[OSM-talk] Voting process (was: Re: Map Features, maxspeed and maplint)

[Dave Stubbs]

On Fri, Oct 10, 2008 at 3:43 PM, Tordanik <osm at tobias-knerr.de> wrote:
> Shaun McDonald schrieb:
>> In my opinion the voting process is broken, as it can potentially vote
>> in proposals that will break backwards compatibility and require
>> extensively more complex processing of the data. Take for example:
>> http://wiki.openstreetmap.org/index.php/Proposed_features/Status
>
> Yes, proposals can break backwards compatiblity. I do not believe this
> is a bad thing at all – if a new concept is better than an old one, then
> maintaining b.c. is the last thing I want, especially in a project with
> as random and insufficient "standards" as ours. That is not to say that
> b.c. isn't desirable, but it should by no means be required for new ideas.
>
> Moreover, the possibility to break b.c. is not limited to proposals –
> the competing concepts of "just add it to the wiki" and "just use it"
> can break b.c. as well, if not easier.
>

Quite true. But the voting process lends a degree of legitimacy and
officialness (where neither really exists) to the uninitiated masses
(and some of the deluded too).
That proposal is quite interesting for several reasons, but mostly
because it's so close, and yet the voting reasons are so varied. One
voter recently voted against it because they didn't like the key name,
compared to lots of people opposing because of the way it breaks
things. Most people supporting it want it because they want a way to
say something is under construction or similar, and don't seem to have
really considered whether it's the best way at all. There's several
comments about improving/degrading rendering performance, which
frankly is a laughable argument either way as it'll make negligible
impact. Then there's the rather nebulous "scalability" argument:
apparently it's scalable, or it's not scalable, I haven't really
figured out what this actually means in the context except that people
are very insistent that it is/isn't.

Anyway, should the vote "pass" by 1 or two votes, some wiki-person
will promote it to map features where all the discussion will be gone.
I'll almost certainly add a section to it's description to discourage
it's use, which will almost certainly be removed by someone else
claiming this is not the place for discussion and that it's an
approved feature don't you know.

Meanwhile the a few of the front page renderings, most of the routing
engines and a pile of other tools you've never heard of will continue
to simply ignore it. Just like they ignore the "disused" tag which has
all the same problems and I wasn't even aware existed (hence why it is
such a big problem).

So while breaking backwards compatibility is not always a bad thing, I
think you do need a _real_ consensus that it's a Good Thing before you
go away and tell everyone to do it.


> I readily admit that voting has its flaws. Looking at the quoted
> proposal, you'll find that two of those who voted against it
> (Nibblenibble and Basemonkey) have only a single contribution in the
> wiki – the vote –[1], have created their accounts the same day they
> voted[2] and have cast their votes within 30 minutes from each other.
> Also, at the time of this writing, no OSM accounts exist whose names
> correspond to these wiki accounts. The two still might be legit voters
> (and when in doubt, it's probably fair to assume they are), but it's
> impossible to tell.
>
> But even this doesn't mean that voting is useless. The RFC+voting serves
> as a way to initiate discussion, collect ideas and encourage sufficient
> documentation. We are not talking about legally binding decisions here,
> it's just a tool for these purposes. And if someone has better tools to
> offer, I will prefer these, of course. It's just that this hasn't
> happened yet.

I'll grant you that the RFC does initiate discussion. And that voting
certainly causes a discussion frenzy on the more controversial items.
Beyond that though I'm not sure what it is trying to achieve. The
effect is "approval" or "disapproval" which actually means very little
except as a large stick to beat anybody who disagrees with the result
when they try to create/edit pages documenting actual usage on the
wiki. It's given people the idea there's such a thing as deprecation
too.

Some better tools would be awesome, but you're right, they don't exist
so we currently have the choice of voting or nothing, and personally
I'd prefer nothing. By all means keep the proposal and RFC parts, and
maybe back them up with TagWatch links.

Dave