.

Tasks for the UT RKF Team

People Priority Project Status
Dan High Grep for element-sequence in the biology subdirectory. You'll see it used in Replication-Origin-Sequence and other files. It doesn't show up in SHAKEN. Should it? Is the problem the use of :seq ? Consider fixes to SHAKEN so it shows that information.
Dan Low The SME dictionary should map the action 'cleave' to our Divide component, but it doesn't. I suspect that Divide is not properly indexed into Wordnet. Do you know how to fix this, or should we try to reach Ken?
James High We need to hide our full component library, the TRUNK, from SME's. Make our "default tree" be the summer branch, and send email to the people in our group describing how to access the tree containing the TRUNK.
Art High As I recall, last week during training, we discussed some issues involving the representation of containers, container walls, contained spaces, and contained objects. Let's settle on a single representation, use it consistently, and document it well.
Art and Paul High Continue experimenting with Shaken to encode the information in Alberts 7.1. Generate questions that are likely to be asked during the summer trials. Use these scenarios and questions to test SHAKEN and KANAL.
UT High Decide whether to include in the KB for the summer trials the representation of properties and values. If `yes', then SHAKEN needs to handle 3-ary predicates. (A Cardinal Value is a numeric value and a unit of measurement. For example, (98.6 *Degrees-Fahrenheit).)
Art and Bruce Moderate Develop better ways of (cross-)indexing the pump priming information. Options include: scanning diagrams and labeling/hyperlinking them with terms in the KB; for each component, including with its description hyperlinks to all other terms that are slot-fillers of that component.
Bruce Moderate Finish writing (what we used to call) sme documentation for the actions in the component library. Also, revise the doc's listed in to-do:rkf folder.
Peter Moderate Eliminate Compound-Action (we now have Scenario, instead). Coordinate with SRI (Vinay) to convert your earlier work to use Scenario.
Bruce Moderate Write documentation (spec's) for some miscellaneous terms. Here's an initial list.
Art and Paul High Take a look at Free-Nucleotide in a SHAKEN CMAP. It illustrates a problem caused by the fact that has-structural-part is a spec-slot of has-part. We need a creative solution.
Art and Peter High Take a look at this problem, which we discovered last week in training: the fully expanded cmap for ADP fails to show the attachment between the Ribose and the Carbon-Atom.
Art and Paul High Art wrote to Paul: I see that when one expands DNA-Molecule on shaken, and then expands "Container" one gets "Genetic-Information-Storage & Be-Contained". The first is a spec of Be-Stored. Any ideas how to unify these two?
Dan Moderate Fix the hyperlinks to slot documentation in our Tree's display of the slot dictionary.
Dan Moderate Fix the problem caused by special characters in the display of 1-liners.
Art and Paul Moderate Paul wrote to Art:

I attempted to unify Ribonucleotide and Free-Nucleotide, which failed because these are sibling concepts. Thus, if I wish to specify a Free-Nucleotide (say in the initial attach step of RNA synthesis), I cannot specialize it to be Ribonucleotide or Deoxyribonucleotide. Any ideas on how to get around this?

James High Element-type causes way more trouble than it's worth. Because it's the only slot that takes a class as its filler, it doesn't work well in SHAKEN.

Wish List from the UT RKF Team

People Priority Project Status
Vinay and Pete High When SHAKEN gives a description, parts of the description are too brief. For example, it might report that some Tangible-Entity is the object of a Be-Attached-To, but not give any details of that state, such as the other object of the attachment! So, for some slot-fillers, SHAKEN should elaborate the description to an extra level of detail, without the SME asking for it. Here's an initial list of slot-fillers to elaborate.
Vinay High SME's need a save-as facility in SHAKEN. This would enable an SME to save their work, to keep multiple versions of their work, and to return to a saved version to continue working on it. We can train SME's to save their work frequently, as protection against crashes and KE decisions they come to regret. done-UT: test it
Pete and Vinay High SME's need a way to remove slots and fillers from a KB.
Pete and Vinay Medium SME's need a way to rename (i.e. assign their own labels to) slots and fillers from a KB. done-needs testing
Pete Medium? There's a general-purpose question template which might enable SME's to pose a variety of questions that they can't pose with the current templates. It's "report the relationship(s) between Thing1 and Thing2". Ideally, the SME could give restrictions or biases on which slots to search through to find the connection.