[Date Prev][Date Next][Thread Prev][Thread Next]
[Date Index]
[Thread Index]
- Subject: RE: Functional objects
- From: "Bilyk, Alex" <ABilyk@...>
- Date: Wed, 22 Sep 2004 12:43:06 -0700
And still this would only apply to those 'objects' that don't have any attributes.
given a table 'object'
table_object =
{
method = function () end,
attribute = 1
}
table_object.attribute = 2
table_object : method()
How can one convert this into a functional one? So, the proposed feature would only apply to objects that effectively are bags of functions with no data. Right? I think, practically, most objects people are dealing with have both methods and attributes. Given an function based object, can one use attributes with them?
like
function_object:method() --> function_object("method")
function_object.attribute --> ???
AB
-----Original Message-----
From: Rici Lake [mailto:lua@ricilake.net]
Sent: Wednesday, September 22, 2004 12:13 PM
To: Lua list
Subject: Re: Functional objects
On 22-Sep-04, at 2:06 PM, Bilyk, Alex wrote:
> With all it's limitations and peculiarities of [mis]use in some
> contexts I fail to see how Lua would benefit from this on one hand,
> while I also fail to see what is wrong with
>
> function_obj("msg", ... )
>
> to begin with. The above form, while marginally longer to type, is
> clear to anyone and works for any string.
Well, that was my first reaction, too. But when I started to think
about it, particularly when playing around with implementing objects as
coroutines, I realised that it is nice to be able to do:
obj:method(arg1, arg2)
without having to know whether obj is a traditional table-based object,
or a function-based object. (In other words, the implementation of obj
could change without having to change consumers of the object.) Of
course, you could write a wrapper for obj to make it work properly, but
Mark's idea seems like a nice way of getting around having to do that.
So (I presume) it would be specifically for that case, in which case it
is not really that limited or peculiar.
R.