Tomasz Pik wrote:
> robert burrell donkin wrote:
>
>>
>> On Monday, May 12, 2003, at 08:58 PM, Tomasz Pik wrote:
>>
>>> robert burrell donkin wrote:
>>>
>>>> i've committed some initial code submitted by phil so there's now
>>>> something to start arguing about :)
>>>
>>>
>>>
>>> One thought - sooner or later somebody wants to pass
>>> matrix through RMI or bind as an attribute to session
>>> in servlet.. Please, make implementations Serializable.
>>
>>
>>
>> +1
>>
>> as a matter of design policy, we could make all interfaces extend
>> serializable or we could just make the implementations serializable.
>> opinions anyone?
>
>
> I'd like to say: interfaces. But I can imagine matrix with backend
> somewhere in database (I know, rather crazy idea).
> So I think - implementation only...
>
> Tomek
>
I agree. Definitely the array-backed RealMatrixImpl should be
serializable, but not necessarily *all* math objects.
>>
>> - robert
>
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: commons-dev-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: commons-dev-help@jakarta.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: commons-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: commons-dev-help@jakarta.apache.org
|