ELMO needs to consider probable future developments for more advanced features, possibly including:
- the work being done by InLOC with learning outcomes and competences
- other work planned around the e-portfolio space such as MELOI
- likely advances in European policy e.g. for Europass instruments
- to support more advanced services that we anticipate, e.g. in the area of employment services.
We want to enable people to describe or specify the learning outcomes of courses at any granularity, not restricted to university courses.
Levels in particular are useful to represent. The requirement is for machine comparison of different levels of the same competence. E.g. when an employer searches for someone with competence at a particular level, the search should find people with competence at higher levels.
There was some discussion on the LRMI list about levels, where it seemed to be accepted that two pieces of information to specify a level was plausible.
One issue is that some elements in the DS profile appear to be misplaced.
- professionalStatus would seem to fit better under qualification, rather than directly under the specification.
This is the single most important aspect of forward compatibility, bearing in mind particularly the InLOC project. The most obvious approaches to representing learning outcome and competence information in ELMO are
- using one identifier for competence and level
- using two identifiers, one for competence and another for level
then one could in theory have
- more than two identifiers
The examples below use the element name <competence> but any will do.
<competence uri="...">human readable descriptive material ...</competence>
This approach looks very simple.
However, in order to do comparisons of levels, this would rely on a series of processing steps:
- there must be a standard structure at the end of the URI
- that structure must be fetched
- the information in that structure must be processed to find the level and relationships between levels
<competence uri="..." level="...">human readable descriptive ...</competence>
This will facilitate several operations with no further fetches:
- compare levels of competence of different courses or learners
- enable search for whether any level of competence has been defined or not
There are two major variants of this approach.
First, for generic competence definitions with well-structured numeric levels
- the competence URI leads to a structure that clarifies how levels are used, the range of levels, etc.
- the level identifier can simply be a number.
Second, for generic competence definitions that do not themselves specify a particular level scheme
- the competence URI leads simply to a competence definition that may not have any relevant structure (e.g. the URL of a pdf file on the web)
- the level identifier is a URI pointing to a level within a generic level structure.
Examples of this kind of level structure include the EQF, with 8 levels in each of three areas (knowledge, skill, competence). However the EQF does not actually have any URIs defined for the levels.
There are then two more questions to follow up.
- Are the level URIs themselves directly comparable, enabling direct comparison?
- Given the level URI, how does one discover what are
- the parent generic definition
- the level number?
<competence uri="..." level="..." somethingelse="...">human readable ...</competence>
One could add, e.g. a human-readable label for the competence, for the level, a URI for an eventual level scheme, etc. Where to stop is a good question. If simplest is best, maybe the trade-offs do not favour more structured information in the spec?