Re: Design help requested: Certificate Chooser UI.

[Bastien Nocera <hadess hadess net>]

Hey David,

On Wed, 2016-05-11 at 13:08 +0100, David Woodhouse wrote:
> Tyagi is working on a GSoC project this year, implementing a
> certificate chooser which will probably live in the GCR library.
>
> We currently have a variety of incomplete and buggy implementations
> of
> this, spread across various places; especially NetworkManager UIs and
> the VPN plugins. All of them get *something* wrong — not implementing
> support for certain types of certificate/key file, or not supporting
> PKCS#11 for access to objects in gnome-keyring and hardware crypto
> tokens.
>
> The point of the project is to have a consistent UI for choosing a
> certificate (and corresponding private key), which everyone can use.
>
> This is https://bugzilla.gnome.org/show_bug.cgi?id=679860
<snip>

As nobody has replied, and you asked me to look at the mail, here it
is. Given the long mail and the incredibly specific application of it,
it's unsurprising that there were no answers, and I'll try to explain
why.

First, this is a bit too "micro" to be something you'd design, unless
you mean "slap lipstick on it". This widget would be used in VPN
plugins and the security configuration pages in the network settings
(nm-connection-editor and I gather GNOME's Network settings). We
probably wouldn't design this widget without seeing how it fits in the
larger network settings.

Second, I'll point you to a number of the existing design pages for an
example, such as:
https://wiki.gnome.org/Design/OS/FileSelection
or https://wiki.gnome.org/Design/OS/Emoji

This lists things like:
- use cases
- non-use-cases (things you explicitly don't want to support)
- a number of possible sticking points (possible technical limitations,
or existing workflows you need to take into account)
- the File selection page doesn't show that, as I guess the prior art
is well-known enough, but the Emoji page has a "relevant art" or "prior
art" section listing what that same widget looks like now, in GNOME and
other OSes (including the mobile ones) and desktops. You don't need to
cover everything, but OS-builtin equivalents for Android and iOS would
probably be useful.

You can use the GNOME Wiki to store all this while you're working on
it, which will help designers that will eventually look into it, to
have a bird's eye view of the problem space and the problem itself.

HTH