> Anyways, based on my limited understanding of HTM, any hardware which is
less than 64-bit (due to it's larger memory addressing capability) might
not be a good fit for real-world usage.

Well, depends of your application. HTM applications related on vision
couldn't be a good fit for poor hardware however simple HTM applications
using a novel of data with one or two data-fields could be.

> May I know which version of C++ is being used to write NuPIC-core?

Currently C++11

On 23 July 2015 at 12:51, Mayuresh Kathe <[email protected]> wrote:

> Hello David,
>
> What you are saying makes sense.
> Anyways, based on my limited understanding of HTM, any hardware which is
> less than 64-bit (due to it's larger memory addressing capability) might
> not be a good fit for real-world usage.
>
> May I know which version of C++ is being used to write NuPIC-core?
>
> Thanks,
>
> ~Mayuresh
>
>
>
> On 2015-07-23 06:35 PM, David Ragazzi wrote:
>
>> Hi Mayuresh,
>>
>> This is my opinion, maybe Numenta guys have other reasons:
>>
>>  ANSI-C would have afforded better portability than C++, and would
>>>
>> have made the whole system quite lighter.
>>
>> I agree but portability is not the main issue in NuPIC (not
>> officialize Raspberry support is an example), I think that when
>> developers choose a language as C++, they tend to want an increasing
>> in performance but keeping certain level of understanding and
>> maintenance. C++ has evolved alot in order to be more readable and
>> continue light (compared to C), and because this I believe that
>> Numenta team has not interest in abandon it as soon.
>>
>>  Is it really difficult to write NuPIC-core in ANSI-C?
>>>
>>
>> Depending your expertise in NuPIC, C++, and finally Ansi-C. Personally
>> my big concern would be Ansi-C, as it requires a lot of work
>> non-related to business logic such as strings and memory manipulation
>> (which is simpler in C++), and this is tricksy to me. NuPIC core also
>> has lot of code that is written in OO that need be converted to a
>> structural approach. It seems that it's not an easy job at all.
>>
>>  If not, now that NuPIC-core is already written and supported by
>>>
>> Numenta, would there be any kind of non-commercial support possible
>> from Numenta towards re-development in ANSI-C?
>>
>> Numenta officially supports bindings for NuPIC-core such as Python
>> bindings but I'm not sure how SWIG (the binding tool) could support
>> pure C code like Ansi-C. About re-development in other languages,
>> Numenta also officially supports Htm.Java, but I don't know say that
>> they have interest in support another re-developments.
>>
>> Best regards,
>>
>> On 23 July 2015 at 03:06, Mayuresh Kathe <[email protected]> wrote:
>>
>>  Hello,
>>>
>>> May I know why the NuPIC-core is written in C++ and not in ANSI-C?
>>>
>>> ANSI-C would have afforded better portability than C++, and would
>>> have made the whole system quite lighter.
>>>
>>> Is it really difficult to write NuPIC-core in ANSI-C?
>>>
>>> If not, now that NuPIC-core is already written and supported by
>>> Numenta, would there be any kind of non-commercial support possible
>>> from Numenta towards re-development in ANSI-C?
>>>
>>> Best regards,
>>>
>>> ~Mayuresh
>>>
>>
>> --
>>
>> David Ragazzi
>>
>
>


-- 
David Ragazzi
Master in Sofware Engineering (University of Liverpool-UK)
OS community commiter at Numenta.org
--
Have you tried *NuPIC Studio*? Just check out
https://github.com/nupic-community/nupic.studio and enjoy it!
--
"I think James Connolly, the Irish revolutionary, is right when he says that
the only prophets are those who make their future. So we're not anticipating
, we're working for it."

Reply via email to