Annotations, comments, field changes, forums, wiki pages: after discussion, they all seem similar to us.
They are text that may have references to semantic objects: zero or many, top level or a specific field.
(Dominic; Vlado made it more direct narrative and cleaned up a bit)
When a user wishes to annotate it could be to:
- Annotate a specific data field or sub-object
Next to each field there will be a button to make a particular comment about the field, e.g. "this may need further examination", or "The reason that this material was used is because....." (a type of comment)
- Suggest another version of the data
Sample use case:
- user disagrees with the the entry for the material and wants to suggest that the material is something different
- user clicks a button next to the field for which the new suggestion is aimed.
- system understands the datatype and any associated thesaurus.
- user selects new material from thesaurus and provides comment (why: reason or justification)
- system records the new version of the data field, together with change data (who, when, why)
- Annotate an object record generally
The annotation is a comment about the object in general.
Should this go against the identifying URI of the object? Vlado: yep, with a button next to the object's top level
- Annotate several objects (i.e. implicitly, their relation)
- Select several objects from data basket, then click a button.
- Start annotation on one object, then include a reference (link) to another object in the comment. For example "This is similar to the situation with object xxxxx...."
Wireframe: currently early draft
The suggestion is that the annotation pane of an object has an embedded discussion forum and this is the method that is used in all annotation tools. A user can create general posts, or more structured annotations. The discussion can be viewed as a chronological discussion or have an annotation view with appropriate filtering to home in on field, link or general annotations. This would fit into the concept that a dicussion can incorporate both general text and more formal annotation and that these formal annotation create links that others can follow.
(Note from meeting with the archaeologist Neal Spencer)
- Although the base data is shared between all project members, it may have sub-groups working towards different research outputs. In this case the annotations and discussions that happen in a sub-group may not be made available to other sub-groups or indeed the overarching project. This means that sub-groups need to specify whether a comment is visible to the whole project, or restricted to a certain sub-group.
- Vlado: which means a lot of complications. Eg:
- I doubt they'd be happy to select individual people for each comment, so they'd need to administer these sub-groups
- How does access propagate downwards in a comment tree (discussion thread)? Does it decrease monotonically, or can it be increased?
At discussions during w/c 26th September the following discussion between Maria Todorova, Vladimir Alexiev and Dominic Oldman.
The discussion forum allows researchers to make comments and discuss issues with other project team colleagues. The researchers can make use of researchspace tools to enhance their discussion post. The specification talks about being items backwards and forwards from research tools through the databasket. This means that a discussion could contain both general narrative and more specific information and data (with appropriate links).
Therefore it is possible that the correct app for annotation could also be the same discussion forum, operating at research tool level with the ability to link with project level discussions. If the discussion forum tool were enhanced in this way then it would provide;
1. A consistent tool across the environment which can be embbeded in a research tool.
2. The ability to swtich between annotation and genenral discussion views.
Problem - Research Tools should be independant apps able to operate outside the DMS. Would this create a dependency for annotation tools?
Participants: Vladimir Alexiev, Mariana Damova, Dimitar Manov, Yana Parvanova, Maria Todorova
Discussion was held based on the real data from Rembrandt DB.
- The group came into agreement that should be supported single page representation of all information for a data object. Thus will be able to annotate intermediate nodes also. Annotation will be done to each node/leaf with URI.
- At the left side of screen will be kept links to main data object nodes (like Basic Painting Info, Frame, Exhibitions, Auctions, Collections, etc.) so that the user will not have to scroll the whole page but be able to collapse and expand the information he is interested in.
- Annotation on empty node will not be allowed
- For each field (node/leaf) on the left side will be shown information for discussions, versions, linked records, etc if any.
- To be supported Delete of existing value - different scenarios for proposing deletion of value in original version of data record and in deleting value created by the user.
- Representation of empty nodes - A user during the research should be able to propose new values for a field. This means that on the screen should be represented empty nodes.