-
Notifications
You must be signed in to change notification settings - Fork 308
Split out identity verification from bank account connecting #3250
Conversation
Tagging |
297a663
to
db11a77
Compare
On a fresh database, when I login and navigate to /about/me/identity, I get this traceback:
|
Identity should be included in the navigation. Right now the only way to get to it is through the bank account page. |
I think fixing Gratipay's navigation needs its own PR. |
The real problem here is that the browsing tests didn't detect that, I figured out why and am working on a PR to fix them. |
db11a77
to
74e8621
Compare
|
|
Needs a rebase? |
- don't float, use `inline-block` for `.half`, prune the `clear` <div>s - drop the unnecessary <div> with the `constrain-width clearfix` classes - wrap the <label>s around the <input>s instead of using the `for` attribute - fix the indentation of the info at the end of the page - fix and complete the i18n of placeholders - prune styles for unused classes (`.disabled`, `.nav`, `.full`)
- i18n of messages - highlight the invalid fields and show the error message underneath it when focused - autofocus the first invalid field in the form - use notifications when we don't know which field to attach the error to
74e8621
to
c305d79
Compare
Confirmed. |
Tests passing locally. Reviewing UI ... |
Interesting that the tooltips I'm seeing on the Identity form are different than the ones screencapped from the credit card form at #3250 (comment). |
I don't see a way to remove my identity verification. |
I think we should add this to the profile nav, between Widgets and Settings. |
Next steps for me are to inspect what happens in Balanced when we post identity, and then read commits. |
|
I'm reading about form validation: http://designmodo.com/ux-form-validation/ |
So ... the inline error reporting we've added here is good for pattern mismatches such as DOB, but it's bad for signaling required fields, because, whereas a pattern can be matched as soon as it's entered, we can't know that a required field is missing until the moment the form is submitted. But the moment the form is submitted, we can and should say when all required fields are missing, not just the first. Firefox has okay behavior here. They highlight the missing fields, with a tooltip on the first: Chrome, however, doesn't highlight the missing fields: Are we doing something in CSS that prevents Chrome from highlighting missing fields? |
Required validation aside, it's shoddy to have some validation styled one way and some a different way. Can we use browser validation for all fields? |
Yes. Having some tooltips styled one way and others a different way is shoddy. Yes, we are quite shoddy in many ways. Let's not make our product worse. |
@Changaco We can punt on further error reporting clean-ups if you want to, since the main purpose of this PR is to separate the identity form from the bank account form. |
In fact, this is a perfect example of the "wagging the dog" antipattern that we've talked about before. To compete with closed companies, we "have to be as good or better on [...] features and UX" (IRC). We're not going to get there unless we really, truly own the difference between the product dog and the technology tail. Wrong: "Hey, I've got this technology available. Where can I use it?" Right: "Here's the experience I want my users to have. How do I deliver it?" |
I really don't get that. Would you have noticed the difference if I hadn't posted a screenshot of the red tooltip? Do you think users really care that one tooltip is white/gray and the other red? I can make ours gray if you want.
Compared to what we have now, this PR is still a big improvement even if the tooltips aren't of the same color. |
Yes.
Yes. Any competent product designer (by definition, so say I) will consciously care that the tooltips are not consistent, because it's a distinction without a difference. The distinction in the product's interface does not encode a difference in the product's mental model. Rather, it encodes a tooling limitation. The tooling limitation has no meaning to the user. The mental model of the product is where the user looks for meaning.
The Stanford study contributes to these guidelines, along with a number of other papers from the same era (1999-2002). Unfortunately, the links to the actual papers are broken. Admittedly, it's also old research, as noted:
|
Hrm. That's a bigger problem, since historically I haven't had much luck getting you to get things you don't get. ;-) I suppose the best I can do for now is to flag problems as I notice them, explain myself as best I can, and correct problems myself as much as I'm able. Hopefully over time we're able to attract and retain competent product designers to balance out our team. |
Yeah, and what's more, the changes to form validation aren't the point of this PR. Splitting Identity out from the bank account page is the point. I'm almost ready to merge, lemme go over it one more time ... |
Split out identity verification from bank account connecting
Closes #987.