Example Creep

[KathiSchleidt]

I was reviewing the 1D_regular example, comparing the various versions, noticed subtle shifts that slowly obfuscate the (admittedly silly) meaning of the original example.

• 40_1D_regular.xml: this seems to be the original source. Here the RangeType has swe:field name="adventskalendertuerchen" (Advent Calendar Doors), the range of the example coverage provides the series of numbers to be displayed on the doors (silly, but makes some sort of sense), while the range is comprised of the 31 days of December 2015
• 40_1D_regular.json: Similar, but unequal. The UoM is still "d" to represent the day number for the door, but the field name has gotten lost.
• 1D_regular.json: a new and de-proved version of the online JSON example, Domain and Range are still the same as in the original example, but now we have "definition": "http://opengis.net/def/property/OGC/0/Radiance" and "uom": { "code": "W.m-2.sr-1.nm-1" }, and the example loses all meaning (providing the values 1-31 as the radiance for the 31 days of December 2015, ah...?)

Would it maybe make sense to start at the beginning, first define a set of examples to be provided as a textual description paired with the content bits in a spreadsheet, then provide these examples in various encodings? Such an approach would surely help both us in creating consistent examples, as well as the poor reader trying to interpret them!

[jerstlouis]

@KathiSchleidt 1D_regular.json fixed the DataRecord which was setting a data type for what should be the semantic definition. Not having a clear registry of valid definitions and completely clueless about the fact that this example was about advent calendars, I put Radiance there as in much of the other examples.

Would it maybe make sense to start at the beginning

it probably would, but would also probably require a lot of efforts. In the meantime, I suggest we try to improve the existing examples, making sure they comply with the schemas and semantic, and if they could also be meaningful that would be a great plus. Thank you for reviewing this!

Also it was in this spirit that I suggested we do a high-level visual overview of the different classes and sub-types/variations of coverages in CIS (and their different encodings), to see what is covered.

[pebau]

precious work! And even more precious it would be to have sorted examples - a good start, Jerome, with the discussion recently and with your analysis, Kathi. IMHO the biggest single step we can do to make coverages better understandable - can we go for it?