alphagov / govuk-design-system-backlog

GOV.UK Design System Community Backlog
31 stars 2 forks source link

Task list #72

Open govuk-design-system opened 6 years ago

govuk-design-system commented 6 years ago

Use this issue to discuss this pattern in the GOV.UK Design System.

bazzle commented 1 year ago

Regarding the application status "Referral incomplete" line, which changes to "Referral complete" line, in our case.

One of our user testing participants was confused by this.

Upon seeing Application complete followed by You have completed 5 of 5 sections, they thought "Application complete" means the form is submitted. And believed at this stage no further action was needed to submit the referral.

image
steve-oconnor commented 1 year ago

@SebHoward Great to hear it's testing well. Do you plan to include the new tag designs?

joelanman commented 1 year ago

@bazzle I would think that if there's a task the user needs to do, that should be part of the task list? If all tasks are complete then the overall task should be complete. So if there's an outstanding task (Submit?) then that should be added, and it would then read 5 of 6 completed.

SebHoward commented 1 year ago

@SebHoward Great to hear it's testing well. Do you plan to include the new tag designs?

Hi @steve-oconnor I'm not aware of the new tag designs? I believe we are using the latest version published on the Design System. Could you drop a link to the tags you mean? :)

querkmachine commented 1 year ago

@SebHoward New tag designs are being introduced in v5 of Frontend. We're planning to provide a pre-release for testing soon™, but in the meantime you can see the new tag designs in our review app: https://govuk-frontend-review.herokuapp.com/components/tag

bazzle commented 1 year ago

Thanks for clarifying, we've been using the pattern incorrectly, in situating the Submit button at the bottom of the task list. We'll think about making Submit its own task, or including it at the bottom of check your answers.

mattia-gobbo commented 1 year ago

For 'Send an SR1 medical evidence form' we've used the Task list pattern and extensively tested its use with users. Overall, user were able to understand what task were involved and check progress, especially because all the section headings were visible above the fold with no need to scroll. From research we have identified the following points that could improve users experience and interaction with its elements:

image (8) Considering the user will interact with the page to understand the tasks that are involved in completing the form, check progress on the form and would never interact with the page when all the sections are completed, the H2 indicating the application is incomplete (in our case framed as 'SR1 form incomplete') doesn't provide much value to the user and might be irrelevant as it won't change status depending on the progress of the user.

image (10) Considering the user must complete the questions of a section in order and mandatory questions cannot be skipped (otherwise these would be flagged with an error), on the questions of the form there should be a exit route to allow the user to not answer the specific question at that time and come back later. For that reason, in our service, we've introduced a 'Return to SR1 form sections' link within a button group as an exit route. When clicked, the link would take the user to the page with the task list and wouldn't save whatever the user inputed in that question. We extensively tested the wording of the link, what worked best was using the form 'Return to + [Heading of the page with the task list]' as a suggestion of where the user would be taken.

image (13) image (9) A link to the 'Check your answer' page is enabled only when previous section of the form are completed to prevent the user to access to that page when the form is not fully filled in. Although, this raises accessibility issues as 'Check your answers' is not picked up by screen readers as link but as simple text and could confuse screen reader users when it's then read as link. It's not clear from the guidance what the status of the tag should be when the user visits the page that was previously disabled and than clicked 'back' to return to the task list page. Shouldn't it be 'In progress’ as the user has started to check their answer since that's the task the user needs to carry out in that page? There is space for improvement in the interaction between the task list and the CYA patterns.

0-1694622984519 image (11) For optional questions, rather than asking a radio question at the end of the task to check if the user has completed the section, we've asked a conditional question upfront to check if the user has that piece of information (e.g. Do you have the patient NINo?). In this way the user is then branched to a question in which can input that piece of information or the next question depending on their selection. The question would be considered 'completed' at the point of the conditional selection.

image (12) We've tested a version of the task list in which all the questions were listed, so the user could see specifically what questions were completed and could start to fill in the form from any point they preferred. Although, this wasn't feasible to be developed using the CASA framework as the code library of this last one is designed so the user can start from the first question of a whole section and complete the following one in order.

CharlotteDowns commented 1 year ago

We've just added some guidance on how to 'share findings about your users' with us 📝. This will help us learn more about how your users use a task list within your service.

Sineaddumigan commented 9 months ago

We are working on the Next Flood Warning Service for England. The task list has been a useful pattern for our service as organisations need to complete multiple steps to set up flood warnings if they want to set up multiple locations and recipients of flood notifications. Additionally, some of the tasks don't have a clear end point - so we tested the designs with the 'Have you completed this section?' question at the end of each task and returned the user back to the task list afterwards.

Initially, we had some issues with this page as some participants were confused about what would happen if they clicked on 'No, I’ll come back to it later' and concerned that if they selected 'No...', the actions they'd completed in that task so far may not be saved.

However, with a change to content - by including a line stating that their changes will not be lost, including the name of the current task when asking if they have completed the action and adding a label of the section status (section completed or in progress) we have seen a drop in any concerns around this.

We have tested a number of versions of the task list across multiple rounds - including a version incorporating a helicopter view of number of users/locations already added to the service, as well as a more simplistic version where sections state if they have been viewed, completed or not yet started. Participants tended to prefer the latter version due to simplicity and assurance that they were making progress on their set up. Below are images of our design so far that is working better for our users. Feel free to get in touch with Andrew Bignell or Sinead Dumigan to discuss in more detail.

Screenshot 2024-01-24 at 15 04 21 Screenshot 2024-01-24 at 15 03 45
richardhattonUX commented 8 months ago

This is duplicate question posted on the slack channel.

With the new changes to the task list pattern, I have a couple of questions:

1 The way to check your answers and submit the application is now a button (it used to be a task). If the user clicks the submit button with incomplete tasks still remaining, should we visually indicate which task needs completing (different to what we have now) or just the error summary box at the top of the page? I think there should be some guidance about how this would work. The button should always be present, as its usually not recommend to disable buttons or to hide buttons.

2 The task/sections completed indicator has been dropped. Does this impact screen reader users? I was under the impression the tasks remaining count allowed screen reader users a way to understand how many task remain overall. Something which many can gauge visually from the tags.

https://github.com/alphagov/govuk-design-system-backlog/issues/72#issuecomment-432652925 (edited)

BoabaFett commented 5 months ago

I'm working on income tax submissions for HMRC, a large, complicated online service with many journeys for taxpayers to complete. The Have you finished this section? feature has generated a lot of debate in our design team but this is my (very opinionated) take on it. We offer our current version of the question on a single page after Check your answers -

Have you finished this section? You'll still be able to go back and review the information that you've given us. Yes / no

image

We've tried different wordings and tried offering this on its own and as part of the 'Check your answers' screen itself.

At this point we've run dozens of research sessions which involve this patterns, and these are these are the responses we get:

  1. most users don't comment on the pattern at all - they click Yes and continue without comment (they don't take advantage of this feature)
  2. a minority of users say something positive like - that's nice, I can come back later
  3. another minority of users find this extra question annoying and frustrating - 'Why are you asking me if I've finished?'
  4. no users at all understand what effect answering yes or no has on their online application

There's one exception to the last group - if the user gets frustrated about why they're being asked the question over and over again, they answer 'No' and see that it controls the status on the task list. Incidentally, I had to do this myself to work out why I was being asked the question. When pressed on what they think the question is for, users give the wrong answer: they assume it 'sends data to HMRC' or it's a compliance function, to confirm the data you've just entered from a legal point of view. A related problem (which other people have mentioned) is that users don't understand if 'section' means the journey they've just completed, or the section their journey sits in.

As we're not in the business of offering patterns which users don't understand, I'd strongly recommend removing this pattern from the design system for scenarios where the user has reached a 'Check your answers' screen.

More detail ... The original purpose of this feature was to help users who leave a journey without reaching the 'Check your answers' screen. I believe it's a mistake to offer this on or after 'Check your answers', when the user has already provided a complete set of answers for that journey.

The power of the task list is that it breaks up longer journeys into manageable ones and offers the 'reward' of changing the status from Not started to Completed. Asking people to confirm they've finished when they've given a complete set of answers creates a blurry mental model. I'm skeptical about whether changing the content or position can help with this fundamental problem. Adding this pattern to the design system lends too much authority to a pattern which doesn't make logical sense.

Our live service doesn't go live for a couple of years, but I'd predict that in practice, this feature will only ever be used by a small minority of users and confuse and frustrate the rest. Web analytics could confirm this.

Cheers all :)

steve-oconnor commented 5 months ago

That's some great insight @BoabaFett

I can see how a "have you completed this?" is superfluous for a task list journey that starts with the questions and ends in a check your answers page.

For clarity, can you give an idea of the full journey? Not every single step if it's complex, but how they get to entering data, when they enter data, when they get to "check answers" and the question, and where they go after that.

What we found in the research is that there are different journeys that involve the task list and they can be quite different.

BoabaFett commented 5 months ago

Hi, a typical journey in our service would be:

Taxpayers go through a 'tailoring' journey to pick which journeys are relevant for their income tax return. This populates a task list where they can see journeys relevant to them. Any example of an individual journey would be:

Income from foo

  1. Did you get income from foo? - Yes / no
  2. How much did you get? - Text input
  3. When did you start getting income from foo? - Date
  4. Did you pay tax on your income from foo? - Yes / no
  5. How much tax did you pay? - Yes / no
  6. Check your answers
  7. Calculation page: how much tax you may owe
  8. Have you completed this section? - Yes / no
  9. Return to the task list

... When the whole task list is complete taxpayers move on to Your final tax overview so see their tax calculations for a year; then they submit the application.

steve-oconnor commented 5 months ago

@BoabaFett I used a "This section is compelte" checkbox in the sub pages of a task list where the interactive elements where on a single sub page linked to from the task list. It wasn't a series of questions and a check answers page. It was also due to the task list being accessed by 2-3 people at once and only one of those people able to say "this section is correct and complete".

It sounds to me like your journey that doesn't need to use the "have you completed" question at all.

frankieroberto commented 5 months ago

@BoabaFett good research, thanks for sharing!

I first used this pattern on the 'Apply for teacher training' service at the Department for Education. It worked pretty well there, but it might not in every service.

For context, the reason we used it in the service was because we had several sections where it would otherwise be possible to tell if the section was truly completed or not. For instance:

We had other sections where it was more straightforward to detect if it was likely 'completed' (eg contact details), but decided it made sense to use the 'Have you completed this section?' pattern on every section for consistency.

Hope that makes sense! If you can tell whether each section is completed without the pattern, and have no evidence that users might want to come back and change their answers later, then you could just set the status automatically instead?

shabana-ali commented 4 months ago

When the font is resized to 200% (screenshot taken on Chrome and font size set to 32 via chrome://settings/fonts), the longer status copy spans across 2 lines. Was there a reason to set max-width: 160px; for .govuk-tag?

Also, on the old task list, the statuses would left align on small screens (similar to summary lists) but remain right aligned now. More of an issue when the task names are a little more wordier and then end up spanning across multiple lines. Again, was there a reason for this?

Complete multiple tasks example with font resized to 200% and incomplete text spanning 2 lines

frankieroberto commented 4 months ago

@shabana-ali good questions!

When the font is resized to 200% (screenshot taken on Chrome and font size set to 32 via chrome://settings/fonts), the longer status copy spans across 2 lines. Was there a reason to set max-width: 160px; for .govuk-tag?

This was required, I think, to force the tag to break onto multiple lines mid-word rather than overflowing outside the boundaries, which would result in horizontal scrolling.

From the tag/_index.scss code:

    // set a max-width along with overflow-wrap: break-word below for instances
    // where a tag has a single long word and could overflow its boundaries.
    // The max-width is necessary as break-word requires a bounding box to base
    // where to break off of.
    max-width: $govuk-tag-max-width;

You could try tweaking the max-width to be a bit wider and see if that improves things for your service?

Also, on the old task list, the statuses would left align on small screens (similar to summary lists) but remain right aligned now. More of an issue when the task names are a little more wordier and then end up spanning across multiple lines. Again, was there a reason for this?

This was a deliberate trade-off as in the old task list, the status would drop underneath the task name on mobile even though there was often space for them to remain left and right aligned (which is easier to scan).

With mobile screens increasingly wide, most of the time it should make sense for the left and right alignment to remain – but it does depend on the length of the task names and status labels, and the text size/zoom!

jeana-adhoc commented 2 months ago

Have you tested this with screen magnifying users? Did they have any difficulties identifying the tags/labels floating to the right?