- From: Dan Murphy <dmurph@google.com>
- Date: Thu, 11 Jun 2026 10:46:39 -0700
- To: Dan Murphy <dmurph@google.com>, Christian Liebel <christian.liebel@thinktecture.com>, public-webapps <public-webapps@w3.org>
- Message-ID: <CA+4qT32o6p1A-V4zMeVAoMqQCSyE0YDh7zQ=WyueC8s1p4usfA@mail.gmail.com>
Notes
-
https://github.com/w3c/manifest/pull/1218
-
having a bool that could be an enum often is something I regret when
coding, as in the future for established APIs this hinders adding new
values, and often the 'enum-label: enum value' is a more readable
description than a bool.
-
Dan
-
Don't care THAT much. Dan has slight preference for enum and no
bool, unless.. maybe 'installed'
-
application-context with an enum feels the most future proof.
-
Christian
-
agreed on enums over bools can be nice
-
also don't care that much
-
why not installed?
-
explanation from explainer
<https://github.com/MicrosoftEdge/MSEdgeExplainers/blob/main/ApplicationContextMediaFeature/explainer.md#alternatives-considered>
-
The name installed suggests a statement about global installation
state. A developer might reasonably expect (installed) to be true
if the app is installed on the device, even when viewed in a
browser tab.
In reality, the feature only matches when running inside an
installed app window. The name application-context makes this
distinction explicit, and describes the current context, not a global
property.
-
Dan: seems reasonable?
-
this also means that it is guaranteed to be interop, as apple
didn't want to reveal installed in the browser tab, and
chrome.. could? but
we likely want to align, and teh getInstalledRelatedApps()
api is still a
way to detect if the app is installed.
-
Christian - sounds good
-
https://github.com/w3c/manifest/pull/1219
-
The unapply thing seems nice to help explain reparenting behavior
that some user agents employ.
-
copilot comments seem reasonable.
-
seems like a good change to update the spec to modern concepts
Received on Thursday, 11 June 2026 17:46:58 UTC