Michael Kirk

We use pelias for location search. Here's the corresponding issue to add overture to pelias https://github.com/pelias/pelias/issues/954

> truncate the precision of my data, and that will automatically get printed as I want... I think that approach will work (1.0 - 0.07) == 0.9299999999999999... percent of the...

Dumb jokes aside, you mean something like this, right? ``` let y = (x * 10.pow(n)).round() / 10.pow(n); format!("{}", y); ``` I think that should work 100 percent of the...

Since this is seemingly a straight forward performance proposal, we should be able to measure it. Can you include any benchmarks using plausible real world data that shows improvement? Related:...

>> There were no measurable perf differences for me with the existing write benchmark. > > We could set up Criterion, I guess. And it might be worth comparing to...

gdal integration leverages the `gdal` crate, geojson leverages the `geojson` crate (etc.). But for shapefiles, all the format parsing logic lives right in the geozero-shp crate, so I'm not sure...

I’d say give it a try!

Just to clarify: You're using 18-dimensional coordinates, is that right?

Ok, I see your example: https://github.com/hamirmahal/rstar-i32/commit/aef70723a73ca76c71387ef93d825b7afcf5a0ea#diff-42cb6807ad74b3e201c5a7ca98b911c5fa08380e942be6e4ac5807f8377f87fc So you are using 18-D coordinates. Just to ask a silly question: Did you really intend to have a single coordinate with 18 dimensions,...

> Do you think R-trees are a suitable data structure for this purpose? Well, I can say with certainty that this particular R-Tree implementation is apparently not currently suitable for...