[OSM-talk-be] Problems with JOSM's unwanted behaviour.
Hi, (Warning: long and difficult subject matter ahead! :-) ) I love to work with JOSM, but I have two problems with JOSM. - When you start drawing a way somewhere in a node, JOSM always assumes you want to continue some way already present. This is very annoying and unproductive, because this is nearly always not what you wanted or intended. - When you split a way, the old id (and by consequence its history) are always in the first part, and the second part gets a new id. This way the history and id can be left by the smallest part. It should always assigned to the biggest part. Togheter this combination defeats totally the wiki nature of OSM by disguising the history of one road to some totally unrelated road in the neighbourhood, also making changes to roads the user never intended to edit. An example is in changeset http://www.openstreetmap.org/browse/changeset/7203063 First there was the Verstrekenstraat. http://www.openstreetmap.org/browse/way/98440181 When the user TAA tried to add a track, http://www.openstreetmap.org/browse/way/82861117 from this crossing http://www.openstreetmap.org/browse/node/298362740 JOSM maked this as a continuation of the Verstrekenstraat It was then splitted and adapted to track by the user who didn't want this long Verstrekenstraat. But as this road had the node, where the drawing took place, as starting node, the track end became the new starting node of the resulting way and got after the split the id of the Verstrekenstraat. After the split, (because the user wanted to add a track and not make a long Verstrekenstraat), JOSM moved the existing id (and history) to the new drawn track and maked a new road out of the already existing Verstrekenstraat by giving it a 0 id. In the history of the track , you see the original way of the Verstrekenstraat. http://www.openstreetmap.org/browse/way/82861117/history An example of a node to test this unwanted behaviour is: http://www.openstreetmap.org/browse/node/803205990 Suppose you want to add a track into the wood as a dead end. You select the node and then clicks the way drawing tool. Do a click in the wood (direction SE) and then escape to end drawing the segment and then you end up with an extended path of Voetweg 48. Oh no, I wanted to add a track! Why is it not a blank road? Oh, never mind, we split the result at the starting node and change properties of the new added part, which has now inherited the id and history of Voetweg 48. And on top of that the original Voetweg 48 will get id 0 and be presented from now on in the database as a new way created by you, while you even had no intention to modify it in the first place!! BTW, also relations are inherited, making tails on it after the split. I find this behaviour totally undesiable! Another example to what mess this can lead is the Jachtdreef in Jezus-Eik (it still needs correction) http://www.openstreetmap.org/browse/way/51768961/history http://www.openstreetmap.org/browse/way/108859153/history http://www.openstreetmap.org/browse/way/97836493/history Conclusion: - JOSM should only make continuations of ways on request by the user, eg by selecting them together with the ending node to draw from. - For the sake of history, JOSM should give the existing id to the largest part of the road of a split and not defaulting to the first part. As stopgap for the second point, I try to do a reverse of the road before and after the split, when I notice that the first part is very small.. What do you think? I opened a ticket for this, at JOSM, but it seems a 'feature' and not a bug and the developer won't fix. http://josm.openstreetmap.de/ticket/6567 Comments invited! Regards, Gerard. ___ Talk-be mailing list Talk-be@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-be
Re: [OSM-talk-be] Problems with JOSM's unwanted behaviour.
Renaud MICHEL wrote: On lundi 11 juillet 2011 at 12:22, Gerard Vanderveken wrote : - When you start drawing a way somewhere in a node, JOSM always assumes you want to continue some way already present. This is very annoying and unproductive, because this is nearly always not what you wanted or intended. Press Ctrl while clicking on the end note, JOSM will start a new way. No, this leads to a double node and the way is not connected to the crossing. You can also start your new way at the second node, and then connect it to the end of the other way. I think I've seen JOSM doing a continuation that way to, but it doesn't seem to happen in the example node. However when you need to make a way between two crossings, it can come from either side. - When you split a way, the old id (and by consequence its history) are always in the first part, and the second part gets a new id. This way the history and id can be left by the smallest part. It should always assigned to the biggest part. For that, I temporarily reverse the way to have the history on the good part, then reverse both part back (but this is only important for ways where the orientation matters, like oneways or rivers). That's what i try to do, when paying attention to it, but I feel JOSM should do it automatically. ___ Talk-be mailing list Talk-be@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-be
Re: [OSM-talk-be] Problems with JOSM's unwanted behaviour.
On lundi 11 juillet 2011 at 13:03, Gerard Vanderveken wrote : Renaud MICHEL wrote: Press Ctrl while clicking on the end note, JOSM will start a new way. No, this leads to a double node and the way is not connected to the crossing. Right, I didn't pay attention this. But If you click on the last node of your way, then press Alt while adding the next node, then you end up with a new way that share its first node with the previous way. For that, I temporarily reverse the way to have the history on the good part, then reverse both part back (but this is only important for ways where the orientation matters, like oneways or rivers). That's what i try to do, when paying attention to it, but I feel JOSM should do it automatically. JOSM can't know what part of the way should keep the history, so the best he can do automatically is to always assign the history in a consistent way. The other solutions is to ask the user which part of the way should retain the history, maybe there is a plugin to do that, but I personally prefer the way it works now. -- Renaud Michel ___ Talk-be mailing list Talk-be@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-be
Re: [OSM-talk-be] Problems with JOSM's unwanted behaviour.
But If you click on the last node of your way, then press Alt while adding the next node, then you end up with a new way that share its first node with the previous way. Exactly, and that was what he was told in the ticket. I entirely disagree with the suggestion to disable autocontinuation, and am very content with the way JOSM currently works in that respect. The example in the ticket (starting from a node in the middle of a way produces a continuation) is convoluted as well. It *doesn't* do a continuation of that way. That's also impossible in the data model. In the example (in the ticket) that node is also the endpoint of *another* way, and it does do a contination of *that*. However, it's made out to appear that selecting a non-endpoint node of a way and then drawing from that will produce a continuation. Not so. In short: use the modifier key when drawing a new way from an existing endpoint node. Don't go through all the hoopla of extending a way, then splitting it, deleting the tags, applying new tags. That only makes life difficult and indeed obfuscates the way history. -- Lennard ___ Talk-be mailing list Talk-be@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-be
Re: [OSM-talk-be] Problems with JOSM's unwanted behaviour.
In the example (in the ticket) that node is also the endpoint of *another* way, and it does do a contination of *that*. However, it's made out to appear that selecting a non-endpoint node of a way and then drawing from that will produce a continuation. Not so. No, you didn't understand the examples. I never sayed that a road was boken up in the middle and then continued. It is just when you start a I was not saying that either. [...] This is exactly what happens when editing at node 803205990. This is exactly what I described. As most often, you intend to add a new road, the default behaviour should be like that. I think it comes down to being of another mindset. For me, clicking an end node to continue that way seems the natural thing to do. If it so happens that end node is part of a crossing, so be it. I have to remember to use ALT if I want to start a new way at the crossing. Of course, I could be entirely spoilt from having worked this way in JOSM for years. I thought Potlatch did the exact same thing. What is the handling in Merkaartor for this situation? (Count once your ways when editing and compare extended existing versus added new ways) That's not a fair comparison. Often enough when adding new ways, you are *not* at a junction with another way endpoint. You might be starting a new way in the middle of nowhere, or branching off of another way, but *not* at a junction. In my experience, when adding new ways, starting one at an 'endpoint junction' is the exception. The splitting and additional obfuscating happens, while this seems the obvious way for most users of getting out of this unexpected alongation. For some messy results see my examples Verstrekenstraat and Jachtdreef. I can actually agree with your other point, being that when splitting a way, the existing ID should stay with the longest fragment. You might modify your JOSM ticket to emphasize that part, or perhaps even better: start a new ticket with just that request (with a reference to the current ticket). -- Lennard ___ Talk-be mailing list Talk-be@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-be
Re: [OSM-talk-be] Problems with JOSM's unwanted behaviour.
Lennard wrote: In the example (in the ticket) that node is also the endpoint of *another* way, and it does do a contination of *that*. However, it's made out to appear that selecting a non-endpoint node of a way and then drawing from that will produce a continuation. Not so. No, you didn't understand the examples. I never sayed that a road was boken up in the middle and then continued. It is just when you start a I was not saying that either. [...] This is exactly what happens when editing at node 803205990. This is exactly what I described. I don't want to go into this futher, but your interpretations of my examples were not right: Quotes: - The example in the ticket (starting from a node in the middle of a way produces a continuation) is convoluted as well. - However, it's made out to appear that selecting a non-endpoint node of a way and then drawing from that will produce a continuation. Not so. As most often, you intend to add a new road, the default behaviour should be like that. I think it comes down to being of another mindset. For me, clicking an end node to continue that way seems the natural thing to do. If it so happens that end node is part of a crossing, so be it. I have to remember to use ALT if I want to start a new way at the crossing. Of course, I could be entirely spoilt from having worked this way in JOSM for years. I thought Potlatch did the exact same thing. No, it does not ! Tested with the example node and an unnamed new road was created (you need shift to start drawing from the selected node) in Potlatch 2. Can you even do continuation of an existing road in Potlatch? What is the handling in Merkaartor for this situation? According to the docs, you have the choice by having the node selected or not, when creating a new road.. http://merkaartor.be/wiki/merkaartor/Documentation#Creating-a-new-way and the paragraph below http://merkaartor.be/wiki/merkaartor/Documentation#Continuing-an-existing-way-at-either-end JOSM will always make a continuation unless stopped by the ALT-key. (Count once your ways when editing and compare extended existing versus added new ways) That's not a fair comparison. Often enough when adding new ways, you are *not* at a junction with another way endpoint. You might be starting a new way in the middle of nowhere, or branching off of another way, but *not* at a junction. In my experience, when adding new ways, starting one at an 'endpoint junction' is the exception. This makes it even more execptional and illogical, that when you are drawing, you get always a new way and suddenly a new way fails and your added way is incorporated in an existing way. But OK, narrow down the counts just for the cases in which JOSM would be doing automatic prolongation and I'm still confident that you would see you are more using the ALT key then not, when adding way segments. The splitting and additional obfuscating happens, while this seems the obvious way for most users of getting out of this unexpected alongation. For some messy results see my examples Verstrekenstraat and Jachtdreef. I can actually agree with your other point, being that when splitting a way, the existing ID should stay with the longest fragment. You might modify your JOSM ticket to emphasize that part, or perhaps even better: start a new ticket with just that request (with a reference to the current ticket). I won't . If no other users are asking, the developer won't change its point of view or anything in JOSM regarding these two problems.. ___ Talk-be mailing list Talk-be@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-be