- From: Jacky Alcine <yo@jacky.wtf>
- Date: Wed, 15 Mar 2023 12:18:28 -0400
- To: public-swicg@w3.org
- Message-ID: <cb4148c6a2795427e0dcdc95918e097965b10a34.camel@jacky.wtf>
On Mon, 2023-03-06 at 09:49 -0800, Johannes Ernst wrote: > I’m not understanding what you mean by “meta level systems”. Do you > mean software that can process any extension because it essentially > extends its type system when it encounters new stuff? (e.g. by > downloading new @context resources etc) > > If so, then sure, but it only helps on the lower levels of the > application stack, like with networking and storage. It does not help > you towards creating a good, (consumer-grade, not geek-grade) UI, > which almost certainly requires “hand coded” elements, otherwise all > you get is the equivalent of property sheets. How should it know that > extension should offer to take a photo, while the other should offer > to insert GPS data — to pick some random examples? Yeah, this actually doesn't sound too unreasonable for a client. In a way, this is done by checking the type hinted at by objects. In the case of needing to add specific semantic information, this could be hinted at as well either by the @context's definition or some note that can give some ideas of mappings to use (like in the case of entering GPS data, that'd be a check-in - in Microformats2 land and its equivalent for ActivityStreams). A consuming and creation UI for each type doesn't necessarily seem unreasonable, with proper hints.
Received on Wednesday, 15 March 2023 16:18:44 UTC