Category: Research

Sorry for the long time lag. So as we are continuing our StoryStarter (Kathy is working on cleaning up the data structure, and revising the PHP code for a first public beta) we’ve been percolating an interesting idea. Once we’re done with the first beta of StoryStarters, we’ll take that code base an begin transforming it into OpenQA for distribution at VRD (feature requests very welcome).

The interesting idea? Tamal Guha, a visiting Fullbright at the Institute has started work on a multilingual version of StoryStarters. He is currently translating StoryStarters into Hindi! We’ve been kicking around the idea of creating a multi-national project to support the development of a multilingual open source digital reference system. The idea would be to emulate the quality study we did with Chuck McClure and Melissa Gross where we identify the project (developing OpenQA) and then seek sponsors for the project. International organizations could join at various levels (right now we are thinking as a function of size and what how much say in the direction of the project). Join at one level and you get a hold of the software and source. Next level you pay for a language adaptation. Next level you drive the software features (we want chat, etc).

This is still fuzzy, but it seems to resonate with folks here. The end result would be an open source, low cost or free digital reference packaged built for multinationals. It would also focus on building software, not consortia or content, hopefully by-passing legal issues.

If you were scratching your head over that last post about IKE (the Inductive Knowledge Engine) and that clustering, fear not. Here is an animation showing IKE in action. The dots you see scattered about are each Story Starter responses. They have been randomly scattered around a 100 x 100 graph. Each frame of the animation has a dot compare itself to another dot and then move (so one dot eventually compares itself to all other dots), then the step is repeated with the next dot. Each frame represents one full cycle of comparison. In each comparison the dots are getting closer to each other if they come from the same blog, and farther away if they are from different blogs. The clustering is dynamic. Entries from the same blog end up clustered together.
THis example is pretty simplistic as it is based on a single static variable (blog title). However, I wanted to give you an idea how IKE was working.

Sorry I have been quiet for a while. I’ve been very busy preparing for a MacArthur presentation next week. I’ve been building the Story Starters site…and having a fun time in source code. The basics of the site are done and almost ready for public play (on the “reference implementation at least). We’ll need to do some re-writing and cleaning up (not good to do database design while you are also figuring out the feature set).
It is also becoming clear to me how Story Starters is going to become the next version of QABuilder – OpenQA.

I hope to have a site up for folks to look at end of next week.

Well, yesterday I finished up a working demo site of Story Starters for the National Science Digital Library. Doing that and teaching kept me too busy to blog. My plan is to make a more general and scalable demo for the world. Hopefully I can have that up by the end of the week. My intention is to also tie in Reference Extract.

Right now in Reference Extract you can search digital reference knowledge-bases and then extend your search to cited references.Like this:

DR Archives –> Cited Searches

No imagine, chaining more specific searches together, so now we can add blogs in Story Starters:

Blogs –> DR Archives –> Cited Searches

Making this capability generic, a local library could add their archives, and then extend to Reference Extract:

My Library –> DR Archives –> Cited Search

And one could continue to add more links:

My Library –> DR Archive –> Cited Search –> Google

The idea would be to make a general utility to tie these searches together in a seamless interface and allow any interface mix and match (and collapse) these chains based on a given user interface. So a library could go:

My Library –> Cited Search –> Google

etc.

Just an idea. Off to start the Story Starter site.

We’re in the process of developing a service called “Story Starters.” It is sort of a fun idea that grew out of an immediate need. The immediate need was a way to connect Blogs with our QABuilder software. Where we’ve ended up (for now) is a general blog starting service. A person or organization posts some list of items they want to be blogged about, and from within blogging software (WordPress to start with), can browse lists and pick the ones they want to blog about. The resulting blog entries are then aggregated and sent back to the posting agent…read more to see why this might be useful.
Continue reading “Story Starters”

So as I mentioned I’m working on linking blogging and digital reference, and trying to figure out how to pass messages between the two without doing a major re-work on QABuilder. Then it hit me (I’m slow sometimes)…this is what we built NISO AZ to do, and why we integrated it into QABuilder. It should be a simple matter to get questions and answers into QABuilder using NISO AZ.
Continue reading “NISO Rises from the Ashes”

We’re busy working with WordPress to link it to QABuilder, our digital reference research package we developed for the NSF’s National Science Digital Library. It is looking quite promising. In particular we’re looking at using WordPress plugins. I also think XML-RPC has some promise to do this as well.
Continue reading “Linking Blogging and Digital Reference”