Improve point documentaion

[timrobertson100]

The documentation for a layer containing only points is lacking. Is it a moveTo with repetition, followed by a series of coords? Or a moveTo and then a lineTo the same spot? The spec appears incomplete for this scenario. A worthwhile addition would be an example for a layer containing points. (A brief answer here would be very gratefully received too ;)

[timrobertson100]

I see the readme has a point example and it is a moveTo. I'd still suggest an example in the proto, which is all I copied into my project.

[springmeyer]

Thanks for raising this - def need to doc this out in more detail and fully explain how single vs multigeometries are handled.

Is it a moveTo with repetition, followed by a series of coords?

Yes

Or a moveTo and then a lineTo the same spot?

No, a point geometry is not going to need a lineTo

To answer your question more detail: all geometry types are encoded the same: First the command (moveTo), then the first delta-encoded and zigzagged coordinate pair (x,y). Next (say with a multipoint or linestring) the next command and coordinates are processed. If the command has not changed (if it is still a moveTo) then it is not encoded directly but rather the count of repeated commands is collected. When a different command is encountered (like lineTo or close) or no more coordinates are found then the first command is updated and we pack the number of repeated coordinates into the bytes that encoded the command: let's call this combined value the "directive". So in the final encoding the "directives" store both the command (moveTo or lineTo) but also the length of repeated commands. This means that when finally encoded a multipoint geometry could look very similar to a linestring: directive,x,y,x,y,x,y,x,y,x,y,x,y.

[flippmoke]

This should be addressed in version 2!