search

ciscoheat / sveltekit-superforms

Making SvelteKit forms a pleasure to use!
https://superforms.rocks
MIT License
2.26k stars 65 forks source link

Better Punish Late, Reward Early validationMethod #469

Open eakl opened 3 months ago

eakl commented 3 months ago

I found out this closed issue and I really think it should be shipped out-of-the-box (punish on submit or blur, reward on change). Let me explain.

@ciscoheat, this article you reference as the validation method that brings the highest user satisfaction is from 8 years ago, and while the idea is IMO still the best one (reward early, punish late), its implementation is not the best.

  1. If the user is entering the data in the field that was in a valid state, perform the validation after the data entry
  2. If the user is entering the data in the field that was in an invalid state, perform the validation during the data entry

The problem with the validationMethod: 'auto' implementation is that it validates late but doesn't reward early enough.

  1. Validating after the data entry avoids distracting the user with annoying error messages while they're typing (because we assume they know what they're doing and that they're going to type it right).
  2. Once the error message has been shown, why should we keep that error message until the field is completely validated? Why don't we assume that the user has seen the typo/error and as soon as they start correcting the data they know what they're doing just as the previous point?
  3. In case the error is repeated or there's another error, we validate again.

IMO, error messages are annoying. I don't want to get them as soon as I start typing in an empty field because I want the validator to assume I'm smart enough to understand the requirement (punish late).

If I made a typo or misunderstood the requirement, I want to get a clear error message that points to my mistake.

But as soon as I understood my mistake and know how I'm going to correct it, I don't want to see that error message. The validator should assume that I know what I'm doing now and I'm going to type it right (reward early).

Currently, validationMethod: 'auto' is holding my hand until I reach validation in an erroneous field and it's annoying because in my mind the error has already been corrected.

This is used at Google and Apple: Google: https://workspace.google.com/business/signup/welcome Apple: https://appstoreconnect.apple.com/login

ciscoheat commented 2 weeks ago

Hello, I'm sorry for the late response. I see what you mean, but my main design principles are:

Make sure that: (1) the user can figure out what to do, and (2) the user can tell what is going on. -- Don Norman, "The Design of Everyday Things"

Rule 2 is why they are kept, to avoid problems after a "sequence break", for example if you step away from the computer and come back, you are put back into context.

Not all users are adept at computers, they need hand-holding until things really work (that is, all the errors are actually fixed). The mental model for live feedback cannot be cheated, then users stop trusting it.