use of http://anzsoil.org/

[meganrwong]

raised in #5 "The http://registry.it.csiro.au/def/soil/ URI is the locator, not the identifier. The identifier is the http://anzsoil.org/... URI."

We've not had the discussion re when it is appropriate that http://anzsoil.org/ is used for a vocabulary that is not an 'Australian standard'. Who decide this?

[dr-shorthair]

Isn't that why you have 'edu' in the path?

[dr-shorthair]

But whatever, http://registry.it.csiro.au/def/ URIs are not identifiers. This merely reflects where they are currently located.

[meganrwong]

You are right Simon. That is why we have 'edu' . We should use anzsoil.org . With the governance in the path (here fed-uni) this will (should) encourage others to similarly create, manage their own organizational vocabularies as required, and help build community/encourage adoption of controlled-vocabularies. I will change identifiers throughout to http://anzsoil.org/ Any comments/thoughts here @ljgregory ? attn @bsimons14

[dr-shorthair]

Cue rant from @bsimons14 about how it is confusing for http: URIs to serve as both identifiers and locators.

Cue W3C rant from Henry Thompson about why there is no need for any novel identifier scheme since we have http.

Cue the whole Pidapalooza thing.

@philarcher may also have some updated reflections, formerly a W3C die-hard, now in bed with GS1

[bsimons14]

@dr-shorthair I’m now a broken man, Simon (do DOIs stand for “death of interoperability “?)

[philarcher]

Tempting though it is, I'll refrain from rehashing tired old arguments (maybe I'm getting old). I'll simply point to the work I lead at GS1 in this area. We have a massively deployed 50 year old ID system for products, healthcare items, shipments, locations and more. Most people know them as 'the barcode' but of course it's the numbers in those barcodes that really matter.

Many different actors have reason to provide info about those items (manufacturers, logistics companies, retailers, special interest groups, regulators). GS1 Digital Link provides a means to include an ID, or set of IDs, in a URL in such a way that those IDs can be extracted without an online lookup. But, if you do follow the URL, you'll get some info from the manufacturer. If you query the URL for a particular kind of info (give me allergy advice) you can get that (if it's available), or you can ask it "tell me everything" and get a set of links back. Because the URL is structured, you can swap the domain name for another one and ask the same question. You might be pointed to different related sources of info but that's OK. You looked the item up in one reference book rather than another. See Developers' intro documentation for more.

So the IDs identify the thing. The resolver is where you can look up info about it. I'm dancing on the head of a pin here but the URL is not a locator for the thing but a look up service for info about it. (This is also how we just about avoid the Http-range 14 issue too).

And this is relevant to science how?

DOIs have a metadata structure so you can provide extra metadata pointing to, for example, research datasets, software used etc. GS1 Digital Link isn't providing something new therefore. However, it might offer an easier interface for it. The principles are documented separately and include an invitation to collaborate.

Time for an example. Imagine "product description page" is the paper with the DOI, "recipe ideas" is the results processing software used etc.

This link will lead you to the product description page https://id.gs1.org/01/9506000134352

But maybe you're looking for a recipe that uses that product https://id.gs1.org/01/9506000134352?linkType=gs1:recipeInfo

Perhaps you're interested in the manufacturer's policies on sustainability: https://id.gs1.org/01/9506000134352?linkType=gs1:productSustainabilityInfo

Or where to buy the product https://id.gs1.org/01/9506000134352?linkType=gs1:hasRetailers

By now you just want to know "what can you tell me about 9506000134352?" https://id.gs1.org/01/9506000134352?linkType=all

[dr-shorthair]

Thx @philarcher that is really helpful.

the URL is not a locator for the thing but a look up service for info about it

That matches how I have explained SISSvoc, where

https://vocabs.ardc.edu.au/repository/api/lda/csiro/international-chronostratigraphic-chart/geologic-time-scale-2020/resource?uri=http://resource.geosciml.org/classifier/ics/ischart/Jurassic

can be read as 'what https://vocabs.ardc.edu.au/repository/api/lda/csiro/international-chronostratigraphic-chart/geologic-time-scale-2020 knows about http://resource.geosciml.org/classifier/ics/ischart/Jurassic. This may be different to what https://vocabs.ardc.edu.au/repository/api/lda/csiro/international-chronostratigraphic-chart/2017 knew about the same thing.

Nevertheless I think we can still say that https://vocabs.ardc.edu.au/repository/api/lda/csiro/international-chronostratigraphic-chart/geologic-time-scale-2020/resource?uri=http://resource.geosciml.org/classifier/ics/ischart/Jurassic is the locator for a description of http://resource.geosciml.org/classifier/ics/ischart/Jurassic

[bsimons14]

The problem is when we have a real-world object that is not an on-line resource. It can never be delivered via a http-uri, only representations of it. I still maintain that using a http-uri as its identifier is confusing and inappropriate.