GSoC 2012: C3 Commons - PhD project

[Shelly Grist]

Hi,

My name is Shelly Grist.  I’m a PhD student and Lecturer at University of Southern Queensland in Australia.  I’ve got a fair bit of programming experience – 9 years working in the U.S. aerospace industry developing communication and satellite test systems (TRW) and then 2 years developing computer games (Philips Media, Dreamworks Interactive).  I moved to Australia in 1999 and since then I’ve worked as a Web developer and taught courses in programming, online business, etc for the School of Information Systems at USQ.   

I'm a relative newbie to FOSS and SilverStripe.  For my PhD project I’m developing a community research and collaboration site.  I chose to use open source tools because the focus of the site is to encourage researchers in the life cycle assessment area to work together to generate the environmental data that is necessary for all.  The philosophy of the site therefore parallels the ethos of the open source community.  I chose SilverStripe because it follows web standards, is built on solid design patterns and I liked its separation of roles within the CMS.  I want to code, not interact with a GUI.

My PhD project is designed to function as a research commons – a place for researchers to communicate, collaborate and connect.  Hence I’ve named it C3 Commons.  The fundamental design includes blogs, forums, publication and data libraries, and areas for researchers to provide tutorials, video links, general knowledge, etc.   It shares some common features with sites like citeulike.org and zotero.org.  Since this is an open source project my intention is to make the code available for anyone to download and use once my PhD is complete. 

For the GSoC I’d like to focus on a few specific areas that could be developed as SilverStripe 3.0 modules.  Two possibilities are an updated Multi Blog Holder module and my Publications Library.  I’ve included a blog module because the current Blog module doesn’t seem to provide the functionality that I need (see question on Blog Module forum 13 March).  I’m not quite sure how the third area, Social Network Elements, could be modularized since I think it will be too site specific but I’m very open to suggestions.

Here’s an outline of my ideas for features and functionality.

Multi Blog: (not started)

·       All members have a blog created for them when they register.

·       Blog entries only editable by member author and admin members with BLOG_MGMT permissions.

·       Member’s ‘Your Space’ only displays member’s blog entries.

·       Community Space displays ALL member blogs ordered/filtered by member name, tag, or search word.

·       Blog Entries can have a pic, title and tags.

·       All member blog entries viewable and searchable.

 

Publications: (under development)

·       All members have a publication library created for them when they register.

·       Publications can be uploaded in RIS format to member’s library, community library or both.

·       Members can add reviews, tags, ask questions or provide answers to any publication in the community library or their personal library.  All additional publication metacontent added to the community library is viewable and searchable by all members.  Metacontent added to a member’s library is kept private.

·       Members can add usefulness ratings to a publication’s metacontent added by other members.

 

Social Network Elements: (under development)

·       Publication tag clouds and member clouds – typical tag style cloud to show number of publications with matching tags, and members with similar publications in their library.

·       Recommendation system for:

o    publications,

o   potential members with whom to collaborate

·       Social Network Analysis map displaying publication/member/author/tag association/relationships using Two-mode networks.

 

Is this a feasible project for GSoC? Any input or suggestions would be very welcome!

 

Thanks, Shelly

[xeraa]

Hi Shelly,

While I like your project ideas, there are 2 questions / problems I can see (I'm pretty outspoken here, don't take this personal):

1) Your modules seem to be very specific. How can you make this relevant to a broad user basis? The number of positions Google grants are limited, so we'll select the projects with the biggest impact for SilverStripe and its users.

2) You're already working on your PhD and have stated, that you want to open source the results once you're finished. So what are we gaining from your project if you're doing and releasing it in any case?

Cheers,

Philipp

[Shelly Grist]

Hi Philip,

No offense at all.  I definitely understand your points.  In response:

1) I think my spec for the Muti Blog module could have a broad user base.  Functionality wise anyone with BLOG_MGMT/ADMIN permission would be able to post to it, which I believe is what the current version provides.  What it would also provide is the ability to restrict postings to the Member Author.  The differences lie more in the implementation.  I envision a site with only a few moderators with BLOG_MGMT permissions, but with each member having their own personal Blog.  Implementation wise it may be possible to also support the current functionality.
 
   The Publications module does have some very specific requirements, but at a higher level its basically allowing people to add a DataObject to a public/private repository and add metacontent.  Members can then tag it, review it, ask or answer questions about it.  The DataObject could be a Publication (my specific use case) or it could be a Movie, a Book, an Event or a Restaurant.  The particulars lie in the specific use case.  I think it would be possible to create a generic module which users could then expand upon (with good instructions and an example or two).

2) As you say, since I'm planning on doing this work anyway and making it open source why apply for GSoC?  I'll be equally open and say my primary motivation was the opportunity to have a mentor.

Despite having worked through Ingo and Steven's book and yours (I'm assuming) I'm struggling a bit with some of my architecture choices.  Most of my site doesn't seem to fit within the examples from the books or the tutorials. Or maybe I'm just not looking at it the right way.  That's certainly possible.

Some of SSBits tutorials - DataObjects instead of Pages - seem to apply but I'd really like to have someone to bounce my questions off of.  I posted to the Blog Module forum and sent an email but haven't received a response yet. 

I understand that the GSoC positions are limited.  If someone was willing to mentor me (without short changing their GSoC student, of course) I'd be happy to match the Mentor fee that Google pays.

Thanks, Shelly

[Shelly Grist]

To All,

Just to clarify, my last comment was meant as an offer to someone to Mentor me OUTSIDE of the GSoC program.  Not in anyway as an inducement to include me within the GSoC program.  I realized after posting that it could be misinterpreted.

Thanks, Shelly

[xeraa]

Ok, now that makes a lot more sense.

IMHO most people aren't mentoring because of the money.

However, I'd assume that you'll get help if people see that you're seriously trying to achieve something.

I can't really say too much about the blog module as I've never really used it. I assume you'll want to take the existing module and enable multi-tenancy (or whatever that should be called). So it should provide a subset of the features of Tumblr or similar platforms?

This is pretty fundamental stuff, so it probably fits best into the mailing list - at least the underlying approach. Maybe you want to add some information to your initial mail so that people will look at it again?

Publications sound like some interconnected DataObjects, but that should be doable. For a powerful search I'd take a look at the Solr module. And you're right, the books are a bit light on complex DO examples, but that's already a pretty advanced topic (especially for my beginner book).

Are you facing any specific problems? But this should probably discussed in the forum rather than on the dev mailing list.

For the recommendation / network analysis system, I'd assume you'll need to finish the previous point to build upon that. Once you've established that, you can try to retrieve the relevant information from it - which is probably betty hard, but rather SilverStripe independent.

Once you've built the underlying model for that, implementation shouldn't be too hard (ask in the forum if you're having any issues). However, if you're having a lot of data, this might get interesting performance wise...

Cheers,

Philipp

PS: Out of interest: What's the major difference between your project an http://www.researchgate.net? They are already connection publications and social network features.

[Shelly Grist]

Hi Philip,

Thanks for the pointers.  I'll have a look at the Solr module.  Regarding the DO issues, right now I'm looking at using the DataObjectAsPage module.  I've already built some of the Publications functionality based on earlier tutorials but I'm considering starting over with the new module because I think it will be a simpler more complete solution.

Regarding Research Gate, besides the sheer scale of what they offer I guess the main design difference is that I'd like C3 Commons to follow the Moodle http://moodle.org/about/ model.  Moodle is an open source CMS "popular among educators around the world as a tool for creating online dynamic web sites for their students".  Anyone can download the code and host it on a server.  Additionally, many institutions also create additional modules which are then available to others.

Please don't get me wrong, I don't have delusions of grandeur here.  I'm not saying I can duplicate the success of Moodle.  I just want to design a tool that follows the same design philosophy of creating something for a community that can be adapted by the community as their needs evolve and change.  Part of my literature review for my PhD included Alexander's 'Design Patterns', Brand's 'How Buildinsg Learn: What Happens After They're Built' and early blogs on Social Software.  The main theme is adaptive design.   And I'll stop blathering now...

Thanks, Shelly

P.S. The offer to pay was not so much for motivation as a recognition that some of my questions probably aren't a 2 minute exercise.  I value other peoples time and realize that I'm totally new to this community.  I'm asking for a lot without having built up any of my own Karma yet.

[Ingo Schommer]

Hey Shelly, Philipp,

I've just remembered that I untangled this "social in SilverStripe" meme

a while ago for internal marketing/roadmap purposes.

https://docs.google.com/spreadsheet/ccc?key=0AmKjIod5x5VIdDBQNEVfQnE3VVBqa003a3ZkN2FzM0E

The spreadsheet tries to separate the different functional areas of "social",

as well as collate existing modules and discussions about it.

A lot of material, and not all the modules will be perfect,

but its worth a look to avoid reinventing the wheel.

In my mind, a lot of what "social in SilverStripe" really means

is just a bit of guidance how to stitch together existing bits through a tutorial,

plus coordination between the various module authors to ensure they work well together.

On the specific GSOC idea, what I'm a bit unclear about:

Are you trying to create a platform+website (like your PHD topic suggests), or a set of reusable modules?

A well integrated platforms (consisting of many modules + custom code) has their merits,

BuddyPress (http://buddypress.org) is one successful example that was built on top of WordPress.

But its also a whole lot of work (both creation and maintenance), probably too big for a GSOC project,

so personally I'd prefer some (achievably scoped) modules.

Thanks

Ingo

[xeraa]

I think this information is outdated - the focus has moved away from thirdparty social networks to a recommendation system within the site - https://groups.google.com/forum/#!topic/silverstripe-dev/WCK_LLga8FQ

Cheers,

Philipp