<br><br><div class="gmail_quote">2011/10/11 Ben Laenen <span dir="ltr"><<a href="mailto:benlaenen@gmail.com">benlaenen@gmail.com</a>></span><br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<div class="im">On Tuesday 11 October 2011 13:05:30 Jo wrote:<br>
> I suppose the ref tag might be more appropriate. It is indeed a bit odd to<br>
> use note for it, OTOH that's the only tag that was used consistently in all<br>
> the 1000s of relations I've seen so far...<br>
> Me removing the names on those route relations made the problem that<br>
> Potlatch does not and will not display them more accute though. When ref is<br>
> used, is it shown in Potlatch?<br>
><br>
> If we reach a decision to change note into something else, no manual labour<br>
> will be involved to change them wholesale. My script can take care of that,<br>
> both for hiking and cycling routes. Consensus between contributors in<br>
> Belgium, The Netherlands and Germany needs to be reached though.<br>
<br>
</div>People here should stop thinking in terms of "does editor X show Y". We tag<br>
the correct information, and if editors support it, then that's an bonus. We<br>
decided on the note tag long ago, and IIRC it was JOSM that started showing<br>
the note tag after we decided to use that tag for it.<br>
<br>
The routes don't have a reference number, they don't have a name, they just<br>
connect two nodes, so we chose not to use the ref or name tag. But since we<br>
obviously needed a little bit of help to know which relation is which, we used<br>
the note tag.<br></blockquote><div><br>Maybe the fact that this discussion seems to recur every so often is an indicator that something is not logical about it.<br>Since these routes ultimately have to be entered by people and these
people make a choice, which is not easily altered (I tried in the case
of RoRay) as to what editor they use, it does matter whether editors
support the way we tag those relations. Ultimately that's why, until
recently, all the route relations had identical note and name tags.<br>
I'm working in The Netherlands now and there I find that they often use
the name of the network in the name tag. I started by changing that to
network:name, but now I''m taking them away, since it's information that
can be deduced from the membership of the network relation. <br></div><blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">
<div class="im"><br>
<br>
> That's what I did, in fact. I don't have the maps published by the tourism<br>
> offices of the provinces. I do have a file obtained from Fietsnet with a<br>
> network for each node, but I used that only as guidance. It took an<br>
> inordinate amount of time to try and assign those nodes to networks. I'm<br>
> relatively pleased with the result, but I don't guarantee that I'm entirely<br>
> done shifting them around...<br>
> I agree that inventing a name like Dijlelandse Kouters (based on Brabantse<br>
> Kouters combined with Dijleland) was not the best thing to do and I'll<br>
> remove that one.<br>
<br>
</div>Except that I'm pretty sure that Fietsnet has been looking at the published<br>
maps as well. And I checked this out in the past by looking for differences<br>
between the published maps and the situation in reality and check them with<br>
Fietsnet, which consistently showed the situation in those maps, even though<br>
some routes would just be impossible.<br></blockquote><div class="im"><br>I used it as a guideline, not as a reference. Until recently I had no
idea of the general whereabouts of Meetjesland, Leiestreek, etc. Many of
the nodes in OSM don't conform to what Fietsnet knows about them. <br><br>
> What do people think of the following names I 'invented' though?<br>
><br>
> Westkust (part of Westhoek)<br>
> Scheldeland Denderstreek<br>
> Voorkempen<br>
> Kalmthoutse Heide<br>
> Hageland Droog Haspengouw<br>
<br>
</div><blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">
Probably as good as any other name. The problem for me is: why even invent<br>
names like these? They're completely arbitrary and don't conform to any<br>
subdivision of the network in real life, so why should we tag them as such?<br>
<div class="im"><br>
> All these names are based on Wikipedia searches I performed. So it's not as<br>
> if they are based on nothing. I doubt that it will be possible to encounter<br>
> them anywhere on the guide posts though.<br>
><br>
> For the other subdivisions of Kust, Westhoek, Meetjesland, Leiestreek,<br>
> Waasland, Scheldeland, Vlaamse Ardennen and Kempen I simply used Noord,<br>
> Zuid, Oost en West to subdivide them.<br>
><br>
> What about splitting up Limburg in 3 or 4 regions?<br>
><br>
> Maasland, Haspengouw, Midden-Limburg and the part that extends into Liège.<br>
> Voerstreek was already separate.<br>
<br>
</div>Why even split it up at all in Belgium? For all I care, use one big network<br>
for every province, given that these networks are controlled by provinces.<br></blockquote><div><br>Well, you seem to agree on something with Gerard. In Vlaams-Brabant
we actually do have names for the regions. And that's where I started.
In the other provinces, those regions seem to a lot bigger.<br>
<br>We have 2 reasons for having an rcn_ref twice in a network relation:<br><br>- split nodes around canals, on roundabouts, on separated cycleways on 2 sides of a road.<br>- the decision of the tourism offices to only use 00-99 (except in Limburg and Hohes Venn and on the hiking networks)<br>
<br>When using 1 relation for an entire province or region, like say Kempen, this network relation will<br><br>1. become a big mess<br>2. contain >1500 members, which becomes impractical to work with.<br><br>If you have a look at the network relations I worked on, you will find (as an example):<br>
<br>- the nodes are sorted on rcn_ref<br>- between 33 and 34, there are 3-4 route relations 33-34, 33-57 and 02-33 (connection).<br><br>You will probably say that it's not possible to keep this ordered like this and you would be right. Somebody adding a route to a network with Potlatch will add that route at the end.<br>
When I work on those relations again, I remove those routes and my script adds them near to their node again, along with all the other routes that were not part of the network yet.<br><br>This helps to know the rcn_ref of the nodes and to quickly find a route that my script reported a problem on:<br>
<br><a href="http://wiki.openstreetmap.org/w/index.php?title=WikiProject_Belgium/Cycle_Routes/Node_Network">http://wiki.openstreetmap.org/w/index.php?title=WikiProject_Belgium/Cycle_Routes/Node_Network</a><br><br>It's easier and more practical to work with relations containing 250 members than to work with relations having 1500 members in case of conflict resolution, etc. (those downloads take forever). Same goes for the uploads, of course.<br>
In The Netherlands I haven't encountered a network relation containing more than 300 members yet.<br><br>Jo<br></div></div><div style="visibility: hidden; left: -5000px; position: absolute; z-index: 9999; padding: 0px; margin-left: 0px; margin-top: 0px; overflow: hidden; word-wrap: break-word; color: black; font-size: 10px; text-align: left; line-height: 130%;" id="avg_ls_inline_popup">
</div>