airavata-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Suresh Marru <sma...@apache.org>
Subject Re: GSOC 2013 Tasks
Date Wed, 27 Mar 2013 12:12:54 GMT
GSoC Students,

It is good to see the enthusiasm and your correspondence on the jira tasks. Based on last
years GSoC experience and in general to align with Airavata goals, we want to rethink the
GSoC projects a bit. Just to give you guys some context, last year gsoc projects did not yield
as much as they should have, not entirely because of the lack of interest or ability of students
but the nature of how Airavata is. 

Key goal of GSoC is to teach students how to participate in open source. Its not just about
coding, its about engaging in a community. Apache Airavata is a general purpose distributed
systems framework. But it is heavily used in scientific projects building workflow systems
and science gateways. And a good number of projects are using Airavata in production supporting
a significant numbers of scientists. These characteristics presents some challenges and also
provide some good opportunities. 

Ideal GSoC projects should be self contained and not be on a critical path of the project
so the student can work and learn freely. But the downside of this approach is the features
may never make it back to the mainstream code base. So we would like this years project to
directly work of the trunk, not in an isolated branch. That way we can ensure the contributions
are immediately incorporated into release and get used in production. But this puts a high
barrier on the quality of code to be written, more importantly well tested code. See Airavata
TDD approach [1]. 

Secondly, Airavata Architecture is still evolving, so its very tough for students to survive
through this turbulence. If you are expecting a well defined water fall model of software
engineering, Airavata is not a place. The development is truly agile and what you start working
may not be what you end up working on. Students should be able to adapt and more importantly
enjoy these challenges. If you get frustrated which changing designs, ideas and so on, then
this is certainly not a good project for you. 

Lastly, there will be a good amount of learning curve and not all design intrincies are documented.
Lot of them, you have to dig through the code and ask the right questions to understand. This
might very well expose you to dive into core of Airavata Services - Workflow Interpreter,
GFac, Registry, Messaging System. Also the Airavata API came a long way but is still primitive.
There is a lot of scope for improvement internals API's like GFac API are way mature and it
takes effort to expose all these capabilities through API

In a nut shell, if you are intending to do a GSoC project just for the sake of money and resume
addition, Airavata will not be a good fit. We need students who can survive through these
challenges and reap the benefits - seeing your code in real-world production, publish papers,
take on challenges, work with other smart people and so on. Lot of these will take up time,
so if you are very interested but will not have time to contribute in next 6 months, please
think again. 

Feel free to ask any questions, but I am trying to set the expectations right, so you will
not get disappointed. 

Cheers,
Suresh

[1] - https://cwiki.apache.org/confluence/display/AIRAVATA/Tests+in+Airavata

On Mar 23, 2013, at 8:18 AM, Suresh Marru <smarru@apache.org> wrote:

> Hi All,
> 
> Please propose GSOC 2013 ideas and label them with gsoc2013 in the JIRA. Note you have
to be a PMC member to mentor a project. But the students and other community members are more
than welcome to propose idea on the mailing list. 
> 
> I will start with few now.
> 
> Cheers,
> Suresh
> 


Mime
View raw message