[Talk-GB] Change Set Error: 3552319

[Peter Childs]

On 21 February 2010 07:14, WessexMario <wessexmario-osm at yahoo.co.uk> wrote:
>
>> If the whole changeset was removed it might erase some correct mapping
>> that he did.
>>
> Isn't the principle supposed to be: It's better to lave less known
> correct mapping, than lots of mapping some of which is complete rubbish.
>
> If there are very, and many, obvious errors, then it's better to revert
> the changeset immediately, than to leave it until someone else comes
> along and corrects it, or adds new and correct information, which then
> complicates any reversion of rubbish, and also means you have to have
> low confidence in anything else mapped in that changeset.
>
> No, far better to revert it, and if there were (in my opinion unlikely)
> some good mapping in there, then it can always be re-entered correctly
> afterwards.
> If he has to reenter the data then he, (like everyone) will learn from
> his mistakes, but let's ensure that only goof information is maintained.
> Mario
>
>
> _______________________________________________
> Talk-GB mailing list
> Talk-GB at openstreetmap.org
> http://lists.openstreetmap.org/listinfo/talk-gb
>


Ok, I'm a local I live in Wainscott, so yes I do have Expert Local
Knowledge, (I've also mapped most of Hoo.....)

There has been some very very good work on Strood in the last few
weeks by a new mapper, which I was trying to find time to map but had
not got around to yet. Its so good I'm now thinking of promoting
Medway off the mapping priorities.

However in this case the data seamed "Random" badly tagged ways and
nodes that made little sence, together with incorrect data., Coupled
with no GPS tracks, and a user who only ever worked for half an hour
over a month ago.

I know that in previous cases (such as liam123) we have had these
problems and "correcting" the data is generally not the right way
forward, and that in these cases reversion is often a better avenue to
try and follow. But what is the best way to deal with this continuing
issue.

I don't mean to get at this user as I suspect he did not realise the
data was going to go live.

I'm thinking we need to document a standard procedure to deal with
this, when the person who notices the problem may not be the right
person to be capable of correcting it, or the area involved is not an
area they know and the area needs flagging for checking by a local.
(which is not true in this case) With established mappers its easy to
just email the mapper but with infrequent random new mapper, this need
to be worded more carefully, possibly with a standard template.

Just a trying to help

Peter.

Peter.