[Talk-GB] Import Progress

Frederik Ramm frederik at remote.org
Sun Mar 19 19:36:11 UTC 2017


Rob,

On 03/19/2017 04:04 PM, Rob Nickerson wrote:
> I don't think any of us are members of the import mailing list and I
> don't see the point of joining any more mailing lists. They represent an
> arcane 20th century solution that allows a few negative comments to
> derail a locally supported project.

Please consider whether your words might offend those who think
differently about mailing lists.

The imports mailing list is *the* place in the project where you can get
a lot of eyeballs helping you iron out problems. That's why we generally
ask for imports to be discussed there before they happen; a large number
of problems have been fixed before, instead of after, an import that way.

A comment that points out a problem with an import might be perceived as
"negative" but that's very short-sighted; even if it halts an import for
a while by pointing out an issue, it is quite likely that the overall
result of that "negatice" comment will be a better import (or a
non-execution of a bad import).

The West Midlands community is one of the oldest and most active local
communities we have in OSM. It would be great if that exposed situation
would make them lead by example, rather than assume that everyone else's
rules don't apply to them.

The question whether a truly local import needs (potentially
international) discussion pops up often, and in many cases it might be
unnecessary - but for every well-executed import by an experienced local
community there's also a botched one by an inexperienced local community
who were just as sure of themselves as the experienced one.

Your idea of local empowerment sounds nice but I don't think it can be
used as a general principle which I would like to illustrate by
stretching that line of thought to the farcial extreme:

"Our local community wants to get rid of black people. When we discussed
this with other people outside of our local community, we were met with
a lot of negativity. We would prefer if those communities who do not
want to get rid of black people would focus their energy on something
positive to them, while we help other communities who also want to get
rid of black people."

With that I want to say that you cannot be constructive and positive
about everything; sometimes someone will want to do something where it
is actually good that they have to listen to those who say no, and
cozying up in your own little bubble without any negatvitiy can feel
nice to you but be bad for the world.

We need to strike a balance here, but "every local community can simply
import whatever they want and it is none of the rest of the project's
business" is certainly not that balance (unless the local community runs
their own database and their own API).

Something we also ask all mappers to do, and again something where I'd
love to see West Mids team lead by example, is to reply to public
changeset comments and not ignore them, for example here:

https://www.openstreetmap.org/changeset/46819034

> Few people replied to Brian's messages on the local list as we had
> already discussed and agreed it.

It occurs to me that the changeset comment I linked above points out a
few potential issues that a wider discussion could have unearthed before
the import happened!

> Far from a bad import i think this is a great example to hold up as a
> good case study. 

I think you might be getting carried away here. The changeset comment I
linked mentions the following issues:

* usual import guidelines not followed - as I explained above, while it
is debatable whether they must always apply to local imports, a "great
example to hold up" would certainly include that.

* just looking at the first tree in this changeset
https://www.openstreetmap.org/node/4732520560 there's a bunch of
duplicate information on it (site name, ward etc.) and the species info
might have been better placed as "genus"

* https://www.openstreetmap.org/#map=19/52.44699/-1.84369 shows an
offset between tree data and other data; has this been discussed?

* unclear plan for keeping imported data up to date

Now these issues don't put the import in the class of "bad imports that
warrant an immediate revert", but they are not issues I would like to
see in an import hailed as an example to be followed by others.

Bye
Frederik

-- 
Frederik Ramm  ##  eMail frederik at remote.org  ##  N49°00'09" E008°23'33"



More information about the Talk-GB mailing list