Change "Submit Dataset" to "Save Dataset"?

[isteves]

Since a "drafts" option has not been implemented yet, I believe changing the submit button to "Save Dataset" can help to prevent PI/user headaches when using the new editor. This happened, for example, to @brunj7 's intern, who was trying out the Editor for the first time recently. Since she wasn't finished with her dataset, she did not submit. When she came back to it a few days later, she wondered where the data went, since she had assumed that the data would be saved automatically.

If we do implement this, we can revise our data team initial contact email to be something along these lines:

Thank you for creating a new data package on the Arctic Data Center. Please let us know when you have finished editing your package, so that we can begin the review process.

I've tried out multiple submits, and haven't run into any issues yet.

@jagoldstein

[laurenwalker]

I think the better fix for this kind of issue is to allow the user to log back in without losing their Editor progress. See #605

[jagoldstein]

@laurenwalker what type of messaging will there be to inform users that their progress will not be lost if they somehow navigate away from the page or are signed out?

Besides token expiration, users may have their submission interrupted by browser crashes, loss of internet connection, etc. Or they may wish to intentionally save a "draft".

Our concern is that in the interim before the 2.1.0 release (many months), we will have to answer many support questions from anxious submitters explaining how to recover their submission draft or that their progress was actually saved after all. Some choice messaging could help to minimize this potentially time-consuming problem, but we would need a better understanding of your proposed solution in order to develop that messaging. Maybe you could explain to us more about HOW your proposed solution (#605) would work in practice.

[laurenwalker]

The main messaging we do right now is to show an alert popup message to the user when they attempt to navigate away from the page that their changes will be lost, and it asks for confirmation before navigating away from the page.

Actually, when we first made the editor, the button said Save instead of Submit. But after UX testing, we got feedback that the word Save was actually causing people to believe that there was either a drafts feature OR that they should click Save after each section of the editor is complete (e.g. after filling out Overview, users thought that they should click Save instead of click on People to go to the next section). So changing the button text to Save seems to actually cause this confusion rather than avoid it. Primarily for this reason, I closed this issue.

This is the same functionality we have in the current registry, so this is not necessarily a new issue, and we are not stepping backwards in functionality. Do we currently receive a lot of emails about lost progress in the registry?

If we do end up getting a lot of emails about this, we can put up some kind of messaging, but I want to avoid filling the editor with so much help text that people stop reading it, so I suggest we wait until we get feedback from users.

[laurenwalker]

About this specifically:

Thank you for creating a new data package on the Arctic Data Center. Please let us know when you have finished editing your package, so that we can begin the review process.

I'm not sure it would be a good policy to require users to contact us (email?) when they are done with their dataset to begin the review process. From a user experience perspective, it seems very cumbersome and confusing to get a success message when I've submitted a dataset, and I can see the dataset online, but then get a email that I need to email when my submission is done...

The more automatic we can make the submission process, the better, for the both sides. I think we'll have to make judgement calls sometimes about whether a submission looks unfinished or not. And we may decide that its safest to always email after every submission anyway, whether it looks done or not.

We can discuss during this week's Arctic team meeting if it needs further discussion.

[jagoldstein]

Thanks so much for taking the time to explain this in detail @laurenwalker ! The convo about the confusion over "Save" is familiar. While we do get some emails about lost progress concerning the registry, it is a relatively infrequent problem. We could speculate all day, but I think we all agree that at this point we need to release the new editor and see how it goes!

I do agree with @laurenwalker that asking submitters to email us is an extra and unnecessary step. arcticbot already emails us automatically and asking the user to do so would likely add confusion and annoyance to their experience.

[isteves]

Thanks for the explanation/history, @laurenwalker. I agree we should give it a go and see how people react!

With regards to the email, I agree that we ultimately want to automate the process, but currently, engagement with the data team via email is a crucial part of the submission process. We can also hedge it with "For your convenience, if you do not respond within a week, we will begin the review process", but I don't know if one extra email is that big of a problem--we want to open a dialogue with them or else their submission will never be published.

[jagoldstein]

@isteves I do agree that it is pretty crucial for submitters to engage in the communication process with us, but I don't think the onus ought to be on them to initiate that conversation. The submission is not the simplest process and ought to suffice as the user's initial engagement with us. I think it is best that we reach out via email first, with a typical customer-service-type approach.

[isteves]

@jagoldstein I think we might have a slight misunderstanding here. Once they submit/save for the first time, I'm advocating for us emailing them to see if they are finished or not. I don't think the onus should be on the submitter to send that first email to support@arcticdata.io

but in any case, I think we'll be testing it out as is, so we can table this discussion until we get more feedback to warrant trying the "Save submission" approach again.