Chebyshev should be different than Ultraspherical{0} with simple conversion?

[dlfivefifty]

Right now julia can't have constructors for typealiases with parameters. This is preventing the addition of types to FunctionSpace, since then Chebyshev requires a parameter. If it was its own type this issue would go away.

[MikaelSlevinsky]

Isn't there a normalization difference between the mathematical definitions of Ultraspherical{0} and Chebyshev? Not that this is enough for me to form an opinion...

[dlfivefifty]

I don't think Ultraspherical{0} is defined, though geigenbauir is..

Sent from my iPhone

On 21 Jan 2015, at 12:42 pm, Richard Mikael Slevinsky notifications@github.com wrote:

Isn't there a normalization difference between the mathematical definitions of Ultraspherical{0} and Chebyshev? Not that this is enough for me to form an opinion...

— Reply to this email directly or view it on GitHub.

[ajt60gaibb]

Yes, Ultraspherical{0} is not defined.

[MikaelSlevinsky]

I guess there's not harm leaving it the way it is then.

Cheers,

Mikael

On Jan 27, 2015, at 4:27 PM, Alex Townsend notifications@github.com wrote:

Yes, Ultraspherical{0} is not defined.

— Reply to this email directly or view it on GitHub.

[dlfivefifty]

I think it's better to wait to 0.4 when typealias can be overriden, and then

typealias Chebyshev{T} Ultraspherical{0,T} Chebyshev(d)=Chebyshev{eltype(d)}(d)

will work fine.