[OpenID] Feedback requested: New OpenID RP login UX prototype

Tue Oct 27 02:49:45 UTC 2009

Just a small note.

Great effort and real-time improvements.   Feel much more comfortable
using these types of libraries when there is such great support and
follow through.

--Mike

On Mon, Oct 26, 2009 at 7:46 PM, Andrew Arnott <andrewarnott at gmail.com> wrote:
> It turns out that nasty redirect behavior that pulled you back to Verisign
> was due to a bug in Verisign's handling of checkid_immediate, which they are
> now aware of and investigating.
> So aside from that redirect issue, which only comes up if you click on
> Verisign at least once, it seems, all fixes this forum has suggested have
> been published.  So please feel free, everyone, to revisit
> http://openidux.dotnetopenauth.net/ and give fresh feedback (or remind me of
> something I may have missed in this discussion).
> --
> Andrew Arnott
> "I [may] not agree with what you have to say, but I'll defend to the death
> your right to say it." - S. G. Tallentyre
>
>
> On Fri, Oct 23, 2009 at 2:20 PM, Rabbit <rabbit at cyberpunkrock.com> wrote:
>>
>> Good improvements, Andrew.
>> Popups look great, plugin prompt did not appear.
>> As for the redirect issue, I don't know if this helps but looking at the
>> headers it checks each of these in sequence:
>> - Google?
>> - MyOpenID?
>> - Yahoo?
>> - Verisign?
>> At which point it redirects me to login at PIP.
>> Perhaps your prototype thinks I am signed into Versign.
>> =Rabbit
>> On Oct 23, 2009, at 2:23 PM, Andrew Arnott wrote:
>>
>> Hey Rabbit,
>> So I fixed the popup window size and the plugin prompt (please verify!).
>>  Also, I happened to run into the same Back->auto-redirect behavior you just
>> described.  It was in IE8 for me, and I have no idea why it's doing that.
>>  It doesn't usually (for me anyway).  But anyway, that's absolutely not by
>> design, and I'm going to chase that down and fix it.  Thanks for letting me
>> know it wasn't just a fluke on my machine. :)
>> Thanks for the extra details.
>> --
>> Andrew Arnott
>> "I [may] not agree with what you have to say, but I'll defend to the death
>> your right to say it." - S. G. Tallentyre
>>
>>
>> On Thu, Oct 22, 2009 at 10:15 PM, Rabbit <rabbit at cyberpunkrock.com> wrote:
>>>
>>> Inline comments.
>>> Peace.
>>> =Rabbit
>>> On Oct 22, 2009, at 10:27 PM, Andrew Arnott wrote:
>>>
>>> Thanks, Rabbit.
>>> Responses inline.
>>> --
>>> Andrew Arnott
>>> "I [may] not agree with what you have to say, but I'll defend to the
>>> death your right to say it." - S. G. Tallentyre
>>>
>>>
>>> On Thu, Oct 22, 2009 at 7:01 PM, Rabbit <rabbit at cyberpunkrock.com> wrote:
>>>>
>>>> I won't sugar coat this. If I encountered this interface in the wild I
>>>> would be furious. I hope you are able to take this criticism constructively
>>>> and improve the user experience.
>>>
>>> Sure, I don't want you to sugar coat it.  Although your tone throughout
>>> this email suggests that you thought I thought this was all finished and
>>> polished.  It's a "prototype", dude.
>>>
>>> I know it's a prototype. Apologies for my tone.
>>>>
>>>> FireFox on OSX receives multiple plugin download requests. Obviously
>>>> these are related to InfoCard but "explainable" from a technical perspective
>>>> doesn't translate to "acceptable" from a user perspective. There was also no
>>>> indication as to what the plugins were for and when I attempted to install
>>>> them it failed. That latter criticism may be an issue with FireFox or OSX.
>>>
>>> I totally agree.  This doesn't happen on Windows, so I didn't see these
>>> problems.  But another mac user reported seeing the same thing.  I have an
>>> idea of how to fix this so please try again in a couple days and tell me if
>>> the problem hasn't gone away.  Yes, it's InfoCard related, but it was not my
>>> intention to throw up all kinds of unpleasantness for those who don't have
>>> InfoCard support.  It's supposed to be a very quiet "light-up" scenario if
>>> you have it, and completely and quietly missing if you don't.
>>>
>>> Realistically, it's not a big deal. In FireFox it appears as a slide down
>>> bar the same that happens when asked to remember a password. I don't know
>>> enough about integrating with InfoCard, the reason I pointed it out is that
>>> there may be some Mozilla-oriented headers that will point its plugin
>>> searcher in the right direction.
>>>
>>>
>>>>
>>>> Almost every popup window was irritating and appeared broken. A standard
>>>> user would most likely think they were tricked into clicking an
>>>> advertisement. This somewhat extends into my general criticism towards the
>>>> entire popup standard being promoted (which I won't get into) but even with
>>>> the popup approach there is room for improvement here. If you know the
>>>> OpenID provider you are sending me to, and they do not offer a simple UI for
>>>> popups, at least size the window to avoid horizontal scrollbars.
>>>>
>>>>
>>>> http://s978.photobucket.com/albums/ae262/rabbyte/?action=view&current=Picture3.png
>>>>
>>>> http://s978.photobucket.com/albums/ae262/rabbyte/?action=view&current=Picture4.png
>>>>
>>>> http://s978.photobucket.com/albums/ae262/rabbyte/?action=view&current=Picture5.png
>>>
>>> Working within the popup UI is still being developed for the OPs, and
>>> each of them are responsible for making it look good.  Google supports it
>>> fully, as I'm sure you noticed.  Yahoo claims to support it but don't get
>>> the window size right according to the UI extension draft spec and I hope
>>> they fix that or the spec (I don't care which).  myopenid and verisign don't
>>> support it at all, and the larger popup window size I give non-supporters
>>> isn't apparently big enough for their large window demands.  I'll see if I
>>> can fix that by just giving them a bigger window.  Ideally, I hope this
>>> encourages these OPs to shrink their UI so it fits in smaller windows.
>>>
>>> imho, I would prefer a new tab being opened if there is no known window
>>> dimensions. Unfortunately, that creates an even more inconsistent user
>>> experience.
>>>
>>> Besides that, myopenid and Verisign aren't likely to be displayed in the
>>> final UI kit that I'm building until they meet the guidelines I wrote up.
>>>  RPs can always add them though.
>>>>
>>>> You'll also notice that the MyOpenID popup is slightly hiding its
>>>> security feature on the top right. There's just no excuse for that.
>>>
>>> I agree.  But see above.
>>>>
>>>> The automatic redirecting is absolutely atrocious! Even just trying to
>>>> test this out for feedback was excruciating. It is impossible to change my
>>>> choice once I have made a choice. Page load tries to redirect. Clicking
>>>> login tries to redirect. My history gets mangled and I can't hit the back
>>>> button. Between the popups and the redirects, a standard user might (and
>>>> should) think they have a virus.
>>>
>>> Now here you've lost me.  Can you explain more precisely what's going on?
>>>  Maybe it's a Mac thing (which I would of course still want to fix), but why
>>> do you say it's impossible to change your choice once you've made one?  If
>>> you click one Provider and log in, you absolutely can pick another provider
>>> the next time you log in (although we make them appear grayer than the rest
>>> to discourage this).  But remember this is targeted at normal users -- it's
>>> not targeted for testing multiple OPs.  So a normal user would want to keep
>>> clicking the same button in order to avoid splintering their identity.
>>>  That's a common complaint about OpenID: "Which button did I click on last
>>> time?"  Or more practically: "I logged in [with the wrong button] and now
>>> all my stuff is gone!"  Although the other buttons are gray, you can still
>>> click them.
>>>
>>> I thought that was an intentional feature. I made a provider choice,
>>> clicked through to the provider, did not login (maybe that makes the
>>> difference?), went back to the original page and each time it would load for
>>> a second then instantly redirect me back to the OP I chose.
>>> I had to load your page, hit stop quickly, click Login, hit stop quickly,
>>> then choose another provider. (You can understand my frustration! hah)
>>> I wouldn't mind the auto-redirects if there were a visual countdown such
>>> as:
>>> "You previously chose X as your provider. Redirecting in 3....2....1..."
>>> with a cancel button to make another choice.
>>> I don't really see how this could be an OS-specific issue. If you have
>>> trouble duplicating this, contact me directly, I'll try to help pinpoint it.
>>>
>>> Now what about this "page load tries to redirect".  What does that mean?
>>>  I've seen sites where the Back button takes you to a page that redirects
>>> you "forward" again, which is very aggravating.  But on my browsers, this
>>> doesn't happen.  The back button works as expected.  Can you elaborate about
>>> what's broken?
>>>>
>>>> It also wouldn't hurt to provide a little information about OpenID since
>>>> it is an option. Even linking to a tutorial site to provide more information
>>>> would be helpful (hey! and you could use another popup!). Ok, maybe that
>>>> last remark was a little mean. I appreciate the effort. I really need to put
>>>> together a demo. I really hope you found my feedback useful.
>>>
>>> Here I disagree with you, but your opinion is appreciated nonetheless.
>>>  This is not a "promote OpenID" design.  Some of the loudest feedback I hear
>>> from users who fail to log into RPs is that they don't know what OpenID is
>>> and they leave.  This UI is designed for maximum user conversions to the
>>> RP's services, not to OpenID.  RPs want users to log in -- they don't care
>>> whether users know what OpenID is, and users don't visit random RPs to learn
>>> about what OpenID is.
>>>
>>> I only suggested it as OpenID is an option. It don't think it would hurt
>>> since you're already expanding the window with an input field to have a
>>> small link saying "What is this?" After all, that could easily be a path of
>>> discovery for many people who just haven't heard of OpenID yet but would
>>> love if they understood it.
>>>
>>> A design requirement is to keep the UI as simple as possible, while
>>> providing flexibility where needed for power users.  So links to learn about
>>> OpenID will be limited to what the RP's attorneys insist on saying for
>>> liability reasons.
>>> Yes, your feedback was helpful -- and can be more so if you can provide
>>> some more detail about the issues I asked for more detail about.
>>>>
>>>> =Rabbit
>>>>
>>>
>>>
>>
>>
>
>
> _______________________________________________
> general mailing list
> general at lists.openid.net
> http://lists.openid.net/mailman/listinfo/openid-general
>
>