interest but
> actually doing the work), is another matter of course. Like a
> library choosing not to repurchase a particular damaged book due to
> lack of popularity, rather than being pressed to remove it from the
> shelves because someone disagrees with what's printed inside even
> though they're never going to check it out and read it for
> themselves anyway.
I think your analogy is worthy. Debian still touts itself as "the
universal OS". That claim, as aspirational as it may be, includes
librarians and people who organize their intellectual lives like
librarians just as much as it does the IT staffs of Fortune 500
companies with well-heeled HR and DEI departments who have been informed
by the executives that if the rosy prospects planned for the quarterly
10-C filing with the SEC is in any way threatened by a civil suit from
an employee, the precarity of their own jobs sharply increases.
At 2023-08-21T19:09:00+0000, Holger Levsen wrote:
> if an image, a png or a jpeg, is considered "software" by us, I'd very
> well also argue that packaging software is communication to the inside
> and outside of our project.
As noted above, it's the "inside" aspect that would be relevant here.
> and if there is disagreement about this, we should extend the CoC.
That's not going to relieve us of the responsibility of deciding what it
is that's being communicated.
> “I'll tell you what freedom is to me.... No fear.” (Nina Simone)
Highly apropos quote--and from one heck of a musician and personality.
Regards,
Branden
[1] https://www.debian.org/code_of_conduct
[2] https://tvtropes.org/pmwiki/pmwiki.php/Main/Jerkass
[3] nroff and troff
[4] https://lists.debian.org/debian-project/2023/08/msg00045.html
[5] https://bugs.launchpad.net/ubuntu/+source/fortune-mod/+bug/1996682
[6] https://lists.debian.org/debian-project/2022/11/msg00037.html
[7] I suppose the irony of deploying the ethnic theory of hypodescent to
a textual corpus in the defense diversity and toleration is
difficult to grasp.
[8] https://www.salon.com/2005/01/25/rice2/
Attachment:
signature.asc
Description: PGP signature