Oliver Schulz

> Many data sets (e.g. Unchunked, uncompressed ones) will be memory mapped Sure - but I'll have to deal with large files with chunked and compressed datasets quite often (and...

> ... then an OffsetArray is handy Thanks, Tim, that could indeed come in very useful. To what extent are arrays with not-one-based indexing are supported by Base and common...

How about making comparisons return an interval? If the result of the comparison is true for all possible values left and right of the comparison operator, the result would be...

@gwater, I like that approach - are you planning to register `NumberIntervals`?

Alternatively - maybe that new interval type could be integrated into `IntervalArithmetic`? @dpsanders, @gwater, what do you think?

Can `NumberIntervals` be used with `IntervalOptimisation`? That would be awesome.

> Therefore I'm not sure what advantage replacing the interval type in IntervalOptimisation would bring. Well, my idea was that it would bring the advantages of `NumberIntervals`: No silent "wrong"...

> I think that the idea of using Cassette is a very interesting one. Uhm, I was actually thinking more on the lines of dual numbers and `ForwardDiff` as an...

Is the current approach to always return true for comparisons if they *might* be true useful in some applications? If not, I wonder if it would make sense to return...

Ah, indeed: ``` julia> autodiff(x -> x.a * x.b, Active((a = 2.0, b = 3.0))) ((a = 0, b = 0),) ```