I must be doing something seriously wrong...

Christoph Wickert christoph.wickert at googlemail.com
Thu May 21 16:11:22 UTC 2009


Am Donnerstag, den 21.05.2009, 08:46 -0400 schrieb Josh Boyer: 
> On Thu, May 21, 2009 at 02:20:52PM +0200, Christoph Wickert wrote:
> [snipped]
> 
> >privileges, so one should also pull his duties. If not,  IMHO the person
> >is not qualified for FESCo, simple as that. (Sorry if this sounds harsh)
> 
> It doesn't sound particularly harsh, but I wonder what privileges you think
> you get when you are in FESCo?

As a FESCo member you are to decide over the direction that Fedora
takes, this is the biggest privilege I can think of inside a FOSS
project. 

> >> >I'm one of the few maintainers who directly is affected by the policy.
> >> >Would somebody - preferably a FESCo member, who voted for the flags
> >> >proposal - please be so kind to answer my questions. TIA!
> >> 
> >> >https://www.redhat.com/archives/fedora-devel-list/2009-May/msg01414.html
> >> 
> >> 1) The rationale was given by spot here:
> >> 
> >> https://www.redhat.com/archives/fedora-devel-list/2009-May/msg01427.html
> >> 
> >> or was that not what you were looking for?
> >
> >Not really. Spot just explained the rationale for writing the policy,
> >but as I already outlined that the policy does not cover all the cases
> >it was designed for. Maybe I should have asked more precisely: What is
> >the use of a policy that does not really solve the problems?
> 
> I see.  I think that is why a ticket has been opened to revisit this.  Stay
> tuned!
> 
> >> 2) The advantage for the project was to codify something that has been
> >> dealt with silently since the RHL days.  That being said, the guideline itself
> >> is going to be revisited.
> >
> >I agree that codification is a good thing, but I really don't see the
> >benefit (see above) and your answer does not solve the apparent
> >contradiction between "follow upstream" and "remove flags". We have two
> >conflicting guidelines here, so which one is more important?
> 
> In general terms, 'follow the Fedora policy' trumps 'follow upstream' (see mp3).
> I think we try to make sure the policies don't conflict with upstream where we
> can.  The flags issue certainly does though.
> 
> >> 3) Your example of keyboard layout selection seems akin to language selection.
> >> In my opinion, under the existing guideline, the flags would not be allowed but
> >> you could ask FESCo for an exemption.  Note, that is just my opinion and I don't
> >> speak for FESCo as a whole.
> >
> >Ok, then please some FESCo members please speak up.
> 
> Honestly, I think it would be best to let the revisit of the guideline happen
> first.  If nothing changes as an outcome of that, then we can start soliciting
> opinions on specific items (or opening tickets for exemptions).
> 
> >> 4) This is a good question.  If you have specific examples of things that are
> >> not obviously 'religion' that you would like to have evaluated that might help.
> >
> >How about Scientology, is it a religion? I don't think so, but I know
> >there are people who see this differently, especially in the US. Also
> >you only answered my question partly, I asked for a definition of an
> >"ethnocultural concept". Even the assumed trivial terms are hard to
> >define. What is "country"? Is Tibet one or Taiwan?
> 
> This is a good question, and I don't have an answer for you off the top of my
> head.

I think Toshio already uttered similar concerns, so FESCo should have
dealt with them *before* ratifying the policy. Looking at the IRC log I
see there was *zero* discussion but only +1s. (This is how I'd expect a
vote in the Communist Party of China but not in Fedora - SCNR)

> >> 5) I don't know the answer to your question, nor do I find it relevant in a
> >> discussion about flags.  It's been pointed out several times that the flags
> >> policy opens doors to madness through defining 'acceptable' content, so let's
> >> not start yet another massive thread about that at the moment.
> >
> >I do think it is relevant. The policy deals with religious flags, but it
> >does not deal with religions symbols. I could draw a square around
> >symbol and call it a flag.
> 
> And I could remove a square from the flag and call it a symbol.  Discussing
> theoreticals at this point is only going to lead us to more and more absurd
> cases.  I don't think we're here to draft policies that cover every possible
> situation whether it will ever happen in real life or not.

I know I'm just to negative sometimes. I think we always need to be
prepared for worst case scenarios, which also includes theoretical
questions. Others say that this or that will never happen because we are
all in the same boat, but I think history has proven different.

> >> 6) How do you make sure users are aware of -docs packages, or -devel packages?
> >> I see no difference here.
> >
> >I do. -devel packages for example show up in comps and can be installed
> >with groupinstall, but -flags can't. While developing or building a
> 
> Why can't they show up in comps?

Maybe this already pisses some states and regimes off? Also showing up
in comps was only half of it, the other one was groupinstall. 'yum
groupinstall flags' won't work. 

> >package you will most likely be pointed to a missing -devel package by a
> >configure script or a failed make, but what will point you to the
> >missing flags?
> 
> >Docs don't affect runtime of a program (as per packaging guidelines),
> >but flags do. Not necessarily the fundamental functions of a package,
> >but they *do* make the program behave differently than intended by it's
> >developers. Just take deluge or my xfce4-xkp-plugin as examples.
> 
> You asked about how we make sure users are aware of -flags packages.  I stated
> that I see no difference from how we make users aware of -docs packages.  Your
> reply doesn't address that.  Instead you described the functional difference
> between -flags and -docs.

Exactly. To show you that your argument is very different and IMHO
doesn't hold up at all. 

> The original question was about advertising to the users.  So how do we
> advertise -docs packages to users, and why isn't that sufficient for -flags?

I still think you put it wrong because IMO you are comparing apples and
oranges. 
     1. In a perfect world (TM) there is no need for docs because
        everything is self self-explanatory. 
     2. At least users *should* not need docs at least none that are not
        installed by default. Our -docs packages are mostly targeted at
        developers and system administrators. 
     3. If someone really needs docs, he will realize this himself,
        because he is missing knowledge. For flags he doesn't. How is a
        deluge user supposed the realize the lack of a function?

> 
> josh

Regards,
Christoph




More information about the devel mailing list