1. EDIT Certification Goals¶
Note: The EDIT certification goals were discussed by the ISTC in 2006 (see 1st ISTC Report: http://cybertaxonomy.eu/blog/filesedit_wp5/ISTC_Report.pdf)_
What are the goals of certification? We need to debate this, because they will determine what certification becomes, and certification could be one of the major and most influential outcomes of EDIT WP5.
Malte Ebach cited five goals in his report of the Berlin ISTC meeting of November 2006, as follows.
1. Assess tools for cybertaxonomy based on a list of criteria
Agreed â€“ without this there is no certification. But are we assessing the tools in the sense of judging or rating them, or are we just recording factual metadata? I prefer to start gently, recording some factual, objective metadata and leaving the users to draw their own conclusions. Once the cyberplatform is designed, up and running and it is clearer what users needs are, we can become more prescriptive.
2. Promote the use of standardised cybertaxonomic work practises worldwide
Hmm, well this may be a little incendiary to any listening taxonomists! But how would certification of tools encourage a standard practise of cybertaxonomy? We already know that many tools may be certified in a given category. The only way I can make a connection is if the cyberplatform recommends â€˜standardâ€™ workflows, or suites of tools. Certification of individual tools will not in itself achieve this.
3. Ensure that all cybertaxonomic data are up-to-date and backed-up
I donâ€™t see how having a list of certified tools can do this! Although a tool can help this, by providing easy import and export facilities, and storing data in accessible ways, keeping data up-to-date and secure is more a matter of human effort and following good practice.
4. Identify the necessary criteria for certification
Indeed we must do this, but it is an activity and not one of the goals of certification itself.
5. Ensure the availability of high quality cybertaxonomic data to all cybertaxonomists
This is two points â€“ quality and availability. Maybe they should be separated.
Certification may help address these issues if we, for example, mandate standards compliance. However, it cannot ensure high quality or data availability as once again other factors dominate. For example, data quality will always be to some degree subjective and dependent on human effort, and data availability depends on IP issues and where the data is eventually stored (not necessarily in a certified tool, unless EDIT intends to certify resources like herbarium catalogues and/or web portals and resources like GBIF.
How about some other goals? For example
a) to provide a list of tools to potential users
b) to capture metadata about the tools which will enable potential users to make informed choices about them
c) to identify areas where new tools are required (e.g. if none are certified)
d) to encourage interoperability of tools
and finally, increasingly contentious perhaps,
e) to encourage tools to interoperate with(in) the cyberplatform â€“ depends really on what the cyberplatform turns out to be
f) to improve the sustainability and suitability of tools for cybertaxonomy
g) to encourage the development of standardised workflows using certified tools
Mark Jackson RBGK
23 November 2006
2. EDIT Certification Process¶
The preliminary document may address the following points (it's a summary of the points we have debated):
Does certification mainly focuses on workflows (integration of existing tools into a common platform) and/or on the data themselve (standardization of the output data provided by the softwares, or assessment of their compliance with existing data models and data standards)?
On which criteria will the certification be based? This question is directly linked with the depth of the certification process we want to set up.
It could be based on the description of the metadata of a tool (such as the type of data standards it uses, its commercial status, etc...) and/or on a serie of tests assessing the quality of data which can provided by this tool. In the second situation, is it possible to implement these tests on a central website such as the bdtracker, or do this validation tests need to be done on a terminal machine having the tool installed ?
2.4 Decision process
How do we control the decision which are taken during the certification? Maybe we should separate in the description of the process and in the bdtracker the publication of the results of a certification from the discussion which are held by the board of reviewers.
2.5 Ratification process
How do we ratify the certification? There are several possibilities: the BDtracker could be the reference "document" for this certification or but we maybe need to ratify it on other documents (maybe printed documents) to ensure that the result of the certification remain independent from platform of diffusion.
_Franck Theeten RMCA
14 August 2007_
a) Existing support, institutional backing, expected "survival duration"
b) CDM compatability. level1 import/export, level2 cdm sync server.
Markus DÃ¶ring 17.08.2007