[Python-Dev] Python3: speed efficiency vs user friendliness (my first experience)

[Terry Reedy]

On 3/23/2011 8:58 AM, Nick Coghlan wrote:
> On Tue, Mar 22, 2011 at 9:57 PM, anatoly techtonik<techtonik at gmail.com>  wrote:

> Python 3 actually chose *cross-platform consistency* over user
> convenience when switching away from the platform IO implementations.

Given that print acted differently on *nix and Windows, there were *two* 
choices, not just one, for consistency: the *nix way and the Windows 
way. In this case, I think the Windows way was/is better and that the 
wrong choice was made. We could and I hope can have *both* convenience 
and consistency.

The *nix choice introduced an new within-platform inconsistency, at 
least on Windows. When a program is run from an IDLE editor window, 
print to screen remaims unbuffered. (This is true on Windows, at least. 
I have no idea about *nix, and hope someone will test the code below). 
That means that I can develop a program like this:

import time

for c in 'Similated 10 cps teletype output':
     print(c,end='')
     time.sleep(.1)

print()

run it, see that it works, and ship it. But apparently, is will not work 
even for Windows users who run it 'normally'.

I would prefer that IDLE not be degraded and am not sure it could be.

>Users may *say* they prefer
> convenience over speed, but that's only true until the lack of speed
> becomes intolerably slow.

Could speed ever really be an issue for print to screen?

-- 
Terry Jan Reedy