[OSM-talk] The OSM Applet - let's ditch it for JOSM (was: Mapping Reading, some impressions)

Andy Robinson Andy_J_Robinson at blueyonder.co.uk
Tue Sep 5 15:42:00 BST 2006


vegard wrote:
>Sent: 05 September 2006 2:41 PM
>To: talk at openstreetmap.org
>Subject: Re: [OSM-talk] The OSM Applet - let's ditch it for JOSM (was:
>Mapping Reading, some impressions)
>
>On Tue, Sep 05, 2006 at 02:22:20PM +0100, Nick Hill wrote:
>>
>> If seasoned users are not prepared to put up with the shortcomings of
>> applet, then perhaps we shouldn't expect new users to do so either. I
>> haven't seen a good argument why new users shouldn't use best of breed
>> tools.
>>
>
>As a fresh osm-user, I second that. I skipped the applet after a brief
>look at it, did my research on where people went, and went straight to
>josm.
>Unfortunately, some bad compiz/Xgl-combination seems to have creeped into
>my
>Linux-box, so I lost the work in josm from yesterday in an X-crash :) But I
>liked the interface far much more than the applet.
>
>> Tom Carden, I am in no way undermining the admirable work you have put
>> into the applet. I believe it has been an important step in the
>> evolution of OSM. I am simply making the point that because it is no
>> longer the tool of choice, and it has not been getting the attention it
>> needs, perhaps Tom Chance is correct in his opinion that presenting it
>> as the default editing interface causes more harm than good.
>>
>>
>
>Conclusion: I don't think josm is too difficult. I believe most osm-users
>are well above average in technical skills, even as beginners. I don't
>think we stand much chance to attract much non-technical people. I might
>be wrong, nothing would be better, but let's be a little realistic.
>
>My impression as a first-time user, trying to think less technical: It
>was far too difficult to get useful results. It has to be much easier to
>take a GPX track, or a part of a GPX track, and convert it to a useful
>way. It was far too much manual labor for an average newbie.
>
>On another note:
>
>Myself, I'm pondering on the following strategy:
>
>1) Use the new "convert to data layer functionality" on the GPX tracks,
>
>and rather
>
>2) Delete the segments and points on both sides of the interesting part
>of the route,
>
>followed by
>
>3) Cleaning of nodes/segments.
>
>This, I found far easier than manually placing nodes and segments.
>
>Is there any good reason not to do this?
>
>And related to this, I do have a feature request: Make it possible to
>delete "a range" of nodes and segments. This would make it very easy to
>create ways....
>
>Is there any particular reason why it can't be that easy? With an
>accurate GPS, who am I to second where it places the nodes, if I haven't
>deviated from the way? A quick control of the result is good, of course,
>but I think it's a far quicker route to more complete maps if this gets
>easier.
>

>From experience a couple of things spring to mind. 

The first relating to converting gpx direct to nodes and segments. I don't
have a problem with this in areas where I know there is no conflicting data
but having had to spend time deleting someone's poor quality upload of
exactly this over the top of an already edited motorway it has to be used
with caution. The frequency of logging and speed of travel has a big impact
on the end result and how close that matches real life.
Even with SBAS correction you will still find that 10 passages on different
days down the same set of streets will reveal differences in track location.
I'm forever going into the data in my area and tweaking and adding new nodes
just to get it to look smooth and regular. A single gpx track, even when you
think it's accurate, is still likely to wander about at least by a few
metres.

Once you have an area heavily mapped already you will find the facility
pretty much useless. Editing within an already dense mat of nodes, segments
and ways needs to be done with some care, especially when you must consider
the attributes that others may have placed on the data already (you never
know what they are using them for). Don't delete anything without checking
for tags first is my motto. If I do find I need to delete something I make
sure I replace the same tags where logical.

>Feel free to flame me if there is a good reason to :)
>--
>- Vegard Engen, member of the first RFC1149 implementation team.
>
>_______________________________________________
>talk mailing list
>talk at openstreetmap.org
>http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk

Cheers

Andy Robinson
Andy_J_Robinson at blueyonder.co.uk 






More information about the talk mailing list