@scott it's really not

this is more about example cases like trwnh.com referring equally to "the web page / on host trwnh.com" or maybe just "the host trwnh.com" or maybe even "the primary subject of the web page / on host trwnh.com, which is the person a"

in RDF you would need different identifiers for each of these, to avoid overloading the term trwnh.com

the difference between Thing and Descriptor Document For The Thing is largely academic but it comes up a lot when making knowledge graphs

@scott so you might say something like

- <https://trwnh.com/#homepage> is a web page
- <https://trwnh.com/#host> is a web host
- <https://trwnh.com/#me> is a person
- <https://trwnh.com/#machine> is the vps on which the http server is running

the uris i pick can be anything but they must be unique

i might host these statements in the same document serialized to multiple formats

for more freedom i can instead use "slash uris" instead of "hash uris", so that statements can be fetched separately

@scott so for example

- <https://trwnh.com/rdf/things/me> is a person
- <https://trwnh.com/rdf/things/my-homepage> is a web page

i can now say something like

- <https://trwnh.com/rdf/things/me> has a homepage which is <https://trwnh.com/rdf/things/my-homepage>
- <https://trwnh.com/rdf/things/my-homepage> has a URL which is "https://trwnh.com/"
- <https://trwnh.com/rdf/things/my-homepage> was last updated on "2024-10-16T04:46:25Z" which is specifically a datetime

but when i GET / on Host: trwnh.com, what am i *really* asking for? me, or the homepage? you need a strategy

@scott one strategy is to say it's always 200 OK for the document for the homepage

another strategy is to check for Accept headers and if it's a linked data serialization, then return 303 See Other and list all possibilities (me, my-homepage)

there are other strategies and they mostly depend on how you want to configure your web server and where you want to store the backing files/documents