[OSM-talk] Voting

[Robin Paulson]

On 09/04/2008, Andy Robinson (blackadder) <blackadderajr at googlemail.com> wrote:
> I haven't expressed my view too much on this aspect of late. I think most
>  know that I'm an advocate of the "let it evolve" approach.

me too. it should evolve - but settling on agreed ways of doing things
does not prevent evolution

>  SteveC pointed me last night to this:
>  http://youtube.com/watch?v=WMSinyx_Ab0
>  If you haven't seen it already its principally discussing the arguments and
>  issues surrounding wikipedia and whether it can stand for truth or not.

there is no truth. only commonly agreed upon values. this can be
applied to society, wikipedia or osm, any group of individuals with a
common aim

>  OSM basically has the same dilemma. There will always be those that think
>  the prescribed approach, and this applies beyond tags too, is the only way
>  the project will be considered authoritative and therefore in the longer
>  term useful/successful. I don't hold this view, and this is why.

prescribed != non-evolving

>  Like another poster I too use international standards in my life as an
>  engineer. But daily I come across poorly conceived standards and differences
>  in interpretation, usage and supposedly equivalent standards in different
>  jurisdictions. I also see standards having to change with time and that
>  these changes don't usually keep pace with technological developments or new
>  research and best practice.

that isn't a failing of standards per se, though, only with their
implementation. some engineers are lazy and can't be bothered reading
how the standard should work (by the way, another engineer here), but
that doesn't mean the standard has no value

i use standards every day too, and plenty of them change every few
years. why don't they change more? well, the standards committees have
to meet, which costs money (which isn't available), work has to be
done researching methods (which is expensive and time-consuming). as a
result, standards committees appear slow to move and out of touch

we have a wiki, and everyone can cheaply get together and
investigate/discuss if a new process is useful or not, so we can
update the 'standard' every day if need be. and we do. every day there
is an active discussion/vote to add, change and remove tags. how is
this non-evolving?

>  The final minute of the above video for me is the important point. An
>  expert, whether it is on tags or anything else, has a high degree of
>  knowledge about the subject, but that is not the only knowledge. Any
>  knowledge, whether from an expert of not is knowledge gained about the
>  subject and has relevance. This is why I think the original wikipedia

absolutely. there is no barrier to joining a discussion. no-one
looking at a tag discussion with a good idea would think ulf, alex, me
or anyone had some higher power. we're very careful on this, and are
aware and promote that every opinion is as valued as another. has
anyone ever said "i've been doing this for x months, i know more than
you, your opinion is worhtless?" not that i can see

>  approach was fine, provided that it would never be considered authoritative.
>  If you want an "authoritative" version, in the same way that the
>  Encyclopaedia Britannica or OED might be considered authoritative, then
>  fine, make your rules and produce your work to standards, each individual
>  then has the option to consider these alongside any other sources of
>  information when making a decision or taking a view about something.
>
>  If we turn this point to OSM we can see that if the community pools its
>  ideas on a point, tags in this instance, then we reach through discussion on
>  the lists/IRC wiki etc a level of general consensus about a tag, it is
>  immaterial whether the consensus reached is right or wrong in the wider
>  context. It's what the community feels is appropriate at the time. The
>  problem comes along only if a subset of the community decide to "approve"
>  the consensus and cast it in stone as an immovable statement. Doing so stops
>  further revision of the community consensus, and thus in my view makes it
>  less authoritative with time.

no it doesn't - anyone can propose changing a tag at a later date.
e.g. i put forward  a proposal to merge cemeteries and graveyards,
someone explained why they were different (showing good sources to
back up their argument), and i retracted it. there is also a proposal
in place to delete sport=football which will probably go through. when
it was created it made sense, now it doesn't - a prime example of
evolution

voting by itself does not give me any confidence that the tags are
'approved' or useful or whatever. but people use them, a lot, which
gives me confidence that they see value in the tags, and thus how they
are created

i don't know the exact numbers, but having looked through tagwatch,
most items are tagged with things in map_features. why would people
use them if they had no value? this tells me we're doing something
right. maybe the people who use them don't know how they got there.
still doesn't matter, they think they are useful

>  As Frederik says, there is absolutely no problem with a subset of the
>  community voting and deciding what it recommends, the result can even be
>  part of the consensus, but I don't see it in any way as a benefit to the
>  project to enforce and standardise the result and make that the "official
>  way".

no-one has said it is official. everyone knows it's a wiki and anyone
can join in

>  I'll make one final point, apologies for some of the nostalgia in what
>  follows, though the background might be useful to record for posterity.
>
>  I drafted the original map features following a pub discussion with SteveC
>  and Alex Willmer in November 2005. It was the first time I had met Steve and
>  he had specifically come up to Birmingham to meet us to get our input on the
>  project, it was early days of course and reaching consensus in those days
>  was often a matter of a beer in the pub. Point 10 on his agenda said "Tags
>  and the data structure. The ideas on the wiki appear unstructured." In those
>  days the community was little meetings and discussions like this. I was keen
>  to show that I had some ideas about how we might display our data beyond
>  just making nodes and segments and adding a single "class" attribute. I had
>  knocked up a list of features in a document called "WAYS". Its here if you
>  want to see what it looked like:
>  http://ajr.hopto.org/osm/WAYS.pdf
>
>  You can see from it where most of the original Map Features came from.
>  Nobody has ever suggested we approve the original map features list. The
>  vast majority of which is still in use today, warts and all. The community
>  acted pragmatically, it saw the list was useful and used it. What has been
>  the real surprise to me is that nobody has done the same thing in a
>  wholesale way for other "non map" aspects of OSM, such as routing.
>
>  So the point I wanted to make is that even if just one individual has an
>  idea and implements it and it gets used buy someone else then its useful, it
>  benefits the project. We need to get back to encouraging and embracing new
>  ideas that individuals or groups come up with. The project has seen a number

we do. anyone can suggest a tag, and then debate with anyone who is
interested whether it's useful or not. everyone gets their say. no-one
is restricted from speaking - that's the point of a wiki

>  of them in the past and they have all been defining moments in the history
>  of the project. We need more of them to keep the evolution of OSM moving. So
>  rather than debating an idea or proposal lets be more pragmatic and get on
>  and finish the map. It may not be the cleanest set of data out there, but

i agree, but creating the tags is a step towards finishing the
geodata. the data has to be structured, or we lose all the edge of
putting it in a database that can manipulate, filter and modify it. we
might as well draw it with crayons if we don't

>  right now it's the only free set there is of its type and we have stolen a
>  march on everyone else. As a human race we continue to evolve. OSM should be
>  the same. Lets keep it that way.