I'd like to learn more about that massive mess and how we can prevent that in the future, Paul.To my mind, most TIGER clean up consists of atomic tasks, which is where MapRoulette would typically come in really handy. (Remember the 70,000 connectivity errors we fixed in 2013/4, and the 100,000+ missing railroad crossings which were also attributable to TIGER .) But perhaps you have something different in mind. I'd like to think MapRoulette can help, also because not everybody prefers the same style of working on large projects.-- Martijn van Exel [email protected]
On Mon, May 7, 2018, at 21:12, Paul Johnson wrote: > MapRoulette really made a massive mess of the lanes situation that a > more systematic and big-picture effort is starting to get a handle on > in Oklahoma. I think MapRoulette works well for smaller-picture > stuff, and is more complimentary to StreetComplete and not terribly > great at directing more complicated projects.> > I've restarted my efforts at a county level to avoid having a huge > number of tasks that fall largely in Texas, since it seems that the > tasking manager is ultimately only capable of dealing with rectangular > project areas even if you feed it a more complicated polygon via JSON.> > On Mon, May 7, 2018 at 9:32 PM, Martijn van Exel <[email protected]> wrote:>> > I’d like to see how TM and MapRoulette could be complementary in this >> effort. I know Clifford has set up a TIGER related challenge on >> MapRoulette, and I have done this in the past as well.>> >> I feel that TM can be good for a general ask like ‘check all TIGER >> residential roads in rural areas in this cell, demote to track / >> unclassified or delete as needed’ whereas MapRoulette may be useful >> for more specific TIGER cleanup related tasks?>> >> Thoughts? >> >> Martijn >> >> > On May 5, 2018, at 12:49 PM, Paul Johnson <[email protected]> >> > wrote:>> > >> > I think it's somewhere between 2000 and 2100. I'm working on >> > eventually handling the entire state of Oklahoma on a TIGER >> > cleanup and enrichment. Ideally, I'd like to do the whole state >> > at once (just for variety's sake and for even coverage), but >> > county by county works, too. If there's a limit for the number of >> > times an area can be split, this could really use some work, too, >> > since 3 (based on tasks2 limit) is not enough. A 4000 or 5000 >> > task limit should be sufficient for a single county (though >> > definitely won't start off with that many tasks, and almost >> > certainly won't hit that many tasks over the life of a project) if >> > the split limit is increased (like, at least 5, possibly higher >> > just to be on the safe side).>> > >> > The idea is to basically keep it in that 75-100 item range per >> > task just to keep it manageable (item count based on the resulting >> > selection when using JOSM search to replace selection and >> > searching for highway=* type:way).>> > >> > On Sat, May 5, 2018 at 12:57 PM, Ian Dees <[email protected]> >> > wrote:>> > I don't know. Do you see a limit somewhere? I'm happy to >> > increase it.>> > >> > On Sat, May 5, 2018, 12:35 Paul Johnson <[email protected]> >> > wrote:>> > What is the maximum number of tasks possible on the US >> tasker, and >> > is it possible to change that?>> > >> _______________________________________________ >> > Talk-us mailing list >> > [email protected] >> > https://lists.openstreetmap.org/listinfo/talk-us >> > >> > _______________________________________________ >> > Talk-us mailing list >> > [email protected] >> > https://lists.openstreetmap.org/listinfo/talk-us >>
_______________________________________________ Talk-us mailing list [email protected] https://lists.openstreetmap.org/listinfo/talk-us
