Hi devs,

I have completed the implementation of the Manager classes using JCR and
wrote test classes to cover basic operations as well.
I have attached the final patch to [0].

Your feedback is highly appreciated.

[0] ManagerImpl_implementations_v6.patch
https://issues.apache.org/jira/browse/PHOTARK-72

Thanks,
~Umashanthi

On Wed, Jun 15, 2011 at 8:51 PM, Umashanthi Pavalanathan <
[email protected]> wrote:

> Hi devs,
>
> I have implemented the PersonManager and RelationshipManager interfaces
> using JCR and submitted a patch [0]. I used JCR Node's STRING properties to
> implement the concept of relationship between two user profiles. I have
> written two test classes to cover the basic functionalities of the methods
> and  was able to run them with success.
>
> I would like to get your feedback on this implementation.
>
> [0] ManagerImpl_implementations_v4.patch
> https://issues.apache.org/jira/browse/PHOTARK-72
>
>
> <https://issues.apache.org/jira/browse/PHOTARK-72>Thanks,
> ~Umashanthi
>
>
> On Sat, Jun 4, 2011 at 11:24 AM, Umashanthi Pavalanathan <
> [email protected]> wrote:
>
>>
>>
>> On Sat, Jun 4, 2011 at 6:28 AM, Luciano Resende <[email protected]>wrote:
>>
>>> On Fri, Jun 3, 2011 at 4:48 PM, Umashanthi Pavalanathan
>>> <[email protected]> wrote:
>>> > Hi devs,
>>> > I am in the process of adding persistence support to the social API
>>> using
>>> > JCR.
>>> > For that first we have to decide on the node structure. I have come up
>>> with
>>> > two options (please refer attached image).
>>> > In option-1, for each username, we have
>>> profile,activity,appdata,messages
>>> > child nodes.
>>> > In option-2, under nodes people,activity,appdata,messages, we have
>>> child
>>> > node for each user.
>>> > Referring to the two options, which is more suitable in your opinion?
>>> >
>>> > Thanks,
>>> > ~Umashanthi
>>>
>>> I think if we have a good understanding of the data access pattern, it
>>> can help us decide which structure to use. For example, if we would
>>> mostly show list of activities by user, and all other data by user,
>>> I'd go with option 1.
>>>
>>
>> Yes; as I understood we would mostly access data by user and my +1 for the
>> option-1.
>>
>>
>> Thanks,
>> ~Umashanthi
>>
>>
>>
>>>
>>> --
>>>
>>> Luciano Resende
>>> http://people.apache.org/~lresende
>>> http://twitter.com/lresende1975
>>> http://lresende.blogspot.com/
>>>
>>
>>
>

Reply via email to