Re: looping
- To: mathgroup at smc.vnet.net
- Subject: [mg106907] Re: looping
- From: Andrzej Kozlowski <akozlowski at gmail.com>
- Date: Wed, 27 Jan 2010 06:24:23 -0500 (EST)
- References: <201001210955.EAA16523@smc.vnet.net> <201001251008.FAA09381@smc.vnet.net> <hjmjja$g0$1@smc.vnet.net> <4B5FDDBD.9070503@cs.berkeley.edu>
On 27 Jan 2010, at 07:31, Richard Fateman wrote:
> Andrzej Kozlowski wrote:
>> On 25 Jan 2010, at 11:08, Leonid Shifrin wrote:
>>> Besides, changes to this seem unlikely because this
>>> is a fundamental mechanism behind scoping in Module, and it has been =
this
>>> way for a very long time (I don't know since which version but I =
wouldn't be
>>> surprised if it was there already in version 1).
>> Actually there was no Module in version 1. (e.g. see the =
documentation for Module). Module was introduced in Version 2. Version 1 =
had only Block.
>> Andrzej Kozlowski
>> (Nice tricks, by the way. I have know about some of them for a long =
time but others are new to me.)
-------
> Unfortunately, if you have several independent programs perhaps =
written by different people and they each generate persistent names via =
Module[{...,r...} ...] and you load them into the same system then THEY =
can conflict.
>
> consider
> g1[x_]:=Module[{r},If[NumberQ[A[r]],Print[r],A[r]=0]]
>
> run it a bunch of times. Save g1 and A in file f1
> start up a new system
> define
> g2[x_]:=Module[{r},If[NumberQ[A[r]],Print[r],A[r]=1]]
>
> run it a bunch of times. Save g2 and A in file f2.
> start up a new system
>
> load files f1 and f2.
>
> the behavior of g1 and g2 will be, so far as I can tell, dependent on
> the (random?) setting of some inaccessible counter in Mathematica.
>
This is very good, but still I think it falls short of the quality of an =
earlier example of why Richard would "shy away from a presentation that =
required the reader to own a free Mathematica player". I mean that he =
might want to write "a paper whose topic was the Mathematica Player =
itself".
I think if not exactly a world record, that post was a record of sorts =
for this forum. This one, while still pretty good, is not really keeping =
up with that standard.
I do hope some day instead of warning us of the possible problems that =
might be caused by his imaginative though somewhat contrived examples of =
"possible" user behaviour, Richard tries to produce an actual, "real =
life" example. Perhaps he could start by trying to generate this =
"possible conflict" by following the procedure he has so kindly produced =
as a warning to all of us. With luck this will keep him occupied for a =
while...
Andrzej Kozlowski=
- References:
- looping
- From: glenn kuhaneck <mcguyver128@yahoo.com>
- Re: Re: looping
- From: Leonid Shifrin <lshifr@gmail.com>
- looping