This is my plan for basic interfaces.

== Vector ==

size()

get(int index)
set(int index, double value)
set(Vector V)

add(int index, double value)
add(double alpha, Vector V)
add(Vector V)

dot(Vector V) instead of the current getCosine() method
norm(Vector.Norm type) : Computes the given norm of the vector

== Matrix ==

get(int row, int column)
getRows()
getColumns()
set(int row, int column, double value)
set(double alpha, Matrix B)
set(Matrix B)

add(int row, int column, double value)
add(Matrix B)
add(double alpha, Matrix B)

mult(Matrix B)

norm(Matrix.Norm type) : Computes the given norm of the matrix

-Edward

On Wed, Aug 6, 2008 at 10:55 AM, joosun yoon <[EMAIL PROTECTED]> wrote:
> I will design *Matrix and vector interfaces*.
>
> A step, simple interfaces will be designed.
>
> Thanks.
>
> 2008/8/6 Edward J. Yoon <[EMAIL PROTECTED]>
>
>> Maybe you can create a wiki page for part of interface design
>> considerations, and See HAMA-12. I'll help you.
>>
>> -Edward
>>
>> On Wed, Aug 6, 2008 at 8:48 AM, Antonio Suh <[EMAIL PROTECTED]> wrote:
>> > Good plans.
>> > I want to choose No.1, Matrix and vector interfaces.
>> >
>> > And IMO, we need more test case and comments.
>> > I see http://issues.apache.org/jira/browse/HAMA-5, comments problem may
>> be
>> > solved.
>> > However, it is better to cover that writing test case in rodmap for
>> > Hama-0.1.
>> >
>> > Hama isn't easy to use, so it would be fine to have enoungh detailed test
>> > case.
>> >
>> >
>> > 2008/8/5 Edward J. Yoon <[EMAIL PROTECTED]>
>> >
>> >> See http://wiki.apache.org/hama/Hama-0.1. If you have ideas to
>> >> contribute, please feel free to add your comments
>> >>
>> >> Thanks.
>> >> --
>> >> Best regards, Edward J. Yoon
>> >> [EMAIL PROTECTED]
>> >> http://blog.udanax.org
>> >>
>> >
>>
>>
>>
>> --
>>  Best regards, Edward J. Yoon
>> [EMAIL PROTECTED]
>> http://blog.udanax.org
>>
>
>
>
> --
> Yoon Joo Sun
> Java Developer Engineer
> Blog - http://www.ologist.co.kr
>



-- 
Best regards, Edward J. Yoon
[EMAIL PROTECTED]
http://blog.udanax.org

Reply via email to