So there, for the rumor mill, is another feature of the CanyonLands
pre-compiler ...
Actually, the memoize function below would work for just about all functions
where you'd ever want memoizing (barring polymorphism and array parameters).
It would work even in a stack-in-use or a back script. Unlike the J function
with the "M." in the function definition, this approach lets the caller decide
whether to apply memoizing. That's plenty general -- for a solution that lacks
a problem.
Any more hints about that pre-compiler?
-- Dick
On Jun 15, 2011, at 11:12 PM, Geoff Canyon Rev wrote:
> This would work for a particular function. I was looking for a general
> solution that would allow memo-ing any function without further
> consideration. In J, for example, you just and "M." to the function
> definition, and you're done. If I were to do something like this in LC I'd
> want to implement something similarly robust and simple to use.
>
> As to why I want to: no particular reason, other than memo seems vaguely
> possible in LC, and a useful thing to have. I like pushing the language
> forward.
>
> gc
>
> On Tue, Jun 14, 2011 at 4:27 PM, Dick Kriesel <dick.krie...@mail.com> wrote:
>
>> Hi, Geoff. Here's a way to memoize without parsing the functions'
>> arguments, so commas won't be a pain.
>>
>> local sMemo
>> function memoize pFunctionName, pParam1, pParam2
>> local tDigest
>> put md5( the params ) into tDigest
>> if tDigest is among the keys of sMemo then
>> get sMemo[ tDigest ]
>> else
>> get value( pFunctionName & "(" & item 2 to -1 of the params )
>> put it into sMemo[ tDigest ]
>> end if
>> return it
>> end memoize
>>
>> Calling MD5 makes the array's keys shorter. IIRC, that's good, although
>> when I tried to find evidence about it in the current docs, I couldn't find
>> it.
>>
>> Does this help?
>>
>> -- Dick
>>
>>
>> On Jun 11, 2011, at 2:23 PM, Geoff Canyon Rev wrote:
>>
>>> http://en.wikipedia.org/wiki/Memoization
>>>
>>> I'm wondering if anyone has experimented with creating an automatic
>>> memoization function within LC? Given that we can access argument lists,
>> and
>>> have "do," it seems possible, but I'm betting commas in the arguments
>> would
>>> be a pain. Something like this:
>>>
>>> function m functionName
>>> local callHistory,callHistoryValue
>>>
>>> -- parse out addition arguments to m
>>> -- concatenate functionName & arguments into functionCallString
>>>
>>> if callHistory[functionCallString] is not empty then
>>> return callHistoryValue[functionCallString]
>>> end if
>>>
>>> -- get the function result
>>>
>>> put it into callHistoryValue[functionCallString]
>>> put 1 into callHistory[functionCallString]
>>>
>>> return it
>>>
>>> end m
>>> _______________________________________________
>>> use-livecode mailing list
>>> use-livecode@lists.runrev.com
>>> Please visit this url to subscribe, unsubscribe and manage your
>> subscription preferences:
>>> http://lists.runrev.com/mailman/listinfo/use-livecode
>>
>>
>>
>> _______________________________________________
>> use-livecode mailing list
>> use-livecode@lists.runrev.com
>> Please visit this url to subscribe, unsubscribe and manage your
>> subscription preferences:
>> http://lists.runrev.com/mailman/listinfo/use-livecode
>>
> _______________________________________________
> use-livecode mailing list
> use-livecode@lists.runrev.com
> Please visit this url to subscribe, unsubscribe and manage your subscription
> preferences:
> http://lists.runrev.com/mailman/listinfo/use-livecode
_______________________________________________
use-livecode mailing list
use-livecode@lists.runrev.com
Please visit this url to subscribe, unsubscribe and manage your subscription
preferences:
http://lists.runrev.com/mailman/listinfo/use-livecode
