|The Clear Button and Forms|
is a clear button necessary anymore
When you first start learning html and you look at some form examples, you usually see a Clear button right next to the Submit button at the bottom of a form. I always assumed it was there because you were filling out the form while impaired, and when you checked to see that all of the information was correct, you realized that almost all of the fields were wrong so you might as well just start all over with the Clear button. I recently asked a colleague of mine, and he replied that if you filled out the form and then decided not to submit it, a person who didn't know any better might believe that leaving their info in the form fields might be magically putting it out into the internet somewhere. So clearing those fields will make them feel better. Either way, it seems very silly to me. The only purpose I see for it is to clear radio buttons, since there is no other way to unselect a group of radio buttons.
So do any of you use clear buttons in your forms anymore? Do you believe anybody uses clear buttons when they are there? Since most forms out there don't use them, do novice users even know what they are?
You should get rid of any "Clear", "Reset" or "Cancel" buttons from forms, because they can actually be harmful rather than helpful, and at best they are a waste of space. Even Jakob Nielsen is against them:
You tend to only see reset buttons on amateur sites these days. If you really need one, then it should at least be styled differently from the submit button.
Thanks for that link, encyclo. Thought I had read Nielsen's whole site by now but here's a 5-year old page I missed. I particularly liked:
|Exception: Use Reset for Repeated Form-Filling |
Reset can be useful for forms that satisfy both of the following criteria:
* the form is filled-in repeatedly by the same user
* the data to be entered differs significantly from one use of the form to the next
Even if a user were to use a form frequently, the reset button would still not be necessary
if the data was mostly the same from one use to the next. In these cases, it would be
easier for the user to edit the old data than to erase it and start from scratch.
That's exactly what I wanted encyclo. I knew they were good for nothin.
|You should get rid of any "Clear", "Reset" or "Cancel" buttons |
|I knew they were good for nothin. |
Since even Nielsen allows for some exceptions (several of which tedster has listed), that's a bit of an overstatement. Better to say that they are of very limited usefulness.
The only one I really have found much use for is the "Cancel" button, which Nielsen also discusses in the page linked above, shortly after the section tedster cited. He recommends using it "sparingly," viz. in a situation where someone using "Back" doesn't work or might be more confusing. (See also the added note about canceling/stopping a download, etc.)
Alright, my colleague is actually my boss, and he strongly believes that people are very paranoid about filling out forms. He wants all forms to use https and he wants them all to have a clear button. He believes that if a user fills out the form and decides not to submit it, they may not want their data in the fields anymore, so the clear button satisfies their paranoia. He doesn't agree with Nielson, saying Nielson "seems to equate usability with herding cattle, i.e. give them only one route down the path of the developer's choosing." And also that Nielson and us developers sometimes lost track of how little people know of how the internet works.
I've argued this to death with him, but can't pursuade him. Anybody have some good arguments I could use? Anybody agree with him?
Sounds like you've argued and lost - and he's the boss. ;) I would acquiesce, but do something like this to carefully distinguish the submit and reset buttons:
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
<input class="sub" type="submit" value="Send the form"> <input class="res" type="reset" value="Clear the form">
Important to note, also: it is not a clear button but a reset button. The distinction is that it resets the form to it's original state and does not "clear" anything.
Example: a member login that loads their data in an edit profile form. They change a few fields and hit reset, and it loads the original data - it does not clear them. The only reason it **appears** to do so on most forms is that the form initially loads blank. Which is it's original state. :-)
|He doesn't agree with Nielson, saying Nielson "seems to equate usability with herding cattle, i.e. give them only one route down the path of the developer's choosing." And also that Nielson and us developers sometimes lost track of how little people know of how the internet works. |
As I see it, his argument is "users know so little that we have to give them more choices".
My natural response is, "since people know so little of how the Internet works, it serves them best to offer the smallest number of choices consistent with accomplishing their tasks". Choices are appropriate for knowledgeable users.
However, it sounds like your boss has made up his mind and isn't really interested. That he thinks of Nielsen as a bureaucratic authority he's going to resist. It probably won't do much good to point out that Nielsen is an authority who actually thinks, researches, reconsiders, and is respected not because he's powerful but because he's astute.
So ... good luck. I can't recommend trying to change his mind. At least in this case his recommendation is a common, familiar element, not something off the wall.
Also, Nielsen very strongly says that users prefer to use Back, and his statements are based on actually observing users. Anyone who claims that users prefer a method other than hitting Back should be prepared with data to back up the claim.
|chadmg: He believes that if a user fills out the form and decides not to submit it, they may not want their data in the fields anymore. |
I agree. What about a kiosk situation where one person might enter information, submit or clear as s/he wishes, and then walk away? If I'm the next person up there, I may be able to get at that information. With either a submit or clear, you can execute code which will leave nothing to find, but using the "back" button leaves your information easily retrievable.
|Nielsen very strongly says that users prefer to use Back, and his statements are based on actually observing users |
This is, of course, foundational to his argument. Unfortunately, on the page discussing these buttons, Nielsen does not lay out exactly how this has been observed nor link to an explanation (testing results, etc). Instead he seems to jump on that generalization to justify his own belief.
Note, I am not saying that Nielsen's conclusion is incorrect! But by failing to include the specifics of these observations at this point, he makes it difficult for us to assess how well they support the specific case he is making.
(It's rather like a politician citing poll numbers to "prove" a point, without giving you the exact wording of the questions, or the "internals" that help you to interpret what people mean by the answers cited.)
So, does anyone know where on his site he does lay out the details of this study (on use of the Back button, etc)?
I agree that a problem with using Nielsen sometimes is that to get access to all his results you have to pay $$$$$. Not a criticism; he's doing it to make a living. He's presented enough explicit results that I expect he has data to support this, but I don't find it in his public papers. Some of his mentions of the back button, such as "hitting the back button with a vengeance", imply anecdotal rather than statistical evidence. Still, he's spent enough time observing users and analyzing their behavior that his statements about a behavior as common as using the back button carry a lot of weight even without stats.
That said, I found two columns with links to external papers about navigation:
The second paper has a lot of references to other work. And there's probably a lot more from recent years.
Those papers appear (I haven't done more than a brief skim of either) to be studying navigation behavior in general and probably say nothing about backing out of forms specifically. Still, the figures they cite for frequency of back button use strongly implies that almost all users hit Back, and hit it a lot.
I use a clear button on my website for my feedback which has a radio buttons and text box's - I always thought it might be userfull. However, after reading this argument it has made me think twice. =\
|Also, Nielsen very strongly says that users prefer to use Back. |
This I believe would be true, or at least I agree - I feel that it is more convenient for the majority of people to click a back link after reading a page as they tend to be at the bottom of pages, and if the user reads the whole page, then they would normally see that button. However, some people like to use backspace, as it is convenient, and quick... I think it depends on the user you're dealing with.
But when we say "the Back button", we (and Nielsen) are talking about the browser's "back" function/button, usually both something clickable and also backspace and alt/opt/cmd/whatever-left-arrow. The user's paranoia would be fully justified if she had to click anything at all on the web page, as even things that look like plain links sometimes submit forms. What Nielsen is claiming (and probably has data for but hasn't cited it in non-paying venues) is that people prefer to use the browser's "back" function, NOT a link on the web page.
I understand - why would people be paranoid? Because of re-posting data or missing something? Or they just are?
I also fear pressing back at all when it comes down to ecommerce and any sort of online shopping. For example, ebay. I worry I might re-add an item, or re-bid. So I use the links provided to go back. You would think that the majority of people would understand that the web designer/s wouldn't put a back button if it was unsafe. I usualy feel more safe with the ones provided anyway, for that same reason. I hardly ever use the browser provided back button.
|The second paper has a lot of references to other work. And there's probably a lot more from recent years. |
Thanks for the references, paleolith. A bit sorry to see that the most recent one is now six years old (though I can't offer anything more recent myself).
I suspect there has indeed been more work on this, and that Nielsen's observations are generally correct. But it still would be nice to see more detail, more recent studies, and particularly some indication of any exceptions to help us judge just when a Back button is NOT the best choice.