On 9 April 2014 14:07, Steven Walling steven.walling@gmail.com wrote:
On Wed, Apr 9, 2014 at 9:33 AM, David Gerard dgerard@gmail.com wrote:
On 9 April 2014 17:30, Brian Wolff bawolff@gmail.com wrote:
That said, we shouldn't be afraid of making changes where we reasonably think they might be a good idea, even without evidence they actually are. You can't have data on everything. I just don't like "Well we are undoubtedly making things better for the reader" used as a counter argument to criticism when we simply don't know what it will do for the average reader.
Yes, it is the sort of statement that probably should not be used without being followed by a link to actual UI testing results.
I should follow up on this and say that no one working on the Beta Feature thinks it's a good idea to try and design typography that only works for people who aren't logged in/don't edit. The design goals listed at http://blog.wikimedia.org/2014/03/27/typography-refresh/ and https://www.mediawiki.org/wiki/Typography_refresh#Summary_of_changes are pretty universal to all users, as they should be.
I don't really think that when it comes to typography, either type of visitor to Wikimedia sites is more or less important when it comes to listening to feedback. Even if Nathan was right, sometimes it's hard for us to balance the two. What I said in reply to Risker is that I don't think there saying the change is a failure is fair or true, based on the level and kind of feedback we've been getting from both readers *and* editors.
Nobody, as best I can tell, has suggested that there be two different typographies, one for logged-in viewing and one for logged-out viewing.
Referring to the 4 points on the blog: I'm not certain why there's an urge to have consistency between the mobile and desktop versions; I can't think of a single mobile site I use that is truly consistent with its desktop version. However, I don't object to it fundamentally, and I think this was pretty much achieved, at least for English. However, given that we do not have reports of typography failures on mobile that we know exist for non-latin scripts on English, "consistency" doesn't seem to be met.
However, it's pretty obvious that there was a failure at #1, since the type didn't work properly on a whole pile of regularly used WMF scripts. We've all seen plenty of screenshots that showed what we can politely call a failure to be beautiful in certain configurations for latin-based scripts, as well; I suppose that would be a combination of points #1 and #3.
And has any testing been done with screen readers and dyslexia readers to test #4? I've not seen any reports of that. We do have some regular users of screen readers who would probably have been right there letting you know if there was a conflict. I have some anecdotal info that it does not interact well with at least one dyslexia reader, but no further details, and the user wouldn't provide me with a screenshot that I could share. Nonetheless, the evidence is lacking that this was tested for; if it's a principle goal, this needs to be done.
So... that sums up to three of the four goals not being met, and one having no evidence that it was met. This first release was a failure. That is okay; learn from it, but before you go further you must first restore the initial functionality to the projects that have identified problems...and any others that seem to have problems, even if they haven't told you about it. Someone should be checking them. Once you've returned their functionality, make fixing those issues in the typography upgrade the priority for subsequent releases. That is failing successfully.
Risker/Anne