Discussion:
[Design] Constructive vs. progressive buttons
May Tee-Galloway
2015-10-21 01:14:48 UTC
Permalink
There is a conversation going on here about consolidating constructive
and progressive buttons: https://phabricator.wikimedia.org/T110555

I want to find out if any of you have used these buttons in your
interface and have done any testing with your users to see if they
understand the difference between the two. Also, if you have used it
in other ways that has been helpful to your users, chime in!

Progressive (blue) conveys to the user that they are starting or
continuing a multi-step process.
Constructive (green) conveys to the user they are completing a single
or multi-step process. In most case Constructive color shows the user
what will happen. In others it is feedback that the action has
completed. For example, thanking a user, adding a page to a watch
list.


--
Related discussion about button consolidation:
https://phabricator.wikimedia.org/T110565


mm
Trevor Parscal
2015-10-21 01:31:35 UTC
Permalink
I would just clarify that constructive buttons are used to indicate an
action creates something, rather than just modifying it. Similarly, we use
destructive buttons to indicate the inverse, that something is actually
being removed. Constructive and destructive don't have anything in
particular to do with multi-step processes, only what the process does.

- Trevor
Post by May Tee-Galloway
There is a conversation going on here about consolidating constructive
and progressive buttons: https://phabricator.wikimedia.org/T110555
I want to find out if any of you have used these buttons in your
interface and have done any testing with your users to see if they
understand the difference between the two. Also, if you have used it
in other ways that has been helpful to your users, chime in!
Progressive (blue) conveys to the user that they are starting or
continuing a multi-step process.
Constructive (green) conveys to the user they are completing a single
or multi-step process. In most case Constructive color shows the user
what will happen. In others it is feedback that the action has
completed. For example, thanking a user, adding a page to a watch
list.
--
https://phabricator.wikimedia.org/T110565
mm
_______________________________________________
Design mailing list
https://lists.wikimedia.org/mailman/listinfo/design
Jonathan Morgan
2015-10-21 14:59:35 UTC
Permalink
Post by Trevor Parscal
I would just clarify that constructive buttons are used to indicate an
action creates something, rather than just modifying it. Similarly, we use
destructive buttons to indicate the inverse, that something is actually
being removed. Constructive and destructive don't have anything in
particular to do with multi-step processes, only what the process does.
Hi Trevor,

May was talking about constructive vs. PROGRESSIVE buttons here. Did you
misread her email, or are you making a secondary point? It's unclear from
your message.

Best,
Jonathan
Post by Trevor Parscal
- Trevor
Post by May Tee-Galloway
There is a conversation going on here about consolidating constructive
and progressive buttons: https://phabricator.wikimedia.org/T110555
I want to find out if any of you have used these buttons in your
interface and have done any testing with your users to see if they
understand the difference between the two. Also, if you have used it
in other ways that has been helpful to your users, chime in!
Progressive (blue) conveys to the user that they are starting or
continuing a multi-step process.
Constructive (green) conveys to the user they are completing a single
or multi-step process. In most case Constructive color shows the user
what will happen. In others it is feedback that the action has
completed. For example, thanking a user, adding a page to a watch
list.
--
https://phabricator.wikimedia.org/T110565
mm
_______________________________________________
Design mailing list
https://lists.wikimedia.org/mailman/listinfo/design
_______________________________________________
Design mailing list
https://lists.wikimedia.org/mailman/listinfo/design
--
Jonathan T. Morgan
Senior Design Researcher
Wikimedia Foundation
User:Jmorgan (WMF) <https://meta.wikimedia.org/wiki/User:Jmorgan_(WMF)>
Trevor Parscal
2015-10-21 15:03:53 UTC
Permalink
Sorry if I'm unclear, but yes I fully understand.

I felt that her description somewhat misrepresented the purpose of
constructive in relation to progressive.

Essentially, I read her description as saying constructive was simply a final
step, when in fact it's designed to provide the user a hint that something
will be created - no matter the step.

- Trevor
Post by Jonathan Morgan
Post by Trevor Parscal
I would just clarify that constructive buttons are used to indicate an
action creates something, rather than just modifying it. Similarly, we use
destructive buttons to indicate the inverse, that something is actually
being removed. Constructive and destructive don't have anything in
particular to do with multi-step processes, only what the process does.
Hi Trevor,
May was talking about constructive vs. PROGRESSIVE buttons here. Did you
misread her email, or are you making a secondary point? It's unclear from
your message.
Best,
Jonathan
Post by Trevor Parscal
- Trevor
On Tue, Oct 20, 2015 at 6:14 PM, May Tee-Galloway <
Post by May Tee-Galloway
There is a conversation going on here about consolidating constructive
and progressive buttons: https://phabricator.wikimedia.org/T110555
I want to find out if any of you have used these buttons in your
interface and have done any testing with your users to see if they
understand the difference between the two. Also, if you have used it
in other ways that has been helpful to your users, chime in!
Progressive (blue) conveys to the user that they are starting or
continuing a multi-step process.
Constructive (green) conveys to the user they are completing a single
or multi-step process. In most case Constructive color shows the user
what will happen. In others it is feedback that the action has
completed. For example, thanking a user, adding a page to a watch
list.
--
https://phabricator.wikimedia.org/T110565
mm
_______________________________________________
Design mailing list
https://lists.wikimedia.org/mailman/listinfo/design
_______________________________________________
Design mailing list
https://lists.wikimedia.org/mailman/listinfo/design
--
Jonathan T. Morgan
Senior Design Researcher
Wikimedia Foundation
User:Jmorgan (WMF) <https://meta.wikimedia.org/wiki/User:Jmorgan_(WMF)>
Bartosz Dziewoński
2015-10-21 17:28:29 UTC
Permalink
Post by Trevor Parscal
Essentially, I read her description as saying constructive was simply
a final step, when in fact it's designed to provide the user a hint that
something will be created - no matter the step.
What May says does seem to better match how these are used in practice.
For example, on https://phabricator.wikimedia.org/T113493 several people
requested the green constructive button on the form on Special:MovePage,
which doesn't "create" anything.
--
Bartosz Dziewoński
Bartosz Dziewoński
2015-10-21 17:32:35 UTC
Permalink
I think I said it elsewhere a few times, but I believe we have way too
many kinds of buttons. We should definitely consolidate "constructive"
and "progressive", especially since judging by this email thread, nobody
really knows what exactly the difference is supposed to be.
--
Bartosz Dziewoński
May Tee-Galloway
2015-10-21 17:48:10 UTC
Permalink
nobody really knows what exactly the difference is supposed to be.
I've had a few conversations on how to use these buttons, people seem
abit confused and I started to wonder myself if the intended impact is
greater than the confusion caused. The way we described its usage
could also be a factor.

However, there may be an alternative use to green buttons that might
still warrant its existence. That's what I'd like to learn from
everyone. I'll report back if I find something too.

mm

On Wed, Oct 21, 2015 at 10:32 AM, Bartosz Dziewoński
I think I said it elsewhere a few times, but I believe we have way too many
kinds of buttons. We should definitely consolidate "constructive" and
"progressive", especially since judging by this email thread, nobody really
knows what exactly the difference is supposed to be.
--
Bartosz Dziewoński
_______________________________________________
Design mailing list
https://lists.wikimedia.org/mailman/listinfo/design
--
mm
Gergo Tisza
2015-10-21 18:59:03 UTC
Permalink
​One problem I have run into recently is that for a complex form you are
not necessarily able to tell which step is the last. E.g. after you submit
the login form and MediaWiki verifies your credentials, depending on your
user settings you might or might not be presented with a two-factor
challenge; so submitting the user name and password might or might not be
the last step of the form. (Arguably login should not be constructive in
the first place, but it is now. In any case, similar problems could be
present with the user registration form, which does create something.)

Personally, I agree with Bartosz that having four button types (five or
more if we include silent buttons) just makes the interface confusing.
Trevor Parscal
2015-10-21 19:05:07 UTC
Permalink
At some point a couple of years ago, the idea of encoding behavior hints
into the colors of the user interface became a major part of MediaWiki UI
styling, and OOjs UI had something similar so the two concepts merged. The
problem has always been when you start using the buttons in practice, the
interface is really loud and confusing because it's blasted with giant
bright colored squares on otherwise very light white and gray controls.
This has always been a challenge, and in OOjs UI we ended up adding
"primary" as an additional designation so only the primary action was a
bright color and other buttons were quieter with only colored text and
outlines. Which is actually a 4 x 2 matrix of buttons. Owch.

I look forward to May returning with some analysis.
Post by Gergo Tisza
​One problem I have run into recently is that for a complex form you are
not necessarily able to tell which step is the last. E.g. after you submit
the login form and MediaWiki verifies your credentials, depending on your
user settings you might or might not be presented with a two-factor
challenge; so submitting the user name and password might or might not be
the last step of the form. (Arguably login should not be constructive in
the first place, but it is now. In any case, similar problems could be
present with the user registration form, which does create something.)
Personally, I agree with Bartosz that having four button types (five or
more if we include silent buttons) just makes the interface confusing.
_______________________________________________
Design mailing list
https://lists.wikimedia.org/mailman/listinfo/design
S Page
2015-10-21 19:46:05 UTC
Permalink
* There are always gray areas in applying a design guide, that's not a good
argument for consolidation. Whether a particular URL should be a link or a
button is often tricky, but we still have buttons.
* Users not understanding the design is no argument for consolidation, I
bet users get an 'D' at best on Material and every other human interest
guideline.

I like Trevor's idea that Constructive makes something. Maybe that's why
the Special:Search button is progressive, or maybe the argument was it
leads to something else. And the way progressive buttons in the middle of a
dialog shift to a constructive button at the [Let's Roll!] stage is a great
UX thing. And Green for Thanks, because in an unfriendly often toxic
"community" it's a nice Constructive thing to do, is a happy byproduct of
the naming conventions. That's at least three uses that go away if we
consolidate Constructive and Progressive.
the interface is really loud and confusing because it's blasted with
giant bright colored squares on otherwise very light white and gray controls

"Confusing" doesn't follow at all from your aesthetic response to the
design. As I type this I'm "blasted with a giant bright colored [Send]
square", but I think Google knows what it's doing. There are many ways to
tamp down the Skittles appearance of the MediaWiki UI and the style guide
[1] is clear: "Under no circumstance should an interface display more that
one *Primary* button colored with *Intention*." If Quiet and Neutral aren't
sufficient Shahyar prototyped at least two more excellent ideas (I can't
find the link right now...)

If any clarification comes out of this discussion, the
http://livingstyleguide.wmflabs.org [2] should be our one and only
guideline. It can acknowledge grey areas without getting bogged down in
details. Maybe it can have footnotes or § links to Solomonic appendices on
edge cases.

Cheers,

[1] http://livingstyleguide.wmflabs.org/wiki/Buttons#Sets_of_buttons
[2] See http://livingstyleguide.wmflabs.org/wiki/Main_Page#Color and
livingstyleguide.wmflabs.org/wiki/Buttons#Button_Styles
At some point a couple of years ago, the idea of encoding behavior hints
into the colors of the user interface became a major part of MediaWiki UI
styling, and OOjs UI had something similar so the two concepts merged. The
problem has always been when you start using the buttons in practice, the
interface is really loud and confusing because it's blasted with giant
bright colored squares on otherwise very light white and gray controls.
This has always been a challenge, and in OOjs UI we ended up adding
"primary" as an additional designation so only the primary action was a
bright color and other buttons were quieter with only colored text and
outlines. Which is actually a 4 x 2 matrix of buttons. Owch.
I look forward to May returning with some analysis.
Post by Gergo Tisza
​One problem I have run into recently is that for a complex form you are
not necessarily able to tell which step is the last. E.g. after you submit
the login form and MediaWiki verifies your credentials, depending on your
user settings you might or might not be presented with a two-factor
challenge; so submitting the user name and password might or might not be
the last step of the form. (Arguably login should not be constructive in
the first place, but it is now. In any case, similar problems could be
present with the user registration form, which does create something.)
Personally, I agree with Bartosz that having four button types (five or
more if we include silent buttons) just makes the interface confusing.
_______________________________________________
Design mailing list
https://lists.wikimedia.org/mailman/listinfo/design
_______________________________________________
Design mailing list
https://lists.wikimedia.org/mailman/listinfo/design
--
=S Page WMF Tech writer
Isarra Yos
2015-10-21 20:20:38 UTC
Permalink
Post by S Page
Users not understanding the design is no argument for consolidation, I
bet users get an 'D' at best on Material and every other human
interest guideline.
If users aren't the ones we are designing for, then who is? Wasn't the
whole point of the colours is to call the users' attention to them, and
the point of differentiating is to tell the users that these are
different things?

Otherwise, what is the purpose of any of this?
S Page
2015-10-23 18:28:10 UTC
Permalink
... every other human interest guideline.
Sorry, "Interface". Human interest guidelines lead to the horror of Daily
Mail Online :-)

If Quiet and Neutral aren't sufficient Shahyar prototyped at least two more
excellent ideas (I can't find the link right now...)
http://area51.yar.gs/wmf/flow1/ To repeat, a page would only use a subset
of these at any one time, ignore the overstuffed appearance.
Post by S Page
Users not understanding the design is no argument for consolidation, I
bet users get an 'D' at best on Material and every other human interest
guideline.
If users aren't the ones we are designing for, then who is?
Obviously design is for users. But having a pleasant productive time using
a design and understanding its nuances are very different things. "Testing
with your users to see if they understand the difference between [two
button colors]" seems crazy (unless I misunderstand what designers mean by
"understand"). Does Google test to see if users understand the difference
between the rounded edge and the shadowed edge in Material? I really doubt
it. Different colors and treatments provide different experiences, and
there are guidelines when to use them.
Wasn't the whole point of the colours is to call the users' attention to
them, and the point of differentiating is to tell the users that these are
different things?
Sort of, and yes.
--
=S Page WMF Tech writer
May Tee-Galloway
2015-10-23 18:39:41 UTC
Permalink
Designers, anything from you guys?
Post by S Page
... every other human interest guideline.
Sorry, "Interface". Human interest guidelines lead to the horror of Daily
Mail Online :-)
If Quiet and Neutral aren't sufficient Shahyar prototyped at least two
more excellent ideas (I can't find the link right now...)
http://area51.yar.gs/wmf/flow1/ To repeat, a page would only use a subset
of these at any one time, ignore the overstuffed appearance.
Post by S Page
Users not understanding the design is no argument for consolidation, I
bet users get an 'D' at best on Material and every other human interest
guideline.
If users aren't the ones we are designing for, then who is?
Obviously design is for users. But having a pleasant productive time using a
design and understanding its nuances are very different things. "Testing
with your users to see if they understand the difference between [two button
colors]" seems crazy (unless I misunderstand what designers mean by
"understand"). Does Google test to see if users understand the difference
between the rounded edge and the shadowed edge in Material? I really doubt
it. Different colors and treatments provide different experiences, and there
are guidelines when to use them.
Wasn't the whole point of the colours is to call the users' attention to
them, and the point of differentiating is to tell the users that these are
different things?
Sort of, and yes.
--
=S Page WMF Tech writer
_______________________________________________
Design mailing list
https://lists.wikimedia.org/mailman/listinfo/design
--
mm
Isarra Yos
2015-10-24 00:02:25 UTC
Permalink
Post by S Page
Obviously design is for users. But having a pleasant productive time
using a design and understanding its nuances are very different
things. "Testing with your users to see if they understand the
difference between [two button colors]" seems crazy (unless I
misunderstand what designers mean by "understand"). Does Google test
to see if users understand the difference between the rounded edge and
the shadowed edge in Material? I really doubt it. Different colors and
treatments provide different experiences, and there are guidelines
when to use them.
That's nuances, details. Nevermind nuances. The question here is, on
whatever level, does the distinction help the user? Have users noticed
that there is a difference? Has it meant anything to them? Has anyone
looked into this?

If it's shown to help, that's all we need. We have a justification for
the maintenance overhead and stuff. And cookies. Always cookies. We
should all go get some cookies.
Pau Giner
2015-10-26 19:43:38 UTC
Permalink
I think one of the underlying aspects here is the idea of side-effects and
safe navigation.

We learnt by using the web that buttons and links had generally different
expectations. Links have been used for navigation purposes, while buttons
are associated with commands which affect the status of the system. These
associations make us to click links without having to fear any unexpected
change in the system. This allows to navigate more fluently when those
assumptions are met, and get into trouble when those are broken (e.g., a
delete link in the middle of other navigation links).

I think the purpose of colour buttons is to set similar expectations. We
highlight the next logical step and indicate whether some new content will
be created (green) or destroyed (red) as an outcome. Creating and deleting
content, even if these actions can be undone seems worth some
considerations in an environment where content is public and edited
collaboratively by many.
So I think that setting these expectations is useful on helping the user to
determine on which decisions to put more thought and which ones to do more
fluently.


I think we can design great experiences with or without this pattern, but
this is a pattern that requires to be applied consistently to work. We need
to consider also that the resulting effects on the user affect more their
intuition than their conscious thoughts. So while I think it can improve
the user experience, it is not something easy to test just by asking if the
user can figure out why the button is green instead of blue, even less if
the user has not been exposed to a consistent application of the pattern
for a while.

If the problem is that it is not easy to be applied consistently, we may
consider clarifying the guidelines to indicate the cases where each one
should use with as unambiguous definitions and clear examples as possible.

Google use of colour buttons has been mentioned. In this talk
<https://vimeo.com/29965463> (at 26:40) the code colour they used is
explained: red for creating something, blue for do/confirm actions such as
search, and green fro actions with an audience such as share. The fact
Google has been using this approach does not mean that we need to follow,
but it illustrates that associating colours to certain types of actions is
not an unprecedented idea in the design of user interfaces.


Pau
Post by Isarra Yos
Post by S Page
Obviously design is for users. But having a pleasant productive time
using a design and understanding its nuances are very different things.
"Testing with your users to see if they understand the difference between
[two button colors]" seems crazy (unless I misunderstand what designers
mean by "understand"). Does Google test to see if users understand the
difference between the rounded edge and the shadowed edge in Material? I
really doubt it. Different colors and treatments provide different
experiences, and there are guidelines when to use them.
That's nuances, details. Nevermind nuances. The question here is, on
whatever level, does the distinction help the user? Have users noticed that
there is a difference? Has it meant anything to them? Has anyone looked
into this?
If it's shown to help, that's all we need. We have a justification for the
maintenance overhead and stuff. And cookies. Always cookies. We should all
go get some cookies.
_______________________________________________
Design mailing list
https://lists.wikimedia.org/mailman/listinfo/design
--
Pau Giner
Senior User Experience Designer
Wikimedia Foundation
Nick Wilson (Quiddity)
2015-10-26 20:16:50 UTC
Permalink
See also https://phabricator.wikimedia.org/T116549 ("Provide a color
palette and design for buttons that are purely highlighted links, to
distinguish them from actual UI buttons") which I think needs designer and
developer input.
Post by Pau Giner
I think one of the underlying aspects here is the idea of side-effects and
safe navigation.
We learnt by using the web that buttons and links had generally different
expectations. Links have been used for navigation purposes, while buttons
are associated with commands which affect the status of the system. These
associations make us to click links without having to fear any unexpected
change in the system. This allows to navigate more fluently when those
assumptions are met, and get into trouble when those are broken (e.g., a
delete link in the middle of other navigation links).
I think the purpose of colour buttons is to set similar expectations. We
highlight the next logical step and indicate whether some new content will
be created (green) or destroyed (red) as an outcome. Creating and deleting
content, even if these actions can be undone seems worth some
considerations in an environment where content is public and edited
collaboratively by many.
So I think that setting these expectations is useful on helping the user
to determine on which decisions to put more thought and which ones to do
more fluently.
I think we can design great experiences with or without this pattern, but
this is a pattern that requires to be applied consistently to work. We need
to consider also that the resulting effects on the user affect more their
intuition than their conscious thoughts. So while I think it can improve
the user experience, it is not something easy to test just by asking if the
user can figure out why the button is green instead of blue, even less if
the user has not been exposed to a consistent application of the pattern
for a while.
If the problem is that it is not easy to be applied consistently, we may
consider clarifying the guidelines to indicate the cases where each one
should use with as unambiguous definitions and clear examples as possible.
Google use of colour buttons has been mentioned. In this talk
<https://vimeo.com/29965463> (at 26:40) the code colour they used is
explained: red for creating something, blue for do/confirm actions such as
search, and green fro actions with an audience such as share. The fact
Google has been using this approach does not mean that we need to follow,
but it illustrates that associating colours to certain types of actions is
not an unprecedented idea in the design of user interfaces.
Pau
Post by Isarra Yos
Post by S Page
Obviously design is for users. But having a pleasant productive time
using a design and understanding its nuances are very different things.
"Testing with your users to see if they understand the difference between
[two button colors]" seems crazy (unless I misunderstand what designers
mean by "understand"). Does Google test to see if users understand the
difference between the rounded edge and the shadowed edge in Material? I
really doubt it. Different colors and treatments provide different
experiences, and there are guidelines when to use them.
That's nuances, details. Nevermind nuances. The question here is, on
whatever level, does the distinction help the user? Have users noticed that
there is a difference? Has it meant anything to them? Has anyone looked
into this?
If it's shown to help, that's all we need. We have a justification for
the maintenance overhead and stuff. And cookies. Always cookies. We should
all go get some cookies.
_______________________________________________
Design mailing list
https://lists.wikimedia.org/mailman/listinfo/design
--
Pau Giner
Senior User Experience Designer
Wikimedia Foundation
_______________________________________________
Design mailing list
https://lists.wikimedia.org/mailman/listinfo/design
--
Nick Wilson (Quiddity)
Community Liaison
Wikimedia Foundation
Gergo Tisza
2015-10-26 21:37:16 UTC
Permalink
Post by Pau Giner
I think the purpose of colour buttons is to set similar expectations. We
highlight the next logical step and indicate whether some new content will
be created (green) or destroyed (red) as an outcome. Creating and deleting
content, even if these actions can be undone seems worth some
considerations in an environment where content is public and edited
collaboratively by many.
The create/destroy/other split doesn't match the "dangerousness" of the
actions well, though. The typical way for a non-admin user to make a mess
is via page move (which usually cannot be reverted without admin rights)
and merge-type actions (wikidata item merge, or adding an interwiki link on
Wikipedia that causes different concepts to be merged). The most dangerous
actions with admin rights are probably user blocking and page history
merge. None of those are constructive or destructive in the sense the UI
uses those terms.
n***@gmail.com
2015-11-10 01:14:37 UTC
Permalink
We highlight the next logical step and indicate whether some new content
will be created (green) or destroyed (red) as an outcome. Creating and
deleting content, even if these actions can be undone seems worth some
considerations in an environment where content is public and edited
collaboratively by many.
Though this is correct, at times, it is very difficult to put an action
into the bucket of "creating" and "progressing" there are grey areas here
since we are talking about abstracted concepts and it's difficult to make
it binary. it is possible, but difficult.

That brings me to the users of these concepts. If i am not wrong, there are
three parties involved.

1. Designers
2. Developers
3. Users

For designers, things like progressive and constructive makes sense but a
on semantic level. for independent developers without any design help, the
distinction between constructive and progressive might make things
confusing and complicated. and as far as a I know, users don't perceive
their tasks to be progressive or constructive right way.

Of course, the semantics that were thought before implementing it our
buttons this way made sense at the time and i think it still can be
justified but it seems like there is a fine line between two which makes it
difficult to convey.

If the value coming from having these distinctions is less than the
confusion that we are causing then maybe it's time not have these different
conventions.
Post by Pau Giner
I think the purpose of colour buttons is to set similar expectations. We
highlight the next logical step and indicate whether some new content will
be created (green) or destroyed (red) as an outcome. Creating and deleting
content, even if these actions can be undone seems worth some
considerations in an environment where content is public and edited
collaboratively by many.
The create/destroy/other split doesn't match the "dangerousness" of the
actions well, though. The typical way for a non-admin user to make a mess
is via page move (which usually cannot be reverted without admin rights)
and merge-type actions (wikidata item merge, or adding an interwiki link on
Wikipedia that causes different concepts to be merged). The most dangerous
actions with admin rights are probably user blocking and page history
merge. None of those are constructive or destructive in the sense the UI
uses those terms.
_______________________________________________
Design mailing list
https://lists.wikimedia.org/mailman/listinfo/design
May Tee-Galloway
2015-12-10 20:46:05 UTC
Permalink
I've responded to the related Phabricator
<https://phabricator.wikimedia.org/T110555#1864571> ticket with my thoughts
and summarization from this thread.

Feel free to keep responding here if you're not too much of a Phab person,
I've already linked this thread there. Here's a copy of the response:


**About constructive buttons: **
used to indicate an action that creates something rather than modifying
provide hint that something will be created, regardless of steps
several people have mentioned that green constructive that doesn’t “create”
anything


**Constructive vs. progressive buttons**

@isarra mentioned: "If it's shown to help, that's all we need. We have a
justification for the maintenance overhead and stuff." Unfortunately we do
not have proper resources to help us figure this out. But the current
situation is that we've had confusions from developers and community
members in understanding when and where to use constructive vs.
progressive. For our users to learn the association of constructive button
being the last action in a process and progressive being the in-betweens,
it "requires [the pattern] to be applied consistently to work" as
@pginer-wmf mentioned. Also, in certain complex processes, it's hard to
tell which would be the last step. An example given by @tgr was:

"
after you submit the login form and MediaWiki verifies your credentials,
depending on your user settings you might or might not be presented with a
two-factor challenge; so submitting the user name and password might or
might not be the last step of the form."

The idea of constructive vs. progressive sounds like an aid for going
through a multi-step process. Its underlying purpose is to keep users
informed of where they are within a process. My thoughts are that this
progressive/constructive solution sounds like it stemmed from the user need
of "knowing where I'm at within a process/workflow."

We can solve this more effectively. @volker_e linked us to a discussion on
SO where a user mentioned: "Apple hardly ever uses Wizards for set-up
processes, and when they do, the Apple set-up assistants are (in my
opinion) always easier to figure out than the Microsoft equivalents. I'd
suggest looking at the differences and try to identify the tricks employed
by Apple." I can agree with this—a better form /workflow experience.

There is also a good amount of rationale as to why we wouldn't need this
sort of variation, which mainly stated complications in applications by the
community members and devs and that we have no resources in place to find
out if the variation complication outweighs the advantages.

@spage suggested that we use "Green for Thanks, because in an unfriendly
often toxic "community" it's a nice Constructive thing to do."


---
-------------


My suggestion after these observations are:


**BLUE**

The main reasons why we use blue button is to (in order of importance):

(1) Help users with clear step to move forward within a page / workflow (if
there is only one way to move on)
(2) To suggest and highlight (if there are multiple options to move on). If
no specific suggestion, they remain uncolored.


Example of (1):
Primary actions

Example of (2):
Log-in form
Others



**GREEN**

There are also a few reasons why we would use green for links or icons

(1) To signal that an action has been taken only when it helps users
navigate a page / workflow
(2) To highlight a suggested action although not the main action
(secondary) on the page or workflow


Examples of (1) & (2):
Secondary actions



---
-------------


Because more colors means more confusion than help, we should have
restrictions such as

Only one color type of button(s) per page / workflow. Either a blue or red.
Never a colored link / icon next to a colored button. Remember, colored
buttons are to help users move forward, more colored links / icons dilute
the intention.
Use green links and icons minimally. Highlight most important only and only
when necessary to help users.


Overall, I'm still skeptical that we can make this a binary decision. I
think there's only not-so-good and good decisions. Good judgement is the
most effective. That is why guides are there to help someone make the best
informed decisions while giving space for creativity.

The most immediate example that comes to mind is the log in form. Both are
possible ways of moving forward and both are equally primary actions. But
say, as a team, we want to drive more sign ups, so instead of highlighting
both, we highlight only one. It's possible that both are highlighted, but
not recommended because when everything is in focus, nothing is in focus,
but this is a less extreme example of that).
In the other example of blue, one can use blue buttons repeatedly because
there is much content around a call-to-action button that warrants a need
for focus.

In conclusion, let's retire the idea of constructive and progressive.
Instead, we have primary (blue), secondary (green), and destructive (red).

Help me poke holes in my thoughts with more use cases or perhaps strengthen
them with more examples and better guides.

mm
We highlight the next logical step and indicate whether some new content
will be created (green) or destroyed (red) as an outcome. Creating and
deleting content, even if these actions can be undone seems worth some
considerations in an environment where content is public and edited
collaboratively by many.
Though this is correct, at times, it is very difficult to put an action
into the bucket of "creating" and "progressing" there are grey areas here
since we are talking about abstracted concepts and it's difficult to make
it binary. it is possible, but difficult.
That brings me to the users of these concepts. If i am not wrong, there
are three parties involved.
1. Designers
2. Developers
3. Users
For designers, things like progressive and constructive makes sense but a
on semantic level. for independent developers without any design help, the
distinction between constructive and progressive might make things
confusing and complicated. and as far as a I know, users don't perceive
their tasks to be progressive or constructive right way.
Of course, the semantics that were thought before implementing it our
buttons this way made sense at the time and i think it still can be
justified but it seems like there is a fine line between two which makes it
difficult to convey.
If the value coming from having these distinctions is less than the
confusion that we are causing then maybe it's time not have these different
conventions.
Post by Pau Giner
I think the purpose of colour buttons is to set similar expectations. We
highlight the next logical step and indicate whether some new content will
be created (green) or destroyed (red) as an outcome. Creating and deleting
content, even if these actions can be undone seems worth some
considerations in an environment where content is public and edited
collaboratively by many.
The create/destroy/other split doesn't match the "dangerousness" of the
actions well, though. The typical way for a non-admin user to make a mess
is via page move (which usually cannot be reverted without admin rights)
and merge-type actions (wikidata item merge, or adding an interwiki link on
Wikipedia that causes different concepts to be merged). The most dangerous
actions with admin rights are probably user blocking and page history
merge. None of those are constructive or destructive in the sense the UI
uses those terms.
_______________________________________________
Design mailing list
https://lists.wikimedia.org/mailman/listinfo/design
_______________________________________________
Design mailing list
https://lists.wikimedia.org/mailman/listinfo/design
--
mm
S Page
2015-10-23 19:50:00 UTC
Permalink
... That's at least three uses that go away if we consolidate Constructive
and Progressive.
Here's a fourth, the wiki login form. You can [Log in], but you can also
create an account (usually phrased as [Join <wikiname>]). They're both
significant actions, so neutral for the second button doesn't quite work.
So they take different colors, and over time a consistent color choice here
and elsewhere should suggest the interaction distinction in the guideline,
even though only a tiny fraction of users will ever be able to articulate
it. The onus is on anyone who wants to eliminate it to prove it's harmful.

I just noticed the [Join Wikimedia button] is taller than [Log in], a
regression. I filed https://phabricator.wikimedia.org/T116427

(I was present when Munaf Assaf and Steven Walling et al. came up with the
new login, the first case for Agora, now MediaWiki UI. Good times.)
--
=S Page WMF Tech writer
Volker Eckl
2015-10-21 19:57:03 UTC
Permalink
I want to try to step out of my role for a moment and think about it from a
user's perspective:
To me the button separation seems rather arbitrary in most current
use-cases, that I've seen so far. If it's a multistep dialog/action
connected, will the user understand that after using the green button
several times and therefore finish his task more easily? If a multi-step
action is needed, there's common agreement
<http://ux.stackexchange.com/questions/9/what-can-be-done-to-make-a-long-multi-step-wizard-more-user-friendly>,
that

- possibly saving user's inputs at every step to be able to go back and
forth,
- give feedback of *where they are* in the process and
- how many steps are left to accomplish the task.

Providing different button colors between the steps and providing one at
the final step doesn't provide much feedback and it's hard to grasp without
explaining it upfront and in-detail.
As we're already providing primary buttons and secondary ("neutral")
buttons, I think we should consolidate the two different primary
(progressive & constructive) ones in order to give user's a clearer focus
on what's the primary action to get "bigger" tasks done.
Post by Gergo Tisza
​One problem I have run into recently is that for a complex form you are
not necessarily able to tell which step is the last. E.g. after you submit
the login form and MediaWiki verifies your credentials, depending on your
user settings you might or might not be presented with a two-factor
challenge; so submitting the user name and password might or might not be
the last step of the form. (Arguably login should not be constructive in
the first place, but it is now. In any case, similar problems could be
present with the user registration form, which does create something.)
Personally, I agree with Bartosz that having four button types (five or
more if we include silent buttons) just makes the interface confusing.
_______________________________________________
Design mailing list
https://lists.wikimedia.org/mailman/listinfo/design
Continue reading on narkive:
Loading...