Re: [gentoo-user] Systemd upower

[Tom Wijsman]

On Thu, 05 Jun 2014 02:34:49 -0400
Greg Woodbury <redwolfe@×××××.com> wrote:

> On 06/04/2014 11:11 AM, Canek Peláez Valdés wrote:
> It is a discussion about technological things, yes, but the art of
> dealing with other people *is* politics [1].

Politics are also about dealing with power, not alone people; it is
possible to give a robot a lot of power, that doesn't imply that the
creator or buyer of that robot has power. The robot will have its own
will; that will isn't necessarily depending on what people tell the
robot to do, but also on what the robot will percept from nature. This
then all boils down to the nature's will; there may be a person, robot
or server with a lot of power, but one day the power of nature decides.

> Systemd *may* well be technologically superior in terms of having a
> better method of doing things. (It certainly makes adding items to the
> mix easier than re-doing all the numbering in SysVinit.)
>
> Unfortunately, the advocates and implementers made some major
> political choices when they (apparently deliberately) chose to put
> the systemd stuff in /usr/lib instead of /lib.  It was pointed out
> that this abrogated certain parts of the FHS, forced those who would
> like to adopt it to *not* being able to continue using their machines
> they way they wished to (I.e. they had to choose between several
> potentially major changes to do so -- don't have a separate /usr or
> be forced to use a kernel initrd/initramfs method in order to do so.)

This is the power of putting things in such places against the power of
the FHS; Gentoo uses its power to allow parts of these powers to exist,
as "Gentoo does not consider the Filesystem Hierarchy Standard to be an
authoritative standard, although much of our policy coincides with it.".

http://devmanual.gentoo.org/general-concepts/filesystem

You note how it abrogates part of the power of the FHS, but you don't
mention its consequences and how Gentoo deals with those consequences;
this highlights a power, instead of how that power affects people.

So, yes, eventually politics deal with people; but it does so through
means of dealing with power, only looking at the power or only looking
at the people deceives one from the total picture. Look at both instead.

> These were not mere technical choices, but highly political/social
> choices.  Early on, the violation of the "principle of least surprise"
> could have been easily fixed by simply correcting the placement of
> things from /usr back to / but the developers doing the work *chose*
> not to see it as a mistake or poor choice, and steadfastly refused to
> accept corrections or patches to improve the work by fixing what many
> saw as a mistake.

Where did we agree with the power of the "principle of least surprise"?

What kind of surprise towards the users are we talking about? Short
term surprise? Long term surprise? How does that affect our users?

It can be a mistake in the short term, but that doesn't make it one in
the long term; things work out well, it seems, where is the problem?

> That placement error was not the only social/political mistake they
> made either.  Other suggestions and improvements were offered and were
> ignored or rejected in rather flammable verbiage.

This paragraph misses a reference to the mistakes and verbiage.
 
> As it happens, some of the parties involved work for companies that
> actually pay them to do work on Linux and FOSS, and have leveraged
> that role to the fullest.

Some people give up on money, to reach something else in their life;
"Hey, honey, I don't want you to move to Silicon Valley; stay with me.".

> Actually, that is not the objection.  Developers do and have always
> done that, but often observed much more concern with a) letting folks
> who use their stuff know what they were doing, and b) giving a bit
> more lead time when introducing major changes.

The "road map" concept exists for that purpose; however, a lot of
developers don't use such thing or use it in some other way (TODOs,
bugs that capture feature requests and important changes, ...); what is
however a more used concept are "change logs", where these kind of
things are mentioned. But can users track all upstream's major changes?

> Mo, you misunderstand.  TINC is/was a humorous reminder that there was
> NOT a real "cabal", but merely the appearance of one in the minds of
> those not involved in the day-to-day operations of real systems and
> networks.  The human mind sees patterns and invents explanations when
> there is not enough information available.  There is no reason to
> ascribe to malice that which is adequately explained by ignorance or
> stupidity (willful ignorance.)

This leaves out "possibilities"; a possible explanation doesn't make it
a factual true explanation of the matter at hand, like how you've
mentioned mistakes and verbiage above.

It might be true to you, because you might have these references; it
could also be a possibility to you, because you think you saw such
pattern but can't bring it back up; ...

Now, to me it will by default be a possibility; because for it to be
a factual true to me, I need to be given the reference I've requested.

> In *your* opinion.  I have heard some surprising folks say things in
> private that they would never choose to state publicly.  And that
> covers a lot of people in over 53 years of programming.

I could've heard the opposite; who of us both will people believe?

> I *do* no misunderstand this at all.  You attribute to folks (myself
> included) motivations or misunderstandings that you simply do not have
> the information or knowledge to know for certain.

Consider that similar attributions are seen from you.

> If someone sees something as a problems that you don't agree is a
> "problem" it may just be that your experience or expertise is
> different.
>
> There is a large amount of ego preservation and self-promotion
> involved in these arguments, and many don't have enough insight to
> recognize that humor and social skill are necessary to succeed.

Boring jokes and social interruptions stall[1] instead of succeed.

> It merely claims to be a meritocracy. But like several other
> *political* models, it boils down to an oligarchy, where those who
> obtain power by whatever means, whether consciously or unconsciously,
> do what they must to preserve it.
>
> And the early days of Usenet was deliberately modeled in a
> pseudo-democratic manner.  An opinion poll was set up in order to gain
> some idea about the potential and perceived use for a topic area.  If
> one wanted a topic group established on a widespread basis one needed
> a fair bit of social skill and perception in order to do so.

Nothing holds you from starting a poll on a poll site and do that
today; you'll however note that such poll, either now or back in the
days, doesn't force people to do things.

Power's will goes hand in hand with people's will.

> Those who has the gold makes the rules?

In Belgium we say "Klant is koning",
which translates to "The customer is the king";
you might have a lot of money, that doesn't mean you can set up
whichever rule you want and expect that behavior to be blindly followed.

No, customers will just move on to the next company; regardless of the
gold a company has. The same goes with people in politics, you might
have all the gold to advertise yourself; but that doesn't mean you get a
majority of votes and a political reformation in your favor.

> > So if you want to change the rules, start writing some code.
>
> Been there. Done That. Have the T-shirt.
> BUT, for *some* reason, I still care.

No, really; if you want to form rules, write some code and get the
people you want to change the rules for to be interested in your work. 
 
> ------------ Footnotes  --------
>
> [1] Those who are politically active constantly deal with the more
> politically naive who complain "there isn't really any difference
> between <group_a> or <group_b> - they all suck."  This can be compared
> to a technologically naive person saying "major software projects can
> be thrown together by a bunch of programmers just sitting around
> together at a coffee shop over the weekend."  (Don't laugh -- a US
> Supreme Court Justice said almost exactly that within the past two
> weeks.)

[1] It really does.

-- 
With kind regards,

Tom Wijsman (TomWij)
Gentoo Developer

E-mail address  : TomWij@g.o
GPG Public Key  : 6D34E57D
GPG Fingerprint : C165 AF18 AB4C 400B C3D2  ABF0 95B2 1FCD 6D34 E57D