Normative requirements regarding the use of the temporal properties (together)

[heidivanparys]

The specification does not contain requirements on how to use the temporal properties (creationDate, terminationDate, validFrom, validTo) together, nor does it contain normative references to other specifications that specify how to use those temporal properties.

How will it be ensured that software systems interpret these temporal properties in exactly the same way, in order to achieve interoperability?

As an absolute minimum, the concepts "valid time" and "transaction time" should be added to the specification, both from [3].

I would also like to see more references to relevant material in the specification. [2] is a must read for anyone who is new to the topic. [3] is good for definitions on the different concepts and a good reference work. [4] and [5] are older and were input to [3].

When it comes to how to use these concepts in relational databases, there is ISO 9075, and related to that e.g. [6] and [7].

Regarding actual requirements: one way could be to add requirements in the specification that make it clear how to use these properties, grouped in one requirements class, that targets software dealing with city models. Another (easier?) way could be to add a reference to an appropriate normative document.

An example of what is missing, is a description of whether or not terminationDate and validTo are inclusive in the interval.

According to [2]:

The preferred representation of a period is a closed-open pair of datetimes.

According to [1]:

The closed/open notation, rather than the open/closed, has become a de facto standard. […] The closed/open convention is also the de jure standard in both ISO 9075:2011 and TSQL2.

In addition, e.g. a requirement regarding non-overlapping valid time intervals should probably be present too.

[1] JOHNSTON, Tom. Bitemporal Data. Elsevier, 2014. ISBN 978-0-12-408067-6. Available from: 10.1016/C2012-0-06609-4
[2] SNODGRASS, Richard T. Developing time-oriented database applications in SQL [online]. San Francisco, California : Morgan
[3] LIU, Ling and ÖZSU, M. Tamer, eds. Encyclopedia of Database Systems [online]. New York, NY : Springer New York, 2018. ISBN 978-1-4614-8266-6. Available from: 10.1007/978-1-4614-8265-9
[4] JENSEN, Christian S. and SNODGRASS, Richard T., eds. TR-90, Temporal Database Entries for the Springer Encyclopedia of Database Systems [online]. Technical Report. TimeCenter, 22 May 2008. Available from: http://timecenter.cs.aau.dk/TimeCenterPublications/TR-90.pdf
[5] JENSEN, Christian S. and DYRESON, Curtis E. The Consensus Glossary of Temporal Database Concepts—February 1998 Version. In: Temporal Databases: Research and Practice [online]. Springer Berlin Heidelberg, 1998. p. 367–405. Lecture Notes in Computer Science. Available from: 10.1007/BFb0053710
[6] ISO/IEC JTC 1/SC 32. ISO/IEC TR 19075-2:2015, Information technology — Database languages — SQL Technical Reports — Part 2: SQL Support for Time-Related Information [online]. Technical Report. International Organization for Standardization, 1 July 2015. Available from: https://standards.iso.org/ittf/PubliclyAvailableStandards/c060394_ISO_IEC_TR_19075-2_2015.zip
[7] PETKOVIC, Dušan. Temporal Data in Relational Database Systems: A Comparison. In: ROCHA, Álvaro, CORREIA, Ana Maria, ADELI, Hojjat, REIS, Luis Paulo and MENDONÇA TEIXEIRA, Marcelo, eds., New Advances in Information Systems and Technologies [online]. Cham : Springer International Publishing, 2016. p. 13–23. ISBN 978-3-319-31231-6. Available from: 10.1007/978-3-319-31232-3_2

[cmheazel]

The standardization target for this Standard is Implementation Specifications, not software. Keep this in mind when evaluating this issue.

[TatjanaKutzner]

Thomas and I discussed this issue, we will to add more information on temporal properties including requirements.

[3DXScape]

Explain that duration is a semi-open interval.
Also plan to explain meaning and use well in Users Guide. Also cross-reference to other standards, such as INSPIRE.

[3DXScape]

Thomas will update in next week - by 10th September.

[TatjanaKutzner]

Has not been updated yet, but will be done by Thomas. This will include in particular an update to the description in 'Chapter 7 Overview of CityGML' as well providing detailed information on temporal properties, identifiers and valid time intervals in the Users Guide.

[cmheazel]

@TatjanaKutzner this should not add any new normative content,. But it would be best not to release a version to the SWG until after Thomas has made his contributions. You never know.
Thanks

[TatjanaKutzner]

@cmheazel This will not add any new normative content, it will rather describe things more clearly. Thus, a release to the SWG is not a problem. Also I don't know when Thomas will find time to update the text.

[cmheazel]

@TatjanaKutzner Thank you, then I think we have our review draft.