Re: [josm-dev] Commit message not empty

2009-04-23 Thread Rolf Bode-Meyer
2009/4/23 Ulf Lamping ulf.lamp...@googlemail.com:

I too am on the position the user shouldn't be forced to a comment as
much worth a useful comment has.

 What we really (also) need to explain is: *what* we want users to enter
 into this field! It's not enough to convince people to do something,
 it's also essential to enable them to actually do it!

 I'm thinking more of users thoughts like: Oh, I have to enter a comment
 here. But what do they expect here?!? Maybe the source of my mapping
 data? Or do I need to mention the area involved? Is there any format
 that I have to follow here? and so on.

That's exactly what I think can be a problem for users.
Users should have an idea of what type of comment is expected and
connected to this what it will be used for, where it will appear.

Getting that information to the user is the problem. While adding that
on a wiki page (JOSM help page) wouldn't be a big deal (and should be
done anyway including a link on the startup notes) it's still an
obstacle and putting a five line label on the comment dialog
elaborating the point of it doesn't help either.

 Whether or not entering a commit comment is mandatory - the complexity
 remains, and the need to explain to the user why we want him to enter one.

 Well, no, having a label like:

 Comment (leave empty if unsure):

 would easily remove the complexity for the first time users ...

Maybe a bit more specific like

Freetext comment on your changes (leve empty if unsure):

but apart from that I'm out of ideas myself.

Rolf

___
josm-dev mailing list
josm-dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/josm-dev


Re: [josm-dev] Commit message not empty

2009-04-23 Thread Frederik Ramm
Hi,

Rolf Bode-Meyer wrote:
 That's exactly what I think can be a problem for users.
 Users should have an idea of what type of comment is expected and
 connected to this what it will be used for, where it will appear.

Let them learn by doing. Let them write bicycle ride or afternoon in 
the rain on their first commits. They will sooner or later find out 
what this is about. And forcing them to enter *something* will make them 
think; will make them bring up the question the next time they have a 
beer with fellow mappers. Allowing them to simply hit enter will make 
them ignore the topic and forget about it.

The leave empty if unsure makes it easy for users to ignore the 
concept and I don't want to make this easy for them.

Bye
Frederik

___
josm-dev mailing list
josm-dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/josm-dev


Re: [josm-dev] Commit message not empty

2009-04-23 Thread Pieren
On Thu, Apr 23, 2009 at 11:39 AM, Frederik Ramm  They will sooner or
later find out
 what this is about.
On Wikipedia, newbie sees what helpful comments are by looking in the
page history or the recent changes report which is one of the most
popular page on any Mediawiki project. It is the same here and I think
that's why the current OSM Recent Changes page will be very popular
as well (might be a bit longer in my opinion).

 And forcing them to enter *something* will make them think;

Someone who just wants to fix small issues on his local home area and
use an OSM editor one or two times in his live, may not want to think
and learn what the purpose of changeset comment is.

 Allowing them to simply hit enter will make
 them ignore the topic and forget about it.

Or make them: Ok, they ask me for a comment. For this time, I may
have or haven't a comment. But why not for next time.
Just asking for comment is already making them thinking.

To finish, I will quote some sentences from our JOSM key developers:
Dirk Stöcker said:
Nah. I'm against restrictions, but warnings are fine.

Frederik Ramm said (about wiki page smoothness):
I have tried to understand what this is about but failed. Obviously
some guy named ChrisCF thinks he owns the place and writes things
like I will not allow this and that - an attitude that will surely
influence my own behaviour should I encounter him on the Wiki.


Pieren, the childish thinking like the Wikipedia admins about empty comments

btw, I also wrote a patch in ticket #2434 which allows empty comments,
following the API documentation.

___
josm-dev mailing list
josm-dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/josm-dev


Re: [josm-dev] Commit message not empty

2009-04-23 Thread Russ Nelson
Frederik Ramm writes:
  The leave empty if unsure makes it easy for users to ignore the 
  concept and I don't want to make this easy for them.

I agree with Fred.  I also agree with other people who say that it's
annoying.  It *is* annoying AND necessary.

-- 
--my blog is athttp://blog.russnelson.com
Cloudmade supports http://openstreetmap.org/ 
521 Pleasant Valley Rd. | +1 315-323-1241
Potsdam, NY 13676-3213  | Sheepdog   

___
josm-dev mailing list
josm-dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/josm-dev


Re: [josm-dev] Commit message not empty

2009-04-23 Thread Martin Koppenhoefer
2009/4/23 Pieren pier...@gmail.com:
 On Thu, Apr 23, 2009 at 3:09 PM, Dermot McNally

 I think that empty comments are better than innappropriate comments.
 guessed comments will be hard to implement and will most probably
 not help better than crappy comments.

well at least for simple operations they could be generated, e.g. just
1 POI added, deleted or moved (comment: moved
amenity=drinking_water)
or for modifications of streets (moved 23 nodes, added 3 nodes of
highstreet)(highstreet changed classification from residential to
tertiary)
ecc.

Martin

___
josm-dev mailing list
josm-dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/josm-dev


[josm-dev] connected nodes in blue

2009-04-23 Thread Jonas Krückel
Hi,
at the moment all the normal nodes (if they don´t have a special tag) in 
ways are displayed in a red color (at least in my JOSM from today)
Especially for debugging of osm data the following feature would be very 
useful: all nodes which are connected to another way are displayed blue.
At the moment i have to move the node away to see if it is really 
connected to the way and then press undo.
If you think this feature would be helpful i will open a trac ticket (i 
have no java skills to implement this feature).

Jonas


___
josm-dev mailing list
josm-dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/josm-dev


Re: [josm-dev] Commit message not empty

2009-04-23 Thread Frederik Ramm
Hi,

Dermot McNally wrote:
 Small|Medium|Large update within x km radius of lat/long
 Update of x nodes, y ways, z relations within n km of town or city

[...]

 What do people think?

Almost worthless, I'm sorry to say. Any automatic message can just as 
well be filled in later by some automatism on some server that compiles 
and anlyses changesets. What we're after is the human bit that no 
machine can replace!

Bye
Frederik

-- 
Frederik Ramm  ##  eMail frede...@remote.org  ##  N49°00'09 E008°23'33

___
josm-dev mailing list
josm-dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/josm-dev


Re: [josm-dev] connected nodes in blue

2009-04-23 Thread Henrik Niehaus
Jonas Krückel schrieb:
 Hi,
 at the moment all the normal nodes (if they don´t have a special tag) in 
 ways are displayed in a red color (at least in my JOSM from today)
 Especially for debugging of osm data the following feature would be very 
 useful: all nodes which are connected to another way are displayed blue.
 At the moment i have to move the node away to see if it is really 
 connected to the way and then press undo.
 If you think this feature would be helpful i will open a trac ticket (i 
 have no java skills to implement this feature).
 
 Jonas
 

I think this is a good idea. I'm using the same workaround to see, if
ways are connected.


___
josm-dev mailing list
josm-dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/josm-dev


Re: [josm-dev] Commit message not empty

2009-04-23 Thread Dermot McNally
2009/4/23 Frederik Ramm frede...@remote.org:

 Almost worthless, I'm sorry to say. Any automatic message can just as well
 be filled in later by some automatism on some server that compiles and
 anlyses changesets. What we're after is the human bit that no machine can
 replace!

A little harsh :(

Lest I've been misunderstood... That you could retrofit such an
automated message is clear. For me, the value of computing it at
commit time is that it can be shown to the mapper as a default. The
intention being to trigger the following kinds of thoughts that he
would not think if faced with an empty text field:

* Oh, I can add a comment.
* Ah, so that's what comments are for.
* Ugh! That commit message is a bit crap. I'll replace it with a better one.
* No, my edits aren't centred on Mistfeld, I was actually mapping mole
tunnels, I should make that clear.

Now, if we think a default message wouldn't have this effect, then I
certainly don't think we should add one just for the sake of the
information it would add to the log.

Dermot

-- 
--
Iren sind menschlich

___
josm-dev mailing list
josm-dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/josm-dev


[josm-dev] Combine and Merge

2009-04-23 Thread Russ Nelson
Can we combine and merge Combine and Merge?

Combine: combines ways.
Merge: combines nodes.

If there's no objection, I'll write the patch.

-- 
--my blog is athttp://blog.russnelson.com
Cloudmade supports http://openstreetmap.org/ 
521 Pleasant Valley Rd. | +1 315-323-1241
Potsdam, NY 13676-3213  | Sheepdog   

___
josm-dev mailing list
josm-dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/josm-dev


[josm-dev] Split Way / UnGlue Ways

2009-04-23 Thread Russ Nelson
Can we combine Split Way and UnGlue Ways?  They operate on distinct
selections except in the case of having one node selected.  Then,
UnGlue Ways creates two nodes, whereas Split Way creates two ways.
This is easily resolved by eliminating the Split Way case -- you have
to say which way you want split and where you want it split.

The new command would be named Separate.


*UnGlue Ways(one node) - assigns original node to one way, and creates a
  new node for other ways.  Copies tags.

  If the node only belongs to one way, the tags
  aren't copied to the node which is left in
  the way.

UnGlue Ways(one way)  -  creates new nodes for all nodes shared with
  another way.

UnGlue Ways(multiple nodes) - error (but why not just iterate?)

UnGlue Ways(one node, one way) - same as UnGlue Ways(one node)


*Split Way(one or more nodes) - locates the way to which the node(s) is
 a member and splits it.

Split Way(one way) - silently ignored (greyed out in menu).

Split Way(one way, one or more nodes) - splits way at each node.



-- 
--my blog is athttp://blog.russnelson.com
Cloudmade supports http://openstreetmap.org/ 
521 Pleasant Valley Rd. | +1 315-323-1241
Potsdam, NY 13676-3213  | Sheepdog   

___
josm-dev mailing list
josm-dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/josm-dev


Re: [josm-dev] Combine and Merge

2009-04-23 Thread Russ Nelson
Russ Nelson writes:
  Can we combine and merge Combine and Merge?

No, I suggest this instead:

Can we combine, join, and merge Combine, Join Node to Way, and Merge?

Combine Way(two or more ways) - combines ways.
Combine Way(two or more ways with any number of node selections) -
 combines ways, ignores node selection.
Merge(two or more nodes) - combines nodes.
Merge(two or more nodes with any number of ways) - ignored (greyed
 out)

Join Node(one node) - joins node into one or more of the nearest ways.
Join Node(one way) - ignored but not greyed out.
Join Node(multiple nodes) - ignored but not greyed out.


We make a new command called Join (which is the mirror of
Separate):

Join(two or more ways) - combines ways.
Join(two or more nodes) - combines nodes.
Join(one node and one way) - inserts the node into the way at the
nearest point.  If it's not near, throw a message saying Not near enough

Join(one node and multiple ways) - for each way: Join(node, way)
Join(one way and multiple nodes) - for each node: Join(node, way)

Join(multiple ways and multiple nodes) - error: Too many things selected

-- 
--my blog is athttp://blog.russnelson.com
Cloudmade supports http://openstreetmap.org/ 
521 Pleasant Valley Rd. | +1 315-323-1241
Potsdam, NY 13676-3213  | Sheepdog   

___
josm-dev mailing list
josm-dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/josm-dev


Re: [josm-dev] Commit message not empty

2009-04-23 Thread Shaun McDonald

On 23 Apr 2009, at 20:18, Frederik Ramm wrote:

 Hi,

 Dermot McNally wrote:
 Small|Medium|Large update within x km radius of lat/long
 Update of x nodes, y ways, z relations within n km of town or city

 [...]

 What do people think?

 Almost worthless, I'm sorry to say. Any automatic message can just as
 well be filled in later by some automatism on some server that  
 compiles
 and anlyses changesets. What we're after is the human bit that no
 machine can replace!


That human bit is a reason for a change. For example where someone has  
put a road in that is on 4 year old Yahoo imagery, however that road  
no longer exists, (let say it has been built over), so you have  
removed it to reflect reality.

This is the kind of thing that is really useful in the changeset  
comments.

Shaun


___
josm-dev mailing list
josm-dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/josm-dev


Re: [josm-dev] connected nodes in blue

2009-04-23 Thread Jonas Krückel
Ulf Lamping schrieb:
 Jonas Krückel schrieb:

 If you think this feature would be helpful i will open a trac ticket 
 (i have no java skills to implement this feature).

 I thought a bit about it, but I have no idea how to implement it in a 
 good way.

 Problem here is, that currently nodes don't kwow any ways and ways 
 don't know each other.
Is that a general problem with osm data or are you speaking about the 
data model of JOSM rendering?

Jonas


___
josm-dev mailing list
josm-dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/josm-dev


Re: [josm-dev] Combine and Merge

2009-04-23 Thread Frederik Ramm
Hi,

Ævar Arnfjörð Bjarmason wrote:
 If you're rewriting this perhaps you could also make it support
 merging two (or more) nodes two the location of a specific node,
 currently merge will merge to the oldest node, which isn't always
 what you want.
 
 Perhaps the combine way has this problem as well?

It has in that you cannot control which of the two ways will be extended 
to contain the other's nodes and which of the two will be deleted, but 
since no nodes are moved in the process it is not visible.

Bye
Frederik

-- 
Frederik Ramm  ##  eMail frede...@remote.org  ##  N49°00'09 E008°23'33

___
josm-dev mailing list
josm-dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/josm-dev


Re: [josm-dev] Combine and Merge

2009-04-23 Thread Russ Nelson
Ævar Arnfjörð Bjarmason writes:
  On Thu, Apr 23, 2009 at 6:58 PM, Russ Nelson nel...@crynwr.com wrote:
   Can we combine and merge Combine and Merge?
  
   Combine: combines ways.
   Merge: combines nodes.
  
   If there's no objection, I'll write the patch.
  
  If you're rewriting this perhaps you could also make it support
  merging two (or more) nodes two the location of a specific node,
  currently merge will merge to the oldest node, which isn't always
  what you want.

I recall that the resultant node used to have the average location of
both nodes, but now it seems to use the first node listed in the
Current Selection box.

Seems like if you don't like this behavior, 

  Perhaps the combine way has this problem as well?

Not really.  The ways are already touching, so nothing needs to get
moved.  The tags from both ways get piled into the new way.  If there
are any tags with different values in the different ways, you get
asked which value to use.

-- 
--my blog is athttp://blog.russnelson.com
Cloudmade supports http://openstreetmap.org/ 
521 Pleasant Valley Rd. | +1 315-323-1241
Potsdam, NY 13676-3213  | Sheepdog   

___
josm-dev mailing list
josm-dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/josm-dev