Right. And it is a little bit stupid, because every professional companies use that.
sometimes i wish for us a kind of software development task force or “software working group”, because some ideas are pretty easy to realize, and when the users see the benefit, and it works - well, there is no more skepsis.
Anyway, I can understand critical voices about such developments. I believe Henry ford said, something like “it is still difficult to realize new ideas because the most people goes in another direction and You have to swim against stream.”
Imho, we need application to extract highway graph, and generalized highway graph. Second thing - we need application to extract addresses and pois registry. I think nobody says smth. against this, but nobody wants to do that. You didn’t swim against, but you swim alone.
True. One reason more for thinking about such task force for software developing. Sometimes I think OSMF should organize money for such things. We have terrible imbalance: big, good database and not enough tools, tools, tools…
M.b. I think, first step is working group. We have no priorities. What we shoud invest first? One more editor, easy and beautiful (new users)? Or we should invest in solid web api: presentation, geocoding and routing in one piece of code (more web projects use us - more fame) or applications for extracting, reorganizing and validating osm data (more fame as data provider). I don’t no what is the best choice, looks like economy-strategy game 
Yes, the working group is absolutely the first… In my point of view we should specify the new OSM Api, realize that and based on this specification define work packages of tools.
You have spoken about some needful things. I add:
- 3D WYSIWYG editor / modeller.
- Half automatized recognition of elements from satellite images. (Areas, rectangles, semiautomatic recogniton of ways).
I have a very good opinion about programmers from Russia. I believe it´s not neccesary to wait for solutions from OSMF board 
OK, challenge accepted 
Not sure about Java or Perl, but open license and github definitely. If I succeed.
I have prepared a presentation with test results.
In short: I write a tool and checked two junctions which Bosch mentioned as problematic. Looks like my tool do better
All download links also inside.
Ну вот. Я всегда думал, что этим в Боше просто лень программировать
Ты на конференцию с этим едешь? 
And what about the turn restrictions?
No turn restrictions.
Вот так и знал что кто-нить спросит. Нет, сначала доделаем.