- Current status of the working group note (use cases, glossary)
- Current status in the two task forces (Task model, Abstract UI)
- Next F2F meetings (June in Pisa, October/November at the Tech Plenary in Lyon?)
- Dave (W3C)
- Javier, Cristina (CTIC)
- Marc (DFKI)
- Jaroslav (Fraunhofer)
- Carmen, Davide (ISTI)
- Gaelle (Grenoble Institute of Technology, Grenoble)
- Joelle (University Joseph Fourier, Grenoble)
- Marc on behalf of Gerrit who couldn't make it today
Current status in the two task forces (Task model, Abstract UI)
No one present to represent UCL.
Carmen: regarding task model spec, we have made some modifications, in particular how to make the meta model accessible. We added a link to the XMI version of the meta model, which is a bit verbose. We need to decide if that is sufficient, or whether to use another approach. We also improved the legibility of the labels. We have included an example of CTT from Serenoa, along with a diagram and a matching text description as a nested list.
Gaelle: why did you use subclasses rather than actors?
Carmen: it gives us better control the categories of tasks.
Gaelle: what about refactoring?
Carmen: you can still do this, if you use an actor attribute you might lose some control.
Gaelle: what about an association between a new class and the task class, so the instance can be changed dynamically and more easily than using sub-classes.
Carmen: we can consider this
Gaelle: I want to enable every aspect of a task to be dynamically changed.
Jaroslav: The distinction between human and computer mediated tasks is substantial, task designed for human processing can't be equally replaced by an automatic one.
Gaelle: in a PC you can allow for flexible interleaving of tasks, but for a more constrained device, this is likely to be replaced by sequences of tasks.
Joelle: the task of saving a file from a word processor can be done by the user or automatically by the application. So the actor can change from a human to the computer. Using an attribute for the actor makes this easy to deal with.
Davide: saving a file is a system task, it is only the trigger that can be assigned, either to the user or to the system.
Gaelle: how would you model this in CTT?
Davide: with tasks for the triggers and a task for the file save operation.
Gaelle: what happens if you have multiple users?
Davide: the question is who performs what action. The action of writing the file is always performed by the system.
Gaelle: the key is who is the actor.
Carmen: this can be considered as different locations, one for the user and the other for the application.
Gaelle: how do you deal with multiple users? We agree with the value of sub-classes, but also want to capture the idea of actor including for one or more users.
Carmen: we have approached multiple users via a means for describing cooperative tasks.
Dave: what about using simple tables as per the discussion on the list as an easier to read alternative to XMI?
Carmen: yes, we plan to look at this as for us it would be a better option.
Davide: we could also introduce hyperlinks between concepts.
Marc: perhaps we ought to approach the discussion on actors/users by considering a set of use cases.
Carmen: we can do that, but in parallel with progressing the spec.
Gaelle: we would support looking at use cases for the task models.
Carmen: we have associative operators (in regards to Gaelle's comment in the spec on operators). If you have further comments please add them to the document.
Marc: shall we proceed to the AUI spec?
Marc asks Javier to explain his comments.
Dave asks if Javier means to reference concept definitions in the glossary on the Note?
Dave: for normative concepts we need normative definitions, and these will have to be in this spec or a normatively referenced spec.
Marc: we need some examples to drive the discussion on the detailed design.
Vivian is unable to reach a regular phone and sends an update via email:
Regarding the AUI document, we are currently working on the meta-model for the AUI, a consolidated version will be send soon. And regarding the requirements that Javier (CTIC) sent, we agree in integrating them in the document, and we invite all to collaborate with more requirements too.
Concerning the question raised by Gaelle last meeting, we believe that completeness is still valid in this context, and maybe one approach to complement this information could be classifying the selected requirements according to their relevance given an application scenario.
The meta-model and a more consolidated version of the AUI model document will be provided in 3 weeks, before the face-to-face meeting, in which we will have the opportunity to discuss them in more details.
June F2F meeting in Pisa
People would prefer to avoid travelling on Sunday.
Carmen: we had been planning for starting the MBUI meeting at 9am on Monday June 11, and ending middday on Tuesday June 12.
Joelle: what about the second half of the week, e.g. Thursday and Friday, as we have a meeting in Metz, France on Monday and Tuesday, and only one of us could make the MBUI meeting in that case.
Carmen: I will ask Fabio and report back next week.
Is this okay for everyone else? [Yes]
Jaroslav: we need to fix the dates and times in order to arrange our travel.
Carmen: we will let you know as soon as possible.
Meeting at the W3C Tech Plenary
Dave: I have made a provisional reservation for an MBUI WG F2F at the W3C Tech Plenary in Lyon, this November. I don't yet know the exact day we would meet just yet.
Current status of the working group note (use cases, glossary)
Jaroslav: I have been working on the glossary. The Cameleon glossary makes heavy use of acronyms, and think this could be a problem for wider understanding. I would prefer us to avoid acronyms where possible.
Dave: Yes, I prefer plain language as it is easier to understand.
Jaroslav: what about synonyms?
Dave: it would be helpful to list synonyms for each definition where applicable.
Jaroslav: I also plan to use cross references.
Dave: yes, that is in the spirit of the Web.
Gaelle provides some historical background to the origins of the Cameleon glossary in previous work on HCI. There have been many projects hence the synonyms as different projects used different names.
We need to clarify our preferred terminology for the W3C work.
Some discussion on the URIs to use for existing glossaries.
Carmen will try to find a more up to date version of the Cameleon glossary.