[python-committers] Timeline to vote for a governance PEP

Tim Peters tim.peters at gmail.com
Sat Nov 3 17:30:33 EDT 2018


[Antoine]
> How does Discourse "work better", exactly?

Several examples have already been given.  You're determined to hate
it, and that's fine.


> The long-winded discussion> on variants of voting systems (with
> close to 100 messages) isn't exactly *important* except for voting
> system nerds.

Yet that discussion was spun off from a _different_ thread, so that's
close to 100 messages that don't show up _at all_ on the thread from
which it was spun off.  Better than a threaded view, if you were
looking at the original thread, that's close to 100 messages you'd
never know even existed.

As to its "importance", the title of the thread you're complaining
about ("Python Governance Electoral System")  made it clear it was
_about_ the election system.  If you don't care about the election
system, why read it at all?  You can't seriously complain that a
thread is full of messages it was intended to be about.


> The subthread you started about the "3-2-1" system was of close to no value,

I disagree.  It may well have been of negative value to _you_, but it
served its intended purpose, and 3-2-1 was approved by close to half
the poll voters despite that it wasn't intended to be a serious
contender at this time.  It got _some_ people to think - "does an
attractive method really need a background in graph theory to even
start to grasp?  display bizarre behavior in simple cases?".  Well,
no.  Which was news to some.

In any case, that subthread consisted of a grand total of 4 brief
messages, including my original post.  That 3-2-1 continued to be
mentioned in _distinct_ subtrees shows the seed I planted sprouted.
Good!

> since you admitted yourself that that system is too young and immature
> to be chosen, and you were only "planting a seed" (and probably enjoying
> yourself a bit in the process).

Yup!

> Yet it seems Discourse didn't discourage you from doing so.  Why?

Because, to me, it was a valuable message thoroughly on topic.


> Well, because people making tangents on topics they like to talk about
> is irrelevant to the discussion system used (and your own behaviour
> proves it).

As above, I disagree.  Knocking people loose from a presumption that a
good voting system "has to be" complex or inscrutable was supremely
relevant to the thread's purpose.  As is, I believe it helped lead to
the final decision:  "pure Condorcet", which is soooooo simple that
it's not even "a scoring method".  It's just a form of ballot, with an
agreement that if there's no utterly inarguable winner (a "Condorcet
winner"), we'll try something else until there is.

> The only way you can prevent tangents is by preventing discussion
> altogether.

Sure.  But in this case, I don't agree that "the tangent" you
identified was a tangent, and Discourse _did_ prevent other tangents
from spinning out of control.  For example, when we got to talking
about the possibility of ties, there was pretty quick consensus that
if we wanted to keep on with that, an entirely _different_ message
thread should be started for it.   Much as there was consensus earlier
that the election system messages should be spun off to their own
entirely different thread.  Which happened - but still leaves you
complaining about it ;-)

> *However*, an important feature of a discussion system is to help
> skipping tangents you're not interested about.  A threaded discussion
> system makes it very easy to ignore a subthread.  Not so much where the
> various subthreads are intermingled in a flat chronologic view.

So we'll apparently continue to disagree here.  To my eyes, there were
close to no off-topic tangents in the thread under discussion, and the
people actually participating in the thread were in broad agreement
that some other _related_ topics should be spun off to a different
top-level thread if people wanted to pursue them.

It worked fine.


More information about the python-committers mailing list