[gephi-dev] GSoC 2011 proposals

[Sébastien Heymann]

Hello @dev,

I'm preparing our application for the Google Summer of Code 2011 (Yeah!).

http://www.google-melange.com/gsoc/

Each year, organization have to give a list of proposals students can apply directly or can be inspired of.

We created a page for this year: http://wiki.gephi.org/index.php/Google_Summer_Of_Code_2011
Last year was: http://wiki.gephi.org/index.php/Google_Summer_Of_Code_2010

Please add new topics before March 11. It is taken into account by Google when they evaluate our ability to become a mentoring organization.

Cheers,
Seb

[Sébastien Heymann]

Hi all,

I created a wiki page to "incube" official proposals:

http://wiki.gephi.org/index.php/Google_Summer_Of_Code_draft

<http://wiki.gephi.org/index.php/Google_Summer_Of_Code_draft>Cheers,
Seb

2011/3/10 Sébastien Heymann <sebastie...@gephi.org>

> Hi Yestin,
>
> I understand that you are really motivated by this topic, however we should
> keep focused on the core improvements. They are listed on the wiki, see the
> Core Evolution section: http://wiki.gephi.org/index.php/Dev_portal
>
> We won't open the wiki page because it is devoted to the official
> proposals, and would confuse people.
> Students have to follow this questionnaire:
> http://forum.gephi.org/viewtopic.php?f=9&t=74
> And openly discuss their ideas here:
> http://forum.gephi.org/viewtopic.php?f=9&t=67
> It works well as it is, even if it is hugely time-consuming.
>
> We also don't foster them to submit original proposals because we only have
> a limited amount of slots allowed by Google: 4 in 2009, 6 in 2010, probably
> 4/5 in 2011.
> Officials proposals are the priority so we warn them clearly that they
> would be put after applicants for official stuff in the queue.
>
> You can apply again as you are still a student of course. However, as
> everybody you'll have to follow the official procedure like others do.
>
> Cheers,
> Seb
>
>
> 2011/3/10 Yi Du(杜一) <duy...@gmail.com>
>
> I made a mistake,and the following email which I send seven days ago seemed
>> send to the wrong person :-P
>>
>> Hi all,
>> I have two advices:
>> 1.We could open the proposal page to all students who want to participate
>> in GSOC. So that we can collect many creative ideas and choose the best.
>> Students can edit the wiki page now, we should also add a tip on the wiki
>> and tell them they can add their own proposals.
>> 2.There is an idea list on the wiki page, we can share it with all the
>> students, which might motivate their imagination.
>> http://wiki.gephi.org/index.php/Ideas
>>
>> Telling about the ideas.
>> 1. Unfortunately, we didn't implementing the refactoring of the Edge
>> during GSOC 2010. So that we cannot develop a layout which can bend and
>> segement(For example, orthogonal layout and edge bundling). I think this
>> proposal can also be put forward. I think we should take the previous
>> application as reference and make some changes.
>> 2.Adding collaboration support to Gephi. SNA and many other kinds of graph
>> analysis are complex processes, sometimes collaboration is necessary. Why
>> cann't we add such support? Two or more users sharing a common graph, if one
>> operate on the graph, others can see the changes. We can also take full
>> advantage of the timeline. The history of many users displayed on the
>> timeline and users can turn back to any time of any person. Then start from
>> a new branch. Sounds like SVN, but support realtime and used in graph
>> analysis.
>> 3.The Gephi toolkit is a Java library, and I saw some real world
>> application such as LinkedIn Maps. We can motivate the students' imagination
>> and let them propose their own ideas on making use of Gephi toolkit. We can
>> choose one or two from them and help them to implement the amazing ideas

[Sébastien Heymann]

Hi Yestin,

I understand that you are really motivated by this topic, however we should keep focused on the core improvements. They are listed on the wiki, see the Core Evolution section: http://wiki.gephi.org/index.php/Dev_portal

We won't open the wiki page because it is devoted to the official proposals, and would confuse people.

Students have to follow this questionnaire: http://forum.gephi.org/viewtopic.php?f=9&t=74

And openly discuss their ideas here: http://forum.gephi.org/viewtopic.php?f=9&t=67

It works well as it is, even if it is hugely time-consuming.

We also don't foster them to submit original proposals because we only have a limited amount of slots allowed by Google: 4 in 2009, 6 in 2010, probably 4/5 in 2011. 

Officials proposals are the priority so we warn them clearly that they would be put after applicants for official stuff in the queue.

You can apply again as you are still a student of course. However, as everybody you'll have to follow the official procedure like others do.

Cheers,

Seb

2011/3/10 Yi Du(杜一) <duy...@gmail.com>

I made a mistake,and the following email which I send seven days ago seemed send to the wrong person :-P

Hi all,

I have two advices:

1.We could open the proposal page to all students who want to participate in GSOC. So that we can collect many creative ideas and choose the best. Students can edit the wiki page now, we should also add a tip on the wiki and tell them they can add their own proposals.

2.There is an idea list on the wiki page, we can share it with all the students, which might motivate their imagination.http://wiki.gephi.org/index.php/Ideas 

Telling about the ideas. 

1. Unfortunately, we didn't implementing the refactoring of the Edge during GSOC 2010. So that we cannot develop a layout which can bend and segement(For example, orthogonal layout and edge bundling). I think this proposal can also be put forward. I think we should take the  previous application as reference and make some changes. 

2.Adding collaboration support to Gephi. SNA and many other kinds of graph analysis are complex processes, sometimes collaboration is necessary. Why cann't we add such support? Two or more users sharing a common graph, if one operate on the graph, others can see the changes. We can also take full advantage of the timeline.  The history of many users displayed on the timeline and users can turn back to any time of any person. Then start from a new branch. Sounds like SVN, but support realtime and used in graph analysis.

3.The Gephi toolkit is a Java library, and I saw some real world application such as LinkedIn Maps. We can motivate the students' imagination and let them propose their own ideas on making use of Gephi toolkit. We can choose one or two from them and help them to implement the amazing ideas. This is not the core of Gephi, I don't know whether it can be selected as a formal project.

I also have a question, can I apply for the project again as a student? I think I have some advantages on some of the proposals.

On Thu, Mar 10, 2011 at 10:41 AM, Yi Du(杜一) <duy...@gmail.com> wrote:

Hi Seb,

  Collaboration is an interesting feature. Just as you said, there are not enough materials on how to build collaboration environment. But what they have already told us is what kinds of feature such environment has, and what it looks like. You words inspired me, there are two kinds of collaboration: asynchronous and synchronous(The CoMotion.swf is a synchronous collaboration environment). In my point of view, asynchronous collaboration environment is easier to build than synchronous one. A paper(see attachement) by Jeffrey Herr, author of Prefuse, told us some considerations and prototypes of asynchronous collaboration.  Most of the prototypes are web based system, maybe some students are good at building such system, and a web based Gephi with collaboration is implemented during this summer :-)  What I wrote in the proposal page is synchronous way, maybe a asynchronous way is also feasible. In fact, I found it much harder and complex than I thought before during writing those words.

  The Edge Refactoring is similar to the Preview Refactoring. The latter proposal wants "execute algorithm before rendering", so edge bundling can be done during this process. There are some difference, that's what I said "orthogonal layout". The difference is that edge bundling needn't relocate the nodes, but orthogonal layout has to relocate the nodes. But if preview refactoring is done, we can first layout the node in the workspace, and then curve the edge in the preview window to implement orthogonal layout. So merging the two proposals are OK.

On Thu, Mar 10, 2011 at 1:33 AM, Sébastien Heymann <sebastie...@gephi.org> wrote:

Hi,

Let me sum up where we are with the proposals.

I have read a paper on collaborative support for data analysis. This is an exciting challenge, but the hard questions remains unanswered by researchers on how to build efficient collaborative tools. At the moment, great examples runs only as platforms on the web. That's why I prefer to wait for maturity in this field before going in this direction.

The Edge Refactoring can be included (is similar to?) the Preview Refactoring I guess :-)

The Maven-based build system is still in question. It would add another layer to our production chain, and it is not clear what the benefits could be compared to a set of scripts doing current manual tasks automatically. Especially I'm afraid of the complexity of this new tool, the time needed to handle it, and how it would replace ANT tasks in an easier way.

The WebGL proposal is also not focused on Gephi itself but would tend to create another specific project. It is however a priority because WebGL 1.0 specs are published and now implemented by most of modern browsers, and mostly because it is a huge need by both research and industry.

A Timeline proposal will probably be added. Other proposals are validated.

Seb

2011/3/2 Cezary Bartosiak <cezary.b...@gmail.com>

Hi,

Currently I have no time to describe it, but the are 2 things that are very crucial to me (and I'm not sure they are "big enough" to be proposals), so I'll mention them:

1) We definitely need some abstraction layer for the Reports. I'm currently working on a project of a simulator thats makes it possible to see how some phenomenons spread in networks (information, viruses etc.). I must report many things and this poor mechanism definitely makes me angry :P BTW, is it connected with the browser idea?

2) Generators API needs to be extended. It would be good to be able to use the same (or similar) graphs API as in other modules. I mean drafts are too simple. It's not so simple to generate more complicated networks with the current API. Moreover I would like to count some statistics during network evolution. Now it is possible but very painful, because I need to make a "special" generator that attributes time intervals to each node/edge and after the generation I must fire the proper statistics for each graph visible (for every step of the evolution). It would be fine to have this automated.

_______________________________________________
gephi-dev mailing list
geph...@lists.gephi.org
http://gephi.org/mailman/listinfo/gephi-dev

_______________________________________________
gephi-dev mailing list
geph...@lists.gephi.org
http://gephi.org/mailman/listinfo/gephi-dev

--
杜一 Yi Du (Yestin) 
IEL, Chinese Academy of Sciences, Institute of Software.

TEL: +86-15810134970
email&gtalk: duy...@gmail.com

--
杜一 Yi Du (Yestin) 
IEL, Chinese Academy of Sciences, Institute of Software.

TEL: +86-15810134970
email&gtalk: duy...@gmail.com

_______________________________________________
gephi-dev mailing list
geph...@lists.gephi.org
http://gephi.org/mailman/listinfo/gephi-dev

[Cezary Bartosiak]

Hi,

Currently I have no time to describe it, but the are 2 things that are very crucial to me (and I'm not sure they are "big enough" to be proposals), so I'll mention them:

1) We definitely need some abstraction layer for the Reports. I'm currently working on a project of a simulator thats makes it possible to see how some phenomenons spread in networks (information, viruses etc.). I must report many things and this poor mechanism definitely makes me angry :P BTW, is it connected with the browser idea?

2) Generators API needs to be extended. It would be good to be able to use the same (or similar) graphs API as in other modules. I mean drafts are too simple. It's not so simple to generate more complicated networks with the current API. Moreover I would like to count some statistics during network evolution. Now it is possible but very painful, because I need to make a "special" generator that attributes time intervals to each node/edge and after the generation I must fire the proper statistics for each graph visible (for every step of the evolution). It would be fine to have this automated.

On 2 March 2011 19:21, Sébastien Heymann <sebastie...@gephi.org> wrote:

[Sébastien Heymann]

Hi,

Let me sum up where we are with the proposals.

I have read a paper on collaborative support for data analysis. This is an
exciting challenge, but the hard questions remains unanswered by researchers
on how to build efficient collaborative tools. At the moment, great examples
runs only as platforms on the web. That's why I prefer to wait for maturity
in this field before going in this direction.

The Edge Refactoring can be included (is similar to?) the Preview
Refactoring I guess :-)

The Maven-based build system is still in question. It would add another
layer to our production chain, and it is not clear what the benefits could
be compared to a set of scripts doing current manual tasks automatically.
Especially I'm afraid of the complexity of this new tool, the time needed to
handle it, and how it would replace ANT tasks in an easier way.

The WebGL proposal is also not focused on Gephi itself but would tend to
create another specific project. It is however a priority because WebGL 1.0
specs are published and now implemented by most of modern browsers, and
mostly because it is a huge need by both research and industry.

A Timeline proposal will probably be added. Other proposals are validated.

Seb

2011/3/2 Cezary Bartosiak <cezary.b...@gmail.com>