Talk:Microgrants/Microgrants 2020/Proposal/Road Completion project
You stated that you need a developer for 10 days, but somehow there’s no indication about the skillset that’s needed to do the job. Also, how did you anticipate bringing new people unfamiliar with the project up to speed? Developing new algorithms sounds like there’s some further research needed. Do you think that’s feasible within the 10 day timeframe? Mmd (talk) 13:44, 17 May 2020 (UTC)
One line on the project MapRoulette website brought a smile to my face: "If in doubt: don’t map it! there are plenty of mistakes in the Wegenregister data." I am generally opposed to big automated imports of dirty data. I worked for government - I know how dirty the data can be. The approach here of not trusting the Roads Register and demanding human validation of every map change is quite correct. I am grateful.
The project is building a system that can be used globally to improve the systems of the OSMF. It is the right sort of project. I have some concerns about the ease of deployment etc, and its ability to use date from most commercial pavement management systems, but I will sit on those concerns.
Process automatically the road networks official datasets and display the comparison in a way that will help the mappers to add or fix the road network in OpenStreetMap.
I had difficulties understanding this sentence (I think "road networks official datasets" had better be written as "official datasets of road networks"). May I suggest an alternative version that I think sounds better?
Automatically compare the road network in OpenStreetMap and in government datasets, making the result available to improve OpenStreetMap.
I interpreted "official" as "government". But is this limited to government data? Maybe "external datasets" would fit as well. Of course, feel free to further tweak any of this if you want. —M!dgard [ talk ] 12:31, 3 July 2020 (UTC)