First Musings on Usage

This is about usage of the working OMRAS2 system that is our goal, rather than about the individual components of the system.

Users

In principle, OMRAS2 aims to provide helpful facilities for every kind of user.

In practice a community of expert users will be involved in the development and testing of OMRAS2. That is, experts in their particular domain of music (or use of music), rather than experts in music information retrieval per se.

The word 'Users' below is intended in this sense of expert users who will be involved in OMRAS2 development and testing.

Types of user

• 'Professional' (UsersProfessional^?)

• Other (UsersOther^?)

Professional users include:

• Musicologists (UsersMusicologist^?)

• Music Librarians (UsersMusicLibrarian^?)

• Record Industry people (UsersRecordIndustry^?)

• Others

Potentially also:

• Video professionals (UsersVideoProfessional^?)

• Composers (UsersComposer^?)

• Others

(Please add to or refine this list!)

Tasks

The number of things that OMRAS2 could be used for is potentially vast. Our design strategy will be to concentrate on certain selected task-scenarios for which we have some evidence that there will be a real demand.

We need to define, at first informally, but later with detailed and rigorous specification, a number of tasks which we can hopefully address during the project. For each task, it would be very convenient to have a good number of (human-expert-generated) examples of successful results which can act (probably with varying degrees of robustness) as 'ground-truth' for testing purposes.

Types of Task

• Search (TasksSearch)

• Browse (TasksBrowse^?)

• Filter (TasksFilter^?)

• Analyse (TasksAnalysis)

• Seek information (e.g. Count) (TasksInformationSeeking^?)

• Others

User Interfaces

• Queries (UserInterfaceQuery^?)

• Results (UserInterfaceResult^?)

Evaluation of Tasks

• 'Ground-truth' (aka 'Relevance judgements') (UseCaseEvaluationGroundTruth^?)

• Precision/Recall curves, etc. (UseCaseEvaluationPrecisionRecall^?)

• System response (UseCaseEvaluationSystemResponse^?)

• Evaluating subjective aspects (UseCaseEvaluationSubjectiveAspects^?)

• • User interface (UseCaseEvaluationUserInterface^?)

• • Presentation of results (UseCaseEvaluationResultPresentation^?)

Examples

• LeitMotif

-- TimCrawford - 19 Jan 2007