Jump to: navigation, search

jens 08:45, 30 August 2010 (UTC) One given task was to evaluate the necessity of having different levels of confidence of nautical information. This site is made to discuss this issue and to share different views on that. The discussion should be focused on the problem not on different portrayal options. The results will be presented for consideration to the DQWG meeting in Rostock Oct 2010. So we don't have much time.

The outcome of various BSH internal discussions is that we have developed two approaches:

  1. Source related
  2. Time related

1. If we consider different sources for Nautical Information and try to generalize as much as possible we come up with following different levels of confidence. The provided items base on German source confidence. Other HOs can add various others to complete the list.

  1. authoritative (Legal bodies)
  2. official (international organizations (e.g. ITU), other federal organizations)
  3. not verified or not verifiable (e.g. reported, websites)

2. Discussing the same problem and taking time dependency into account. We might evaluate several options all with several pros and cons. One can follow the life cycle of the relevant publication. That can be very different from one HO to another and nobody is checking all information in a pub every time when a new edition is scheduled. Only changes are recorded. The other can be a time line indicating when the source was recorded or revised. The latter makes sense if we refer to very old sources in particular and it separates the information from the publication. That is our preference. The SORDAT approach can be adapted. Using the chart-chart datum is source datum, using the HPD source datum is source datum. For an initial upload SORDAT of printed information and database can by similar. That will change afterwards.

  1. 0-5 years old
  2. 5-10 years old
  3. older than 10 years.

It is unlikely that it will be necessary to track information older than 10 years more detailed.

DavidAcland 15:30, 3 September 2010 (UTC)

I agree that quality of information can be generalised to the source. I would not argue with the first two classes mentioned above. The third clearly exists. I think there are others which are probably worth considering.

I would take the word of a Master Mariner, who bothered to make a report about something, in which his or her vessel was not involved. If it had been, like a grounding or an accident, there could be reasons to treat the information with some care. Similarly, Ports can generally be relied upon and probably deserve a status a bit higher than "not verified", even though there may not be any other supporting evidence beyond the report from them that a lock is out of operation or a berth or basin is no longer in use.

As discussed at SNPWGs, time or date is vexed. A single old report of a quiet sheltered anchor berth, might be invaluable to Masters. I accept that date is interesting but I would not make the decision in the HO to remove the information just because it was old. Conversely recent information does not necessarily make it right.

I think we want to keep clear of the "wisdom of crowds". I would not support an approach commonly seen in websites where users are invited to rate the information provided. "Was this information useful? Answer 1-5".

jens 18:43, 3 September 2010 (UTC) It was clear to me and expected that we have to discuss that.
I agree with not introducing a rate of information provided.
Can we confirm that we need a statement of the time line? That would be one indication. We can extend the discussion about the quality for a while.

It is worth discussing the items 1.1 and 1.2 in combination with time line. Assuming information provided by an official authority (state or county) was recorded 10 years ago and nobody has ever checked it again. How reliable is that? Similar to your update status discussion I remember some nice discussions about planned harbour extensions at the Hebrides recorded 10 years ago as "planned" and not been updated. It is only a fact not a discussion who is responsible for the update. So pls be not disappointed.

Actually it is planned to visit SMM and several shipping companies and ECDIS manufactures at Hamburg next week. I hope to come back wiser.

following is off topic, only to record my mind:
"However, one solution can be to state "SNPWG does not need any confidence status of their information". But is that what is really requested?"
keep in mind that CATZOC is also dealing with both vertical and horizontal accuracy; parallel to time and liability.

DavidAcland 11:34, 14 September 2010 (UTC)

HSSC2-03C paper [1] has some interesting factors which we might like to consider. An extract, which may help our thinking, follows:

"The DSCC (Data Supply Chain Certication) Standard proposes to use these seven (7) characteristics to express integrity and quality of data: accuracy, resolution, assurance level, traceability, timeliness, completeness and format. These are the same characteristics used in RTCA /DO-200A Standards for Processing Aeronautical Data, from which DSCC standard has been based.

 Accuracy – The degree of conformance between the estimated or measured value and its true value.

 Resolution – The smallest difference between two adjacent values that can be represented in a data storage, display, or transfer system.

 Assurance Level – Quantifiable value that communicates clearly what level of trust a user can place on the assessed data.

 Traceability – The degree to which a system or a data product can provide a record of the changes made to that product and thereby enable an audit trail to be followed from the end-user to the data originator.

 Timeliness – The degree of confidence that the data is applicable to the period of its intended use.

 Completeness – The degree of confidence that all of the data, needed to support the intended use, has been provided.

 Format – The process of translating, arranging, packaging, and compressing a selected set of data for distribution to a specific target system. A result of this process is a data structure that fulfills the characteristics of data quality."

I think what we have discussed above is the "Assurance Level" characteristic.

jens 07:02, 26 October 2010 (UTC) ok, I see the point. If I interpret the HSSC2-03C paper correctly and their transition between current hydro and aviation standards I assume we have our problem solver without the need to introduce new ideas.
SORDAT and SORIND can be used to describe from when and from whom the information came from.
M_COVR can be used to describe if NPUB information exists for the particular area.
ja, I think using existing features and attributes is a good idea.

jens 07:39, 27 October 2010 (UTC) further attribute values which can be used are QUAPOS (reported, not confirmed) and (unreliable); further values have to be considered depends on source