- From: Michael Douglass <mikeadouglass@gmail.com>
- Date: Sun, 26 Mar 2023 16:21:45 -0400
- To: Sergey Ponomarev <stokito@gmail.com>, Julian Reschke <julian.reschke@gmx.de>
- Cc: "w3c-dist-auth@w3.org" <w3c-dist-auth@w3.org>
- Message-ID: <69a8faa1-61a3-2f4f-2ce7-4a290dab3ebb@gmail.com>
This looks very much like https://datatracker.ietf.org/doc/draft-ietf-calext-subscription-upgrade/ This was proposed to allow advertisement of a more efficient form of subscription rather than repeated download of an ics file. I've no idea why it hadn't occurred to me that it would make sense on a web page as headers. On 3/26/23 13:13, Sergey Ponomarev wrote: > Thank you, Julian, > So there is no official protocol schema. But unofficially the dav:// > and davs:// become more supported e.g. in GNOME Files (Nautilus) and > soon in KDE Dolphin. > I also created a ticket for Nautilus to support the webdav:// schema > just because users may not remember a proper schema. But it may take > a lot of time for them to implement. > > Interestingly, just the http:// should also work for Gnome Files but > it looks broken. Windows mount works only by the http:// > It's a good question if the separate protocol schema is needed. Even > if a user clicks on the http:// link the browser itself may try to > detect the dav folder and render it as a folder. > Still having an explicit protocol would be better. So you add it > somewhere in spec or somehow confirm that you are fine with usage of > dav:// and davs:// as a schema? > > > > Speaking about some header added to all GET request to let know about > ability to watch the folder as dav. > We have few options: > 1. DAV header. It would be easier to implement because it's already > added for OPTIONS. It can be big for SVN or something like that but > this is a rare case and anyway HTTP2 HPACK should solve the problem. > 2. Link is less obvious and not described in the dav spec so it needs > for more work. > 3. Alt-Svc is also something that may be used. Not sure if this would > be a proper usage. > > The preferable view as webpage or as a webdav may be an > overcomplication. Maybe user's browser have a dav less useful that > automatically generated directory listing. Most time I expect oposite. > Here a browser can show a button on top of the page and propose to a > user to decide and remember it's choise. This would be a simplest > solution and if needed we may back to this on future. > > Read only view also doesn't have a good solution so let's keep it out > of the scope. Only ACL or sonething like that can solve the problem. > > > So to summarize: does anyone have any ideas or concerns about using > the DAV header on GET requests? > If no then I'll implement it in my browser extension and I'll ask > existing servers to add the header. > > Please vote with + or - > > Thank you > > > -- > Sergey Ponomarev <https://linkedin.com/in/stokito>, > stokito.com <http://stokito.com> >
Received on Sunday, 26 March 2023 20:22:02 UTC