3 comments

  • Thomas Michael SemmlerThomas Michael Semmler, over 2 years ago

    This is a complex matter and it is not just solved be "just use an autocomplete text input".

    Many things can go wrong with this:

    • the list of options is not very small, you force the user to go from a "selecting / taping" action to a "keyboard input" action.
    • the person is forced into your phrasing or understanding of concepts. If your label doesn't describe the invisible list of contents accurately and appropriately enough, your users might search for something that they can never find
    • the person might have troubles typing but might be fine just taping a few options
    • selects input-types are treated differently by the OS. Whereas there is no native autocomplete that behaves consistently throughout OS' or that we as developers have enough control over.

    This post highlights more a different thing you should not do - assume users share the same mental model for something.

    If this form's users are not just Americans, but say, europeans, and there is a state field, then that might mean different things to different people. To people from India, it might mean "state", to people from Germany it might mean their country, because that is the word that is being used to refer to it - but in actuality, you wanted their province. That problem would have ben solved easily with a select, that prompts the user whatever is available and lets them know what kind of state you want from them.


    TLDR: There is a place for selects and a place for autocompleted text inputs, neither are bad or good, they are contextually appropriate and effective, or not. But one does not replace the other in general, not even in most cases. It's bad to think of a possible solution as in general better then the other one.

    2 points
    • Stuart McCoyStuart McCoy, over 2 years ago
      • the list of options is not very small, you force the user to go from a "selecting / taping" action to a "keyboard input" action.

      Actually, they go to a tap/swipe/focus to ensure right location in menu/swipe some more/repeat/tap action.

      • the person is forced into your phrasing or understanding of concepts. If your label doesn't describe the invisible list of contents accurately and appropriately enough, your users might search for something that they can never find

      This is why the argument was made that this is to be used with known input selections (i.e. states/countries, car manufacturers, etc.). If there is a hint of the unknown to the user then don't use this select option, use a proper, prefilled dropdown.

      • the person might have troubles typing but might be fine just taping a few options selects input

      If the dropdown is that small, then yes, that would be a better option. I think the inference from this article was that the list was long but known (i.e. states/countries, car manufacturers, etc.) and typing the first two or three character would trim the menu options down to a handful that are easier to then tap.

      • types are treated differently by the OS. Whereas there is no native autocomplete that behaves consistently throughout OS' or that we as developers have enough control over.

      This is a developer/designer problem that needs to be sorted out earlier in the process. If they know the data will support an autocomplete method then they need to design/develop the solution with this in mind. Sure, it might make some choices harder or impossible but that's not necessarily a bad thing.

      0 points
  • Philipp Christen, over 2 years ago

    Definitely agreeing with Thomas here. Adding to that, I've been cought off guard by text inputs that look like selects more often recently. Being from Switzerland, I know the country's place in the dropdown list of countries - I usually press T (for Taiwan) and the up twice to go to Syria and then Switzerland.

    0 points