Thank you, Kevin for your thoughtful and rather complete reply!

Mark Wagner <mark+...@carnildo.com> wrote:
> Of course, this only works for ordinary relations.  If the way you
> clicked on is shared by two or more relations, you need to go
> through the far more complicated method of playing with the
> relation-editor dialog.

It is no secret that using either iD or Potlatch for relation editing is 
difficult and error-prone, nigh unto impossible for novice editors to either 
understand or perform easily.  By contrast, while it does take some practice to 
learn, I find JOSM's relation editor to be a straightforward method to edit OSM 
relations.  In short, the "four pane dialog" (not strictly correct, but it 
pedagogically suffices) consists of "key-value pairs on top, (left and right); 
member elements and selections on bottom (left and right)."  Along with the 
buttons to the left of and between the bottom two panes (sort, reverse, select, 
move, ...), you have all you need to edit relations.  This is a modeless (not 
modal) dialog, meaning that while the relation editing window is open, 
selections (e.g. click, drag a selection box...) and operations (e.g. split or 
join...) can/should be performed on underlying data in the geography editing 
window.  Taken together, these are the seeds of learning how to effectively 
edit relations in JOSM.

Plug-ins that offer "power tools" beyond that?  Well, caveat usor.

I wholeheartedly agree with much said in this thread:  both polygons and 
multipolygons are perfectly valid data structures to use in cases where 
choosing one or the other is a matter of taste, preference, use-case, or all 
three.  (Some uses absolutely require multipolygons, and that is that, other 
uses offer a choice of polygons OR multipolygons, where one or the other are 
equally correct).  Notwithstanding JOSM's current paradigm noted above, our 
tools have a ways to go before they present simple methods to edit relations 
(type multipolygon or others) so that all and sundry are comfortable editing 
them.  OSM gets better at this, though it is taking some time to get there.

I believe a most important result from this thread is that there are many use 
cases where either polygons OR multipolygons are correct.  Really, we are not 
very far apart from rather fully agreeing with one another.

SteveA
California
_______________________________________________
Talk-us mailing list
Talk-us@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-us

Reply via email to