-
Notifications
You must be signed in to change notification settings - Fork 4
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Revise Login Architecture & UX #53
Comments
Found a few bugs in different places recently. Starting to compile a list here. UI
Server
|
@jbuck @simonwex @cadecairos @secretrobotron Feel free to add notes here, my list above is by no means exhaustive. |
@adamlofting See last item (instrumentation) above. |
@thisandagain Can you explain more about this item? The adapters were built with de-duplication in mind. |
Ah! Good to know... I think I might not quite understand the architecture as well as I thought. @simonwex and @jbuck had some thoughts on this as well, but the general idea was to deprecate the |
👍 |
Suggested re-scope for pre-MWC:
@thisandagain @simonwex considering that we're pulling out auth from webmaker-app for MWC, above list look ok for this sprint? @jbuck will write up another bug to address post-MWC login plans |
Milestone has been created in the mozilla/webmaker.org repo: https://github.com/mozilla/webmaker.org/milestones/Login |
Note: Using something like BriteVerify would require (likely significant) vendor review before trial, and we'd probably want to look at the larger scope of email verification in the context of mailing list systems (not that they need to use the same system, but there are two use cases with similar challenges). |
@davidascher do you know if there has been any similar work done by Moz teams to do part of what BriteVerify does? Thunderbird auto-configuration comes to mind. |
Status update:
|
I doubt we have alternatives to this particular service, as it requires
effectively tracking email delivery. There may be alternatives that we are
using as part of some of the email campaigns that @valianttry has been
doing.
|
Just to note @valianttry is on PTO. Though I'm pretty sure we don't have anything like this for our BSD mailing lists right now. |
Yet another reason to work through vendor review with whoever is best in
class for bounce management.
|
@secretrobotron For the kickoff meeting let's include @hannahkane @HPaulJohnson and @simonwex so we can confirm alignment w/ the teach site and getting vendor selection / review rolling. |
Summary for easy reading:
|
Architecture question which impacts analytics (and UX):
This is a consideration for how we track login for user on *.webmaker.org vs a user on teach.mozilla.org (GA account IDs change and so on). |
Follow-up note from our 1:1 – Given that we have some content that needs to vary here as well (value proposition & logo) we may want to consider having one service that listens on two domains or URLs. That would make distinguishing between referring links easier as well as possibly resolve some of the branding issues ( |
Notes from hangin' out with @hannahkane, @jbuck, @thisandagain, @Pomax & @cassiemc
Deliverables:
|
It's worth mentioning that we should add the teach.mozilla.org site to a whitelist. This means the user won't be prompted with a decision as is often the case in typical (and problematic) OAuth flow. @adamlofting do we have a preferred method for measuring user-perceived response-time for the login steps? By choosing the redirect approach, I think this will be a very important factor wrt conversion rate. |
We don't have a preferred method for perceived-load time, but I'd encourage a dev iteration to work specifically on this issue if we can afford it. Our own perception as users, along with some browser-rendering-time data should make a good dent on this. Slightly slower loading time can lose a percent or three of the conversion rate, so it's important, but it's a lesser impact than UX, messaging and brand. Really slow response times are more problematic, but our standard of dev work is high by default, so I don't have that as a concern 👍 Over time, I'd love to see this response-time 'polishing' as a dev step in all the interactions we're building. |
@secretrobotron I don't think the milestone listed above (https://github.com/mozilla/webmaker.org/milestones/Login) is showing all open issues here -- they seem to just be design tickets. @alicoding and I were discussing this morning and would like to break the design tickets into more issues so that it's clearer what is finalized and what is still WIP. I can do that, but not sure if this is the right milestone. Can you direct me to the right repo / milestone, please? |
Yeah, my bad. I assigned a couple to you and marked them with design. Here's the milestone though: https://github.com/mozilla/id.webmaker.org/issues?q=is%3Aopen+is%3Aissue+milestone%3A%22Teach+Implementation%22 |
Great, thanks @secretrobotron! I updated the ticket overview in the first comment with this link and am going to move design tickets over there / break them down more. |
Spoke with @thisandagain and @hannahkane today about what we'll ship this week.
|
Tracking next steps specific to the teach.mozilla.org implementation in #404 |
After reviewing the performance of Login v3 for a few months, we've observed a few issues specific to user experience (conversion rate) and developer ergonomics. In order to more easily resolve both, this sprint is to address some of the issues with the architecture, developer ergonomics, and user experience of login.
User Experience Issues
Information Requirements
Architecture Requirements
angular
andvanilla js
versions of Login to a single non-framework (vanilla) dependent moduleMilestone
https://github.com/mozilla/webmaker.org/milestones/Login
https://github.com/mozilla/id.webmaker.org/issues?q=is%3Aopen+is%3Aissue+milestone%3A%22Teach+Implementation%22
RACI
Phase: Build / Ship
Owner: @secretrobotron
Decision: @davidascher
Lead design: @cassiemc
Lead dev: @jbuck
Dev: @Pomax
Dev: @cadecairos
Quality: @simonwex
Metrics: @adamlofting
The text was updated successfully, but these errors were encountered: