[08:31:07] <bawolff>	 I kind of feel like the biggest mistake in XML/RDF/SPARQL is the choice to have namespaces have a protocol instead of using protocol relative urls
[08:38:39] <Nemo_bis>	 bawolff: does the protocol even matter? (in normal XML)
[08:39:01] <bawolff>	 It doesn't, but it confuses people
[08:39:17] <bawolff>	 officially, namespace URIs do not have to be dereferencable
[08:39:33] <Nemo_bis>	 "confuses people" – definitely https://chaos.social/@steely_glint/109528649928353793
[08:40:51] <bawolff>	 I'm pretty sure that thread is a bit confused, because the url in an xmlns is not the location of the schema
[08:41:06] <Nemo_bis>	 yes https://mamot.fr/@nemobis/109530344018319061
[08:41:06] <bawolff>	 They're probably more thinking of DTD definitons where that was a thing
[08:41:15] <Nemo_bis>	 could be
[08:45:28] <bawolff>	 I don't think i have ever in my life actually seen an xsi:schemaLocation attribute, which would be the way its done for xml schema
[08:46:48] <bawolff>	 DTDs are weird though, because they sort of have to be fetched because they don't just have schema related stuff, but also entity reference definitions, which you need to parse the document
[08:49:02] <bawolff>	 reminds me of http://web.archive.org/web/20080206214640/http://blog.netscape.com/2007/01/16/to-dtd-or-not-to-dtd/ 
[09:03:12] <bawolff>	 anyways, i just think having things like //w3.org/1999/xhtml would be so much less evocative of something dereferencable, but still have all the benefits