Navigation:
Documentation
Archive



Page Tree:

Child pages
  • January 2011 Meeting - Work Area Break Out Session Day 2 Notes and Report Outs

This wiki space contains archival documentation of Project Bamboo, April 2008 - March 2013.

Skip to end of metadata
Go to start of metadata

Upload presentations and meeting notes about Work Area Break Out Session Day 2 Notes and Report Outs to this page.

Work Spaces

Collections Interoperability

Corpora Space

Scholarly Services and Platform

Places-Texts

(Steve gave brief overview; cf. Places-Texts page)

Morphology & Syntactic

  • storing result sets
  • storing indexes of pointers to results of previously processed texts
  • dependencies on
    • external morph services
    • text repositories – CTS ref index service
  • interface for service: this quarter
  • BSP dependecies:
    • storage services (platform or elsewhere), notification services
    • person services to associate fed-back corrections to the underlying morphological engine to attribute contributions

Philologic

  • 5 services (cf. Oren's documentation, to be provided)
  • Oren: conceptually, a batch service
  • Bridget: output from Morph as input to Philologic? Oren: good use case to write up
  • Q: SLA re: storage of indexes (R: indexes want to live and run for months)

Characterizing Bamboo Platform hosted services

  • AuthN
    • Tufts: who are they is more important than metering
    • PL has access concerns re: license-controlled materials ... need then to know what ARTFL needs vis-a-vis AuthN, level of trust, who is the audience of users who needs to be Authenticated.
    • Oren: walk before running. initial collections of texts and AuthN that might apply to those ... will think about use cases

Notes on final discussion of this session:

Seth's notes

[this is Seth Denbo's rough notes which might be useful for the discussion of principles this afternoon - please edit if anything is misrepresented]

Bob Taylor - the need to get back to texts is crucial - this should be a principle for building tools - will differentiate Bamboo

Tim Cole - benefits - break down into operations, understand the impact of differences in documents on this, discover what will be necessary to do to the objects to make the use of the tool possible.

Robin Valenza - lots of scholars do not know that this is what they want to do, but once you show scholars that they can do this with their texts. we will get pictures out of it which can be used for showing off what Bamboo will make possible

Martin Mueller  - How far ahead of your users do you want to be?

Martin Wynne - it's good to have the things that are a bit ahead of the user, but the more mundane are also important

John Pybus - disconnect between web services and the UI interfaces. Scholars need good UIs that are clear to work with and we aren't thinking about how to make this possible to enough of an extent.

Tim Cole - Role of Work spaces in making JP's issue

Dave Lester - If a service is a tool without a UI then do the tools have to always work through the WS or can they interact directly with services.

Steve Masover - Bamboo could potentially make it possible for tools to be used within Bamboo or outside in their own environments.

Steve's notes

Work Spaces

(Bruce reporting out)

Success is demonstration of end to end use of the Bamboo ecosystem in the Work Space environment.

Unrealistic to expect highly sophisticated tools in work spaces, but what we make must be of credible value.

Audience for Work Spaces:

  1. Scholars who will use the tools provided; and
  2. Developers who will build tools for these environments.

Here's the challenge: WS has been about providing a platform into which tools are integrated; yet we need for success that there be tools that can be integrated, to demonstrate the value of this proposition. "We need demonstrator tools that we can deploy."

If we can pull off OpenSocial integration, that provides an avenue ...

So the question includes, which tools ought to be brought into play, where do we get them, how do we deploy them, how will we integrate tools that are deployed outside the Work Spaces (whether on BSP or elsewhere).

Scholarly use cases must drive and undergird whatever story we tell about Bamboo generally and Work Spaces in particular.

Discussion

Neil Fraistat: Propose that we create principles for tool selection, as we did for collections yesterday. We want to be able to explain the principles by which we came to our selection decisions, this is important to tell the world. Secondarily, we need to talk about where in the ecosystem tools ought to exist. Articulating a clear explanation to that question is key to CS work.

Jim Muehlenberg: HUBzero – can integrate tools – work from principle of integrating tools that support our initial, local installations.

Rick Jaffe: BSP required to authenticate and form Bamboo Groups. Need to abstract AuthN.

Steve Masover: On our roadmap. We need to discuss timing.

Collections Interoperability

(Tim Cole reporting out)

CI management – CMIS connectors into collections we discussed yesterday. Question: timeline requirements, esp. for Hathi and TCP (we thought those would be collections to start with); and, where do these services sit in the Bamboo Ecosystem – Work Spaces directly? BSP and called by Work Spaces? We will focus on those questions early.

Firm up CI interoperability standards. Access, metadata – tease out the difference between these, and how far we're going to get in defining which and how deeply we plan to commit to standards in each of these areas.

Collection Descriptions: would like to exploit automated techniques developing (esp. @ ANU) to do this; manual collection descriptions/profiles may be the place we start, but with an eye toward applying automated techniques more widely as we get traction with this effort.

Operations on content: E.g., what kinds of transformations are needed to prepare collection materials for application to services/tools that are in Bamboo's area of focus.

Phase Two oriented inquiry (the survey Harriet Green spoke about yesterday).

TSR: we think this should include some collection information, we're working on the assumption that some RDF description will be our approach. However, we'll need to discuss with TSR team how they are modeling collection descriptions.

Discussion

Bob Taylor: TCP texts are probably going to be our first opportunity to work with Work Spaces to demonstrate the end-to-end ecosystem functionality in Work Spaces. At the same time, discussion will continue re: how Hathi Trust expects / requires communities and technologies like Bamboo and its products to engage with its materials. It is key for David Greenbaum and Berkeley team (at PI level) to move forward with formal engagement with Hathi Trust communities.

Neil Fraistat: Urgency of Hathi engagement for Corpora Camp.

Tim Cole: Can we use the API on the public domain documents for Corpora Camp. Hathi will also be a very good place to bring services to.

Scholarly Services
Steve - Places & Text

Text mining for place using a GeoParser; hooking up Gazeteers; using this linking to do things like map or discover artifacts in material culture related to texts of interest.

Cf. Scholarly Services - Places in Texts

Bridget - Morph & Syntactic services

Morphology and Syntactic Analysis services.

Cf. Scholarly Services - Linguistic Analysis ... a less technical explanation is coming soon.

Question from Martin Wynne: New morphological services?
Bridget: No, not rewriting Morpheus, but providing a toolkit for interacting with Morpheus. Mediated access in a standardized way. To start, Morpheus for Greek; Buck Walter for Arabic to demonstrate diversity; Latin – not sure which; open to others.
Tim Cole: Assumes some architecture of BSP ...

Oren - Philologic services

Three services available on BSP:

  • Concordance – user supplies criteria and gets back a list of occurrences of contextualized of words or phrases in text
  • Frequency – frequently occuring words in text
  • Colocation – building on frequency, here are words that occur most frequently in context with a set of frequently occuring words in a text

Martin Wynne: Where will services be deployed and results stored?

Oren: Much discussion to be had. RESTful proxies to ARTFL services running at Chicago. Details to be worked out.

Martin Mueller: Passionate request for another service ... post processing result sets that are larger than a dozen – extraordinarily helpful to the user.

Steve Masover: faceted browsing on arbitrary search results is a great thing to aim for, but perhaps difficult to implement given the need for an ontology of categories to present to a user (what's an ontology for arbitrary everything)

Marlita Kahn: Can we make a statement about what a user's experience is in the Bamboo ecosystem. Are we ready to make a statement about a collection, a service, an operation, in an environment (Work Space)

Neil Fraistat: Principles on selection of tools might be a better precursor to a discussion like that.

Tim Cole: Common thread among tools presented, but not necessarily obvious. PRinciples will help elucidate those ties. Also, consider that an individual user won't come to Bamboo for general reasons, but will come to work on particular materials or with particular functions that are of particular interest to that user/scholar.

Marlita Kahn: Let's talk principles in the afternoon.

Corpora Space

(Neil Fraistat reporting)

Not much has changed in terms of how we describe success. Perhaps that will evolve as we get deeper in to this meeting and in our group meeting later this afternoon.

Corpora Camp an important even to evolve principles for design from the "tinkering" we will do.

In email during this discussion this morning, with Robert McDonald's help, we have gotten a commitment from Hathi for 50,000 texts. With our additional collections that we've identified as in-scope for Corpora Camp, we will implement one function across three collections.

Preface all: we don't know whether what we will do in Corpora Camp will become a part of Bamboo. We are not arguing that what we produce in Corpora Camp is part of Corpora Space – it's what we abstract from the process that is the important outcome from Corpora Camp.

Travis Brown: Functionality for Corpora Camp:

  1. must be doable in 3 days (length of camp)
  2. implement representative parts of the system
  3. of interest to community of scholars who are audience for Corpora Space

Functionality we'll look at is capturing relationships between documents in a visual or exploratory way. Methods that work across simple streams of text, or texts with very simple structure. E.g., look at three documents at the level of paragraphs, mapping them visually vis-a-vis some features (two or three choices) of texts.

Neil: Principle articulated by Martin Mueller – want to be able to get from the visualization to the underlying texts that are visualized. For every step up and away, we need to be able to take a step down and in. Perhaps this is a principle – to the extent possible given IP constraints – for CS functionality.

Travis: three decisions – these documents, this level of granularity, these document features – visualize this

Neil: Result of Corpora Camp will be the best set of principles for design going forward that we can extract out of the experience. As Robin Valenza suggested, the Corpora Camp engagement will also help to forge a team ethos that will improve our remote/virtual collaboration over the remainder of the grant period and beyond.

Robin Valenza: Faculty may not broadly know that they are interested in this functionality, but it is an appealing way – with pictures – to explain what Bamboo will be able to do to scholars who don't see what's in it for them ... yet.

Martin Wynne: How will all the pieces being articulated here tie together? There are questions to be discussed here.

John Pybus: Two ways to think about tools are happening here, and I think they're disconnected: how to use (UI) vs. how to build and deploy the capabilities expose by the UI. Missing layer – to demonstrate how the service capabilities can be applied to interfaces that scholars can get at in the Bamboo-ecosystem

Neil Fraistat: Can we say that services are tools without interfaces?

This right-side column is an artifact of "old-style" navigation. Please remove sections, columns, and content of the right-side column when this page is next edited. Cf. the Wiki navigation changes - July 2011 page for a 'how-to' on removing old-style page markup.