-
Notifications
You must be signed in to change notification settings - Fork 27
List of UI Annoyances
This is a list of inconsistencies or annoyances/irritations in the UI. We are listing them with the intent of mounting an offensive against them soon to make Conductor much more user-friendly.
-
“Launch with errors” makes it sound like it will cause errors on purpose
- maybe rewording to something like: “Ignore errors during launch” would do
- Random orange alert shown when creating a deployment , no text
-
Create a pool: Should probably default to enabled, not disabled
-
“Are you sure you wish to deploy”Epigaea repens" to the Staging Pool? Doing so will utilize 0% of your overall deployment quota." -- Maybe we shouldn’t show “will utilize 0% of your… quota” when the quota is unlimited?
-
[FIXED b54f2b7 (seolus-configure)] `@font-face` with Interstate or whatever we use is missing in RPMs (wrong paths in httpd conf in aeolus-configure)
-
[IN PROGRESS]“Xml Deployable XML file doesn’t resolve valid XML” is the worst error ever
- it is worded and uses the word “XML” three times in one sentence
- it doesn’t actually indicate what’s wrong with the XML -- we might as well just say “Your XML is no good! Try again!”
-
The Pool Deployments toggle arrow on overview page stops working when backbone updates
-
Applied filters in some views are not persistent to backbone updates
-
flash instead of flash.now is used on same pages in combination with rendering instead of redirection -> flash message is displayed for next request too which is very confusing
-
[FIXED posted/in progress]“name availability checking” when launching a deployment: 1) “next” button on new deployment page is disabled by default (no description why). 2) after changing and leaving name input field name checking is done on bg - there is no loadmask or description what is being done, user just sees disabled “next” button which is so confusing. 3) when user goes to this page by pressing “back” button on launch time params page, “next” button is disabled again even if name field is filled in. It would be best to remove this enable/disable of “next” button, information about name would be just displayed next to name field.
-
The provider show page is not accessible from anywhere in the UI, consider moving the propertes to a tab in provider edit page and remove the show view, route and related controller code.
-
Remove New Provider link from the provider-select dropdown.
-
Provider select should be probably removed from the New Provider page. It’s a good idea to keep provider select in edit page, only the label should be more something like “Switch to provider:”
-
The order of the providers in the ‘Choose a provider’ dropdown is changing after a provider was enabled/disabled. The order should be persistent
-
There’s no page where the user can overview which providers are enabled and which are disabled. The new providers#index page could be good place for it. Now it can only be checked on the show page of the individual provider
-
In case of validation error the form on provider_accounts#new seems to properly persist the key and certificate field over the request. But after fixing the error and submitting the form again, it complains about the missing key and certificate
-
The form of deployables#new and images#new have different default value for editing the xml file
-
On Launch time params page ‘Cancel’ is capitalized ‘back’ is not
-
Inconsistent deployment pretty view cards on pools#index and deployment#show. Why do we need 2 different kinds of deployment card?
-
[FIXED 5ebb17e] Remove unnecessary “Logged in successfully” message
-
/pool_families uses both “Environments” and “Pool Families” terminology
- In fact, “Environment” and “Pool Family” mean exactly the same
thing here (aeolus-cli also uses the term ‘environment’ for
pool family. The reason for this is that we had just begun the
process of changing “Pool Family” to “Environment”. The idea was
to use Environment instead of Pool Family because 1) it’s what
Katello uses for a similar concept and 2) Pool Family is a bit
confusing -- Environment actually matches the function of Pool
Family fairly well. We hadn’t actually changed any of the
codebase yet, though -- we got ‘Environment’ in the UI as part
of the UI design work, and the CLI work was new, so we started
with ‘Environment’ there. The next step was to change it to
‘Environment’ everywhere, when we heard that PM didn’t want to
call it Environment but ‘Cloud’ instead. Without a term that
everyone agreed on, changing everything to ‘Environment’ didn’t
make a ton of sense at the time, but pretty much everyone on the
dev side agreed that a vague term like ‘Cloud’ would be a bad
idea for the upstream codebase, so we were left with the
original term (for now). Lets leave this as-is until we can have
some more targeted conversations about terminology used both
upstream and in the product -- hopefully we can get more
convergence. It’s possible that ‘Environment’ will be acceptable
for a future product version, but I’m reasonably sure that ‘Pool
Family’ will never be approved.
- Hard to tell that a pool is disabled; this is not shown anywhere on Pool list whether Pool is disabled should be distinguishable in Pools pretty list, Pools filter view and Pool show page
- Pool show page Properties page has no content
- we need to decide whether to remove it or if it should hold any information
- View a new environment / pool family: “Images cannot be imported, as no Provider Accounts are currently enabled for this Environment.” -- wtf? Why? How do I fix it?
- In fact, “Environment” and “Pool Family” mean exactly the same
thing here (aeolus-cli also uses the term ‘environment’ for
pool family. The reason for this is that we had just begun the
process of changing “Pool Family” to “Environment”. The idea was
to use Environment instead of Pool Family because 1) it’s what
Katello uses for a similar concept and 2) Pool Family is a bit
confusing -- Environment actually matches the function of Pool
Family fairly well. We hadn’t actually changed any of the
codebase yet, though -- we got ‘Environment’ in the UI as part
of the UI design work, and the CLI work was new, so we started
with ‘Environment’ there. The next step was to change it to
‘Environment’ everywhere, when we heard that PM didn’t want to
call it Environment but ‘Cloud’ instead. Without a term that
everyone agreed on, changing everything to ‘Environment’ didn’t
make a ton of sense at the time, but pretty much everyone on the
dev side agreed that a vague term like ‘Cloud’ would be a bad
idea for the upstream codebase, so we were left with the
original term (for now). Lets leave this as-is until we can have
some more targeted conversations about terminology used both
upstream and in the product -- hopefully we can get more
convergence. It’s possible that ‘Environment’ will be acceptable
for a future product version, but I’m reasonably sure that ‘Pool
Family’ will never be approved.
-
Getting to edit/show pool_family view is accessible only by clicking the pool family title which is rather user unfriendly
-
“No catalog exists! Please create one.” error pops up on new pool when you try to create a deployment. You can’t add one on that page.
-
Why are “Images” under “Environments”? Aren’t they Content?
-
The whole Monitor/Administer distinction is arbitrary.
- The current workflows cause in many cases skiping from administer to monitor sections bsed on the which section the current view belongs to which is very confusing.
- as Matt pointed out let’s make the administer tabs first level and add Overview tab at the beginning.
- Based on changed top tabs, let’s make a dropdown shortcuts for the folowing tabs if the page under the tab is tabbed (Users, Environments, Content, providers (make a new provider link) [2], this also brings some space to utilize changes for better workflows which Brian works on. Display tabs and subtabs based on permissions of user.
-
You have to jump between Monitor/Administer to do common things, like import and launch an image.
-
On the add a provider page -- I feel like provider dropdown should be first field on page, since it changes what fields you see *** the create provider action button should propably exist separately from providers dropdown ** Templates site -- we’re thinking that terminology (“Template”, “Deployable”) is not intuitive… Maybe fold that back into our app too?
-
Logs UI needs polishing and figure out proper Logs link placement (original designs puts the link (icon) between pretty/filter togle buttons [1]), resolve places where logs sections should be also present (eg. related only to pool, deployment, instance, user, provider…) see logs section design: [1]
-
Launch from Catalog (dropdown) is not clear what it does (should be something like Launch NEW DEPLOYMENT from CHOOSE CATALOG)
-
Settings tab is kind of empty (is needed?) the first page just points to another with link
-
Breadcrumbs - enable them everywhere, original point - maintain applied filters (viewstate), use it to help with the navigation (plays nice with dropdown tabs), we need to figure out proper placement and design
-
Notifications, the current placement of notifications is wrong and doesn’t fit the current Header style. UXD Wiki says nothing about notifications look and placement either [3]
- flash[:error]s are very easy to miss; should be more prominent (Katello does this much better)
-
Forms do not reflect whether an attribute is optional or required
-
Error reporting should be improved. In many cases we just return the name of the exception to the user which is a bad UX. We should not interact with the user via exceptions but process the exceptions and set a custom message based on it.
-
The Overpass font does not align vertically to it’s line height.
[2] http://file.bne.redhat.com/~afitzsim/headers/SE-header-CSS-HTML/katello-header.html