[asterisk-dev] Regarding Project in Google Summer of Code

Akshay Bhardwaj bhardwaj.akshay07 at gmail.com
Wed Mar 24 09:04:41 CDT 2010


*Sir,*

*AIM: Implementation of astobj2 container using a sorted data structure.*

I just went over the code and documentation. *I think we can do following
things in the project :*

*Operations that should be present corresponding to new container (*For
reference I have used the name of operations for present astobj2 container)*:
*

1. There would be *Container definition* for new astobj2.
2. I will have to *make corresponding constructor and destructor for
allocating memory and then making it free* *respectively*.
3. *ao2_container_alloc(args)* -> that allocates memory for the data and
stores it, in a binary search tree format. We can also consider using a
btree if the amount of data storage is large.
4. *ao2_find (args)* -> that traverse through the binary tree. Over here i
think that Binary tree would score over the hash tables.
5. *ao2_callback(args) *-> This would include features like unlinking,
searching multiple entries etc.
6. Also I will have to *include getter and setters* *for other data
types*so that data could be easily stored in other data types when
necessary. I
would go through entire code again to check what all measures should be
taken.

*Please correct me if I am going in wrong direction.* Also *I will like to
ask whether my suggestion of using the binary search tree is correct?* I was
thinking over btree as well but it is generally used for large data storage.
So could you tell me what will be limit of data allocation?

Please try to correct me If I would have misunderstood problem and would
have taken wrong decision regarding the data structure




On Tue, Mar 23, 2010 at 4:08 AM, Akshay Bhardwaj <
bhardwaj.akshay07 at gmail.com> wrote:

> Sir,
>
> *I would prefer the first option, where aim is to include implementation
> of another data structure. *
>
> I think that *Binary Search Tree would be good option for sorted data
> structure implementation *also this would make implementation easier. *I
> think my next aim would be to study what all container types are their and
> what connectors(support) I have to write. *
>
> But going to* negative effects of implementation of Binary Search Tree*. *It
> would make execution slow as compared to hash function*. This could be an
> bottleneck that we ought to come over.
>
>
>
> On Mon, Mar 22, 2010 at 9:48 PM, Russell Bryant <russell at digium.com>wrote:
>
>> On 03/19/2010 03:45 PM, Akshay Bhardwaj wrote:
>>
>>> Sir,
>>>
>>> I am B.Tech. 3rd year student from Indian Institute of Technology
>>> Roorkee. I
>>> had worked on asterisk and its integration with sphinx and hindi ASR with
>>> ex-professor Randolph Wang, Princeton university. This project was done
>>> for
>>> Digital Study hall an organization for poor people in India. I think I
>>> have
>>> sufficient background knowledge of Asterisk and could contribute to it.
>>>
>>> I have special interest in two projects that you have listed on your
>>> reference site:
>>> 1. Improvements to the Asterisk object model
>>> 2. Improve Call Conferencing Features
>>>
>>> Also I have started to prepare a small proposal for the two projects.
>>> Could
>>> you please guide me little regarding the above projects and what would be
>>> your requirements.
>>>
>>
>> Thank you very much for your interest in Asterisk for the GSoC program.  I
>> also appreciate the insight into some of your previous Asterisk experience.
>>  Those with existing experience with Asterisk make the best candidates for
>> working on the code.
>>
>> Regarding the two projects:
>>
>> 1) For improving astobj2, I would like the proposal to include the
>> implementation of another data structure other than a hash table.  To be
>> able to implement another data structure, refactoring to support other
>> container types will be necessary.  Regarding which data structure, I'd like
>> for it to be a sorted data structure, but beyond that, I'd like to hear your
>> ideas.
>>
>> 2) For this, the best place to start is by looking at the output of "*CLI>
>> core show application MeetMe" and "*CLI> core show application ConfBridge".
>>  MeetMe has a lot more features than ConfBridge.  I would like to see a
>> proposal for implementing some of these features in ConfBridge.  If you have
>> any detailed questions, don't hesitate to ask.
>>
>> Also, as we start discussing the details of your proposal, we should
>> probably move our discussion to the asterisk-dev mailing list.  That way,
>> more people can provide feedback and help improve the chances of success.
>>
>> Thanks,
>>
>> --
>> Russell Bryant
>> Digium, Inc.  |  Engineering Manager, Open Source Software
>> 445 Jan Davis Drive NW   -    Huntsville, AL 35806  -  USA
>> jabber: rbryant at digium.com    -=-    skype: russell-bryant
>> www.digium.com -=- www.asterisk.org -=- blogs.asterisk.org
>>
>
>
>
> --
> Akshay Bhardwaj
> B.Tech 3rd Year
> Department Of Biotechnology
>
> Indian Institute Of Technology Roorkee
> India
>
> Mobile : +91-9319308834
> Skype : akshay.bhardwaj8907
>
>


-- 
Akshay Bhardwaj
B.Tech 3rd Year
Department Of Biotechnology
Indian Institute Of Technology Roorkee
India

Mobile : +91-9319308834
Skype : akshay.bhardwaj8907
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.digium.com/pipermail/asterisk-dev/attachments/20100324/2658e980/attachment-0001.htm 


More information about the asterisk-dev mailing list