Beginner needs advice

[Steven D'Aprano]

On Fri, 27 May 2011 15:40:53 -0500, harrismh777 wrote:

> Steven D'Aprano wrote:
>> Would you care to revise your claims?
> 
> No.
> 
> 
> You have erected a straw-man... once again.

You keep using that term, but it is clear to me that you don't have the 
foggiest idea of what the straw-man fallacy is.

A straw man is not when somebody points out holes in your argument, or 
unwanted implications that you didn't realise were there. It is when 
somebody makes claims on your behalf that you did not make to discredit 
you, not because you don't understand the implications of your own 
argument.

You stated that Python 2 and Python 3 are COMPLETELY INCOMPATIBLE -- your 
words, not mine. You have repeated that claim, in this very post, and 
others, despite having been shown that they are not completely 
incompatible at all, that it is possible to write both forward and 
backward compatible code that works in every version of Python from 1.5 
through 3.2 inclusive.

Yes, it is true that you can also write code that works in 2.5 but not 
3.2, but so what? You can also write code that works in 2.5 but not 2.6:

    raise "some error"  # works in 2.5 and older

Or 2.4 and 2.5:

    True = 23  # works in 2.4 and older

or 2.3 and 2.4:

    None = 42  # works in 2.3 and older


Do you think that Python 2.3, 2.4, 2.5 and 2.6 are four completely 
different languages, and if not, why not?

Python 3 is not the first backwards incompatible version of Python.


[...]
> All of these things are for the better, I must add.  But, the point is
> that 3.x is completely incompatible with 2.x in real ways.

And you've done it again, despite the fact that you can write compatible 
code that works in all versions of Python from 1.5 to 3.2, and easily 
write non-trivial code that works in 2.7 and 3.2. For larger projects, 
it's probably better to keep a separate 2.x and 3.x fork, but that's for 
convenience, nothing more: numpy, for example, supports 2 and 3 out of a 
single code base.

http://www.mail-archive.com/numpy-discussion%40scipy.org/msg26524.html


Perhaps you don't understand what "completely" means and are confusing it 
with "slightly".



-- 
Steven