Table of Contents
Here are a few new ideas that we can use for directing our conversation:
- David would like us to focus on what we can demonstrate live for Mellon on December 14
- ideally include TSR
- ideally include Oxford work
- Let's ask ourselves:
- What do we have already built?
- For BSP and CI Hub infrastructure, what do have that we can demonstrate from a user/viewer point of view?
- Is there more than one type of user for whom we can demonstrate value? Do we need more than one demonstrator to do so? Take into consideration non-Bamboo partner type users
- Are there any demo ideas that are feasible BUT that are potential land mines in terms of a Mellon demonstration - we should still include these in our list, but note the dangers we perceive
The purpose of this meeting is to develop recommendations for at least one "large" demonstrator and one "small" demonstrators to create by early December. The recommendations will be reviewed by Steve Masover and by the PEG prior to implementation. We will also request that our team scholars provide feedback.
ideally we will identify additional demonstrators that we have the bandwidth to produce during this time frame.
We also will probably be able to identify demonstrators that can be produced in Q1 2012.
Another outcome of this meeting will be to reach agreement regarding roles and responsibilities for creating the demonstrations. We will not present recommendations to Steve or the PEG without having identified who will do the work.
Context and Considerations
The purpose of the demonstrators as well as the process and decision making criteria are described on the parent page to this one and the demonstrator recommendations are based on ideas generated at the July Partner meeting in Evanston and subsequently expanded and refined by team members. These are also available on the parent page.
Additionally the Mellon Foundation has requested to view a demonstration of our work to date during our meeting with them mid-December. Therefore it would be ideal if we could identify demonstrators that would serve this purpose as well as our general audience communication purpose. However, we will not try to "force fit" the two types of demonstrators.
Marlita Kahn, Fernando Alvarez, Noah Wittman (UCB)
Emma Millon, Travis Brown (UMD)
Bruce Barton (Wisconsin)
Bridget Almas (Tufts)
Tim Cole, Mike Tang, Loretta Auvil (UIUC)
Bill Parod (Northwestern)
Quinn Dombrowski (Chicago)
- Review goals, context and considerations
- Review and discuss demonstrator ideas
- Agree on demonstrator recommendations
- Agree on roles and responsibilities for creating the demonstrators
Meeting Notes and Decisions
- What do we have already? What can we readily build upon? Looking to demonstrate "end to end".*
Starting with Zotero file, go to CI Hub to retrieve an object via BSP calls/service and storing the object back into work space in local object store (Fedora instance).
- need attachment of provenance meta data with any object insertion into object store
- if transformations: need source object
- we have the piece working re Zotero book file, deposit into CI Hub and can view what HUB exposes of that object
- doing the Fedora integration now, unclear how much of that story available in Dec but expect to have the basic, and hopefully marshals the object from the store and applies the tools to it
- even if don't have storage or re-ingest pieces working, we have a lot - we are marshaling objects using CMIS means we can marshal objects via CI HUB
- We have lots of services, even if not on Platform yet
- we can use Bridget's services that are on the BSP
- tool is exposed in the workspace and user doesn't experience where the service exists
- it is simple to build a UI to Bridget's services
- given stress on curatorial activities should we look at something that is specific re curation
- an Indiana tool - an annotation repository, open annotation model that's been working on for a while, Mellon funded
- what kind of objects can Bridget's annotations service work on?
- this a safe opportunity let's peruse that and see if Tim's idea can bear fruit in time, Mike Sang will work on and if nothing else can use for public demo
- Travis: another curation thread they've been doing; Martin's problem of text that has gaps in it, Martin has built lexicon about how the words can be corrected and gives recommendations,
- Bill Parod already has something that identifies gaps but not exposed via CMIS yet
- by early December Bridget can get a template for showing the annotations work in (need more information)
- TBD if there is bandwidth
- MK: confident we’ll have something that uses HUBzero that uses CI HUB. Do we need to know which repository we’re going to retrieve from? Can we retrieve from more than 1? Is there another tool/service in addition to the annotation service that we can commit to?
- Travis: what does the user actually see? XML w/ Morphadorner tags or some other visualization? Are the texts being run on the fly through Morphadorner on the BSP?
- Bridget: no, those are morphological services not Morphadorned.
- Fernando: the BSP isn’t doing any TEI-related work.
- Tim: this is pre-process.
- Bridget: we can use morphology service.
- Tim: Woodchipper could work on TCP & some Perseus.
- Morphological services work on Perseus & maybe TCP texts.
*There is a dependency between tools + collections.
- What does user see if we use Morphadorner?
- A tool that does some visualization is a key
- Phil Burns skeptical about using Morphadorner because of the time it takes to process
or stored as morphadorned objects on the BSP - they are stored in the TCP repository in Illinois
- there are morphologic tools on the BSP
- Collections available:
- Hathi Trust CI adapter not up to latest BookModel
- Woodchipper could be on TCP and some Perseus
- Morphological tools could be on TCP
- Morphadorned could be on TCP
- i.e. dependency between tools and collections
- Quinn: Philologic seems a bit out of left field versus these items, but this could be an "in addition"
- Laura Mandel has some TEI texts of Poetess Archive
- Set up a philologic server, we crunch the data and return the results as a Bamboo Drupal model
- This would be something that would demonstrate something with an 'outside'
- this can be tabled for later
- we need to demonstrate what Mellon is paying for
- TSR demo
- Quinn: should be quick to create a screen cast version once data model approved
- Good idea to tip our demonstration to what we are heading towards in phase 2 cf curatorial
- We need to evaluate how easily a tool is "plugable" into workspace or in a java applet - seems difficult per Loretta's experience - Rapture toolkit is complex; but there are other options for HUBzero - so "web-like things' would work with Joomla component better. Loretta will give Bruce examples.
- The end user experience would be a determining factor for what to include
- Bridget: Alpheous tools should be easily hooked up to Workspace in the time frame
Big picture we agreed to create an "end to end" demonstrator using HUBzero, the CI HUB, and the BSP along with at least one scholarly tool and/or service that has a UI within the HUBzero interface.
We would like to show curatorial activities since this is our future strategic focus.
We also acknowledged that we will have a more powerful demonstration if we choose scholarly tool and/or service that works with at least one of the collections for which we have a CI adapter running; i.e. we have a dependency.
- Lorretta and Brigdet send along candidate tools for integrating into HUBzero and specific working examples so Bruce can study what it will take to put them; then a conversation after Bruce has done some research
- Make collections decisions
- We have CI Adapters that work on TCP and Perseus with the current book model
- CI Adapter for Hathi Trust is out of date vis a vis the current book model
- Make tools decision
- Make visualization tools decision - is there one we can implement within the HUBzero interface within our time frame?
- Make curation tools decision
- Travis: could propose gaps highlighted, as per his demonstrator, he could maybe do the work himself later in October; Mike Tang also could help
- This would not necessarily be a part of the "end to end" demonstrator
Once we have the evaluative materials and Bruce has had a chance to review, we will have a follow up meeting to finalize our demonstrator recommendation.