Re: Order of function load matters ??

Dave B will need to weigh in here, but I wanted to say that with
python, the functions can be listed in any order, but they have to be
listed before any main code that references them.

Here's an example:

def two(y):
   y2 = one(y) + 2
   return y2

def one(x):
   x2 = x + 1
   return x2

z = 5
z2 = one(z) + two(z)
print z2

Note that function "two" (defined first) references function "one",
but python doesn't complain about this. If you moved function "one"
and/or "two" after the main code, it would complain.

So, yes, it would be nice if you could do something similar in NCL and
not have to worry about function ordering. I don't know what the
expense would be for speed however. Arindam makes a good point on
this.

--Mary

On Sun, 28 Jun 2009, Arindam Chakraborty wrote:

> Hello All,
> probably 'load' of NCL defines all the functions in memory including their
> dependencies and therefore the functions being used have to be predefined. One
> way out could be just to list the functions, which are defined within the
> files, on to the memory while 'load'-ing (say, after syntax checking, note the
> file name and line number where the function starts). A function can be loaded
> to the memory only when it is called. This way the dependency to a function
> which is being 'load'-ed later can be avoided. A possible drawback of this
> could be decrease of performance since NCL needs to do more work during
> execution of the job.
>
> regards,
>
> arindam
> --
>
> ---------- Original Message -----------
> From: "Saji N. Hameed" <saji_at_apcc21.net>
> To: Dennis Shea <shea_at_ucar.edu>
> Cc: ncl-talk_at_ucar.edu
> Sent: Sat, 27 Jun 2009 22:57:47 +0900
> Subject: Re: Order of function load matters ??
>
>> Hi Dennis,
>>
>>> Unlike a language like (say) fortran which has a compile/link/load
>>> sequence where the linker sets up pointers to any callable
>>> functions, NCL does not.
>>
>>> I think languages like python work th same way.
>>
>> Maybe not, this order dependency is not there in Ruby and I guess
>> not in Python as well. Within a script I could load two libraries
>> lib_a and lib_b in any order, even when lib_b is dependent on lib_a
>>
>> saji
>>
>>>
>>>
>>>
>>> http://www.ncl.ucar.edu/Document/Manuals/Ref_Manual/NclStatements.shtml
>>>
>>> Jack Glendening wrote:
>>>> Using NCL 5.0.0, I have sets of functions in two files which I load
>>>> immediately one after the other ala
>>>>
>>>> load "FunctionsFile_A"
>>>> load "FunctionsFile_B"
>>>> NCL COMMANDS WHICH CALL FUNCTIONS IN THE ABOVE FILES
>>>>
>>>> I just altered one of the functions in "FucntionsFile_A" to utilize a
>>>> new function "nintspan". When I placed the new function at the top of
>>>> "FunctionsFile_A" it worked. But if I moved the new function to
>>>> "FunctionsFile_B" I got the message "fatal:Undefined identifier:
>>>> (nintspan) is undefined". So apparently the the ordering of the
>>>> function definitions matters ?! This does not make sense to me, since
>>>> one cannot always know the ordering of the function calls.
>>>>
>>>> Jack Glendening
>>>>
>>>> _______________________________________________
>>>> ncl-talk mailing list
>>>> List instructions, subscriber options, unsubscribe:
>>>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>>> _______________________________________________
>>> ncl-talk mailing list
>>> List instructions, subscriber options, unsubscribe:
>>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>>>
>>
>> --
>> Saji N. Hameed
>>
>> APEC Climate Center
>> 1463 U-dong, Haeundae-gu, +82 51 745 3951
>> BUSAN 612-020, KOREA saji_at_apcc21.net
>> Fax: +82-51-745-3999
>>
>> _______________________________________________
>> ncl-talk mailing list
>> List instructions, subscriber options, unsubscribe:
>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>>
>> --
>> This message has been scanned for viruses and
>> dangerous content by MailScanner, and is
>> believed to be clean.
> ------- End of Original Message -------
>
>
> --
> This message has been scanned for viruses and
> dangerous content by MailScanner, and is
> believed to be clean.
>
> _______________________________________________
> ncl-talk mailing list
> List instructions, subscriber options, unsubscribe:
> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>
_______________________________________________
ncl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
Received on Mon Jun 29 2009 - 07:49:02 MDT