Re: [geometry] is2D

On Jun 7, 2014, at 7:57 AM, Rik Cabanier <cabanier@gmail.com> wrote:

> 
> 
> 
> 
> On Fri, Jun 6, 2014 at 10:45 PM, Dirk Schulze <dschulze@adobe.com> wrote:
> 
> On Jun 7, 2014, at 7:12 AM, Rik Cabanier <cabanier@gmail.com> wrote:
> 
> >
> >
> >
> > On Fri, Jun 6, 2014 at 10:06 PM, Dirk Schulze <dschulze@adobe.com> wrote:
> >
> >
> > On Jun 7, 2014, at 6:57 AM, "Rik Cabanier" <cabanier@gmail.com> wrote:
> >
> >> All,
> >>
> >> is2D [1] is a function that takes no arguments, doesn't change the state and returns a boolean.
> >> Is there any reason why we can't turn it into a readonly attribute?
> > What is the benefit and cost for either of both?
> >
> > It's likely the same. It just seems like this should be attribute since the function has no side effects when you call it.
> 
> Most of the informational methods[1] do not heave any side effects and don’t take arguments. I think it looks weird to make some of them attributes. IMO it is easier to understand if we stick to just having m11 to m44, a to f being attributes. We can discuss it on the next call.
> 
> transformPoint() is not an informational method so it shouldn't be in that category.

We can move the functions to the section of immutable methods. That should be ok.

> toFloat32Array() and toFloat64Array() returns copies of the internal state so those should be functions. (Shouldn't they start with 'get' ?)
>  

This is more in line with stringfiers (toString) and similar functionalities. ‘to' is more expressive as well. ‘get' doesn’t say what you get from. matrix.toFloat32Array() makes it more clear.

Greetings,
Dirk

Received on Saturday, 7 June 2014 06:36:25 UTC