[OSM-talk] GPX trackpoint stacks

[Nick Hill]

I do favour the use of a 32 bit int. In fact, I am trialling the GPX 
database based on exactly that right now :-)

Integers have many benefits over floats. There are many useful 
offset-based clever things you can do with integers which due to 
rounding errors in float calculations are impossible.

Not only do you get a better accuracy for a given memory usage, machines 
need fewer steps to perform magnitude comparations using int compared to 
float.



Robert Hart wrote:
>> 2^24 would be the significant accuracy of a single float.
> 
> Why store the value in a float when you know it has a fixed range of +/-180
> degrees? 
> With a 32-bit int, you could get accuracy of 360/(2^32), for the same memory
> usage.
>  
> Rob
> 
> 
> This message has been scanned for viruses by MailControl - www.mailcontrol.com
>