Help on project design questions?

[Axel Hecht]

I looked at the web, curled up and tweeted. Way too many options, and
all look like heavyweight silos.

I have a project of projects, and need some help to get things off the
ground and on the web.

My first iteration is black-n-white tables like
http://mozilla-l10n.github.io/mqm-issues/.

Context:

We want to talk about quality of localizations, qualitatively and
quantitatively. Our starting point is MQM,
http://www.qt21.eu/mqm-definition/definition-2015-12-30.html, which
basically says:

There's a tree of issue types (see the first link for an overview), and
if you find an error in a localized product, you can assign that an
issue type, and a severity.

This gives us a basic language to exchange and gather quality data.

Thus far the good stuff. Now on to why I need more brains than mine.

We have some 180 issue types. So drop-downs don't work. This is a UX
problem, and there are no good solutions yet. I want something that
allows us to use different implementations of that UX and compare them.

We have multiple use-cases for this:

- classify bugs in l10n (like [0])
- classify suggestions in pontoon
- classify suggestions in pootle
- classify errors found in screenshots, like in
http://people.mozilla.org/~sarentz/fxios/screenshots/
- classify issues reported through transvision
- classify just translations, string by string, going through looooong lists

There are different facets of UX, too:
- determine with issue type applies
- cheat sheets to quickly help with those classifications
- docs, onboarding tours to those classifications
- use issue types in style guides for l10n

For the use in reviews of just strings and screenshots, we also want
UX/UI telemetry. Studies show that raters suffer from fatigue, but it's
unclear if that's just bad UX, or a general problem. Being able to track
rating-related events together with results might help to evaluate UX
and guidelines for testers. In particular, we're interested to learn
when we start getting errors in the error reports.

I'd love to get opinions on how to tackle all of these. Preferably in an
incremental fashion.

This is less about the actual UX of the individual pieces, but on which
foundation to implement the individual parts UX and put them together,
and how to get to a visual UI that let's us evaluate the options
properly, and not an html table badness.

Thanks

Axel

[0]
https://bugzilla.mozilla.org/report.cgi?x_axis_field=resolution&y_axis_field=component&z_axis_field=&query_format=report-table&short_desc_type=allwordssubstr&short_desc=&longdesc_type=allwordssubstr&longdesc=&bug_file_loc_type=allwordssubstr&bug_file_loc=&status_whiteboard_type=allwordssubstr&status_whiteboard=[transvision-feedback]&keywords_type=allwords&keywords=&bug_id=&bug_id_type=anyexact&votes=&votes_type=greaterthaneq&emailassigned_to1=1&emailtype1=exact&email1=&emailassigned_to2=1&emailreporter2=1&emailqa_contact2=1&emailtype2=exact&email2=&emailtype3=substring&email3=&chfieldvalue=&chfieldfrom=&chfieldto=Now&j_top=AND&f1=noop&o1=noop&v1=&format=table&action=wrap

[Axel Hecht]

And made it too easy to miss the point.

I don't care about user experience at this point, I care about developer
and project manager experience.

As a developer, I want to modify a score card and have people use it.

As a developer, I want to use a score card against a different type of
data to rate against (bug, string, screenshot, TBD)

As a project manager, I want to see user metrics and compare different
score cards against each other.

As a developer, I want to integrate a score card and a cheet sheat in my
php app.

As a developer, I want to integrate a score card and a cheet sheat in my
django app.

Things like that.

Axel

[Peter Bengtsson]

Just a little 2c drop; A drop-down with 180 options doesn't sound like
"drop-downs don't work" to me.
180 is not a big number. I can, on my desktop and my iphone scroll through
a list like that really fast. Also there's widgets like select2 [0] which
makes it really easy to turn that into something you can also filter down
by freetext search.

A problem might arise, performance wise, if you have 180 SQUARED or
something as your HTML will get enormous, but that's easy to solve with
some AJAX autocomplete on the drop-downs.

[0] select2.github.io

> _______________________________________________
> dev-webdev mailing list
> dev-w...@lists.mozilla.org
> https://lists.mozilla.org/listinfo/dev-webdev
>



--
Peter Bengtsson
Mozilla Tools & Services