RE: ISSUE-184: Is an dqv:ServiceLevelAgreement a kind of certificate, or a standard?

Nandana,

 

Yes this is sometimes a problem with URIs that are “human-readable” – the human reader my infer the meaning from the URI without making the effort to actually read the full definition. I once got the question why Dublin Core has two properties dct:format and dct:hasFormat; the person asking did not understand the difference because he/she did not read either definition.

 

Imagine that the URI was foo:XYZ instead of dct:Standard; on the basis of the definition, there would be no discussion whether it could be the superclass of dqv:SLA. (Although I haven’t yet seen the formal definition of dqv:SLA, and I am just inferring the meaning from the URI ;-)

 

Makx.

 

 

 

From: Nandana Mihindukulasooriya [mailto:nmihindu@fi.upm.es] 
Sent: 24 September 2015 11:22
To: Makx Dekkers <mail@makxdekkers.com>
Cc: Data on the Web Best Practices Working Group <public-dwbp-wg@w3.org>
Subject: Re: ISSUE-184: Is an dqv:ServiceLevelAgreement a kind of certificate, or a standard?

 

Hi Makx,

 

Yes, I agree as it is defined in DCTerms it can be a superclass of agreement as it is defined much broader than the "standard" that is commonly understood as a result of work in “standardization bodies”. So probably with respect to the semantics, though it might lead to some confusion for human-readers. Having said that, I am *not* strongly opposing the use of it and ok with what the WG decides. 

 

Best Regards,

Nandana

 

On Thu, Sep 24, 2015 at 11:11 AM, Makx Dekkers <mail@makxdekkers.com <mailto:mail@makxdekkers.com> > wrote:

Nandana, all,

 

I am not quite sure why a dqv:ServiceLevelAgreement cannot be a subproperty of dct:Standard. 

 

The definition of dct:Standard is “A basis for comparison; a reference point against which other things can be evaluated”. The criticism of this definition could be that it does not say what it *is* but only how you can *use* it.

 

However, I think that this definition is sufficiently general to include the idea of an SLA. An SLA will provide information, like KPIs, against which a customer can evaluate a service. The definition of dct:Standard can also include “certificate” and “contract” if you want. 

 

An instance of the class dct:Standard does not have to be “standard” in the sense of the result of work in “standardization bodies”.

 

Makx.

 

From: Nandana Mihindukulasooriya [mailto:nmihindu@fi.upm.es <mailto:nmihindu@fi.upm.es> ] 
Sent: 24 September 2015 10:36
To: Data on the Web Best Practices Working Group <public-dwbp-wg@w3.org <mailto:public-dwbp-wg@w3.org> >
Subject: ISSUE-184: Is an dqv:ServiceLevelAgreement a kind of certificate, or a standard?

 

Hi all,

 

I read what has been discussed re: ISSUE-184 and I agree with general opinion that dqv:ServiceLevelAgreement is neither a certificate, nor a standard. 

 

Based on the definitions of  ITIL v3 [1] and ISO 20000 [2], I understand a ServiceLevelAgreement [3] as an agreement between two parties such as the service provider and the consumer. It can be part of the contract/charter and describes (or promise) the relationship or the level of service expected by the provider. 

 

With respect to finding a class to base dqv:ServiceLevelAgreement, I agree with Antoine that ODRL Policy [4] could be an option. We could also look at its more concrete two sub-classes Offer [5] and Agreement [6]. But I have the impression ODRL is bit more focused on rights that the consumer get rather than the level of service the consumer expects from the provider. I will talk to Víctor Rodríguez Doncel from my group (who is also an author of the ODRL ontology) to see if they can provider an example of how to model a service level agreement in ODRL. https://schema.org/Offer is a bit more less specialized as Antonie says yet more specific than a general document, so it is a good option too.

 

Other ontologies that model similar concepts:

http://lov.okfn.org/dataset/lov/terms?q=Agreement

http://lov.okfn.org/dataset/lov/terms?q=Offer

 

ITSMO IT Service Management Ontology models the concrete concepts SLA and OLA. 

http://ontology.it/itsmo/v1/itsmo.html#term_Agreement 

http://ontology.it/itsmo/v1/itsmo.html#term_SLA

http://ontology.it/itsmo/v1/itsmo.html#term_OLA

 

Beyond the base class, dqv:ServiceLevelAgreement could be related to quality metrics too because the service level agreements might include KPIs as such as quality metrics, for example, that the provider promises >99% availability, 100% completeness with respect to a agreed criteria, etc. It could also have some relation with dqv:QualityCertificate because it might be able to prove the fulfillment of the SLA with a certificate about the levels of metrics included in the SLA. However, I am not sure whether those relations are generic enough to be reflected in the model. 

 

So, regarding the proposals

* I think SLA is not a standard and the current subclass relation "dqv:ServiceLevelAgreement is rdfs:subClassOf dcterms:Standard" can be removed.

* We can send an email to the authors of the ODRL Ontology to check if they think ODRL fits for describing service level agreements.

* or use schema:Offer 

 

Finally, whichever we choose I think we can provide one good example in the document to illustrate their use so that readers have guidance on how to use that class and also we can further evaluate the goodness-of-fit.

 

Best Regards,

Nandana

 

[1] http://www.best-management-practice.com/gempdf/itil_glossary_v3_1_24.pdf

[2] http://www.praxiom.com/iso-20000-definitions.htm#3.29_Service_level_agreement_(SLA)

[3] http://www.city.ac.uk/__data/assets/pdf_file/0007/133936/Service-Level-Agreements.pdf

[4] http://www.w3.org/ns/odrl/2/ODRL21#term-Policy 

[5] http://www.w3.org/ns/odrl/2/ODRL21#sec-example-2

[6] http://www.w3.org/ns/odrl/2/ODRL21#sec-example-3