deriving from float or int

[Russ]

Does it ever make sense to derive a class from a basic type such as
float or int? Suppose, for example, that I want to create a class for
physical scalars with units. I thought about deriving from float, then
adding the units. I played around with it a bit, but it doesn't seem to
work very well. Am I missing something here? Thanks.

[Sybren Stuvel]

Russ enlightened us with:

Does it ever make sense to derive a class from a basic type such as
float or int? Suppose, for example, that I want to create a class
for physical scalars with units.
That makes sense.
I thought about deriving from float, then adding the units. I played
around with it a bit, but it doesn't seem to work very well. Am I
missing something here? Thanks.

Yeah, you forgot to include a magical 'make the reader psychic' text
;-)

An explanation of what you did and what went wrong would be useful in
answering your question.

Sybren
--
The problem with the world is stupidity. Not saying there should be a
capital punishment for stupidity, but why don't we just take the
safety labels off of everything and let the problem solve itself?
Frank Zappa

[bearophileHUGS]

Probably this is interesting for you:
http://home.tiscali.be/be052320/Unum.html

I think its API can be improved, but it can be used.

Bye,
bearophile

[Russ]

The problem is that when I derive a new class from float, the darn
thing won't let me create a constructor that accepts more than one
argument. I need two arguments: one for the numerical value and one for
the units. But when I try to give the constructor two arguments, I get
this when I call the constructor:

TypeError: float() takes at most 1 argument (2 given)

In other words, python doesn't seem to want to let me "extend" the
float type. I don't understand the reason for that, but I assume there
is a reason.

[Robert Kern]

Russ wrote:

The problem is that when I derive a new class from float, the darn
thing won't let me create a constructor that accepts more than one
argument. I need two arguments: one for the numerical value and one for
the units. But when I try to give the constructor two arguments, I get
this when I call the constructor:

TypeError: float() takes at most 1 argument (2 given)

In other words, python doesn't seem to want to let me "extend" the
float type. I don't understand the reason for that, but I assume there
is a reason.

http://www.python.org/2.2.3/descrintro.html#__new__

--
Robert Kern
ro*********@gmail.com

"In the fields of hell where the grass grows high
Are the graves of dreams allowed to die."
-- Richard Harter

[Sybren Stuvel]

Russ enlightened us with:

The problem is that when I derive a new class from float, the darn
thing won't let me create a constructor that accepts more than one
argument.

Use __new__, not __init__. It's the function that's called when a new
immutable object is created.

Sybren
--
The problem with the world is stupidity. Not saying there should be a
capital punishment for stupidity, but why don't we just take the
safety labels off of everything and let the problem solve itself?
Frank Zappa

[Terry Hancock]

On Tue, 21 Feb 2006 15:01:22 -0600
Robert Kern <ro*********@gmail.com> wrote:

http://www.python.org/2.2.3/descrintro.html#__new__

Curiously, __new__ does not appear in the index of
the Python 2.3 language reference!

It is fixed in Python 2.4, though -- I just checked.

--
Terry Hancock (ha*****@AnansiSpaceworks.com)
Anansi Spaceworks http://www.AnansiSpaceworks.com

[gene tani]

Russ wrote:

Does it ever make sense to derive a class from a basic type such as
float or int? Suppose, for example, that I want to create a class for
physical scalars with units. I thought about deriving from float, then
adding the units. I played around with it a bit, but it doesn't seem to
work very well. Am I missing something here? Thanks.

you could look at how sciPy does it:
http://starship.python.net/~hinsen/ScientificPython/

[Robert Kern]

gene tani wrote:

Russ wrote:

Does it ever make sense to derive a class from a basic type such as
float or int? Suppose, for example, that I want to create a class for
physical scalars with units. I thought about deriving from float, then
adding the units. I played around with it a bit, but it doesn't seem to
work very well. Am I missing something here? Thanks.

you could look at how sciPy does it:
http://starship.python.net/~hinsen/ScientificPython/

ScientificPython != SciPy

ScientificPython's unit package does not subclass from floats since it tries to
be agnostic about the kind of value you can assign a unit to.

--
Robert Kern
ro*********@gmail.com

"In the fields of hell where the grass grows high
Are the graves of dreams allowed to die."
-- Richard Harter