Login vs Log In
stackoverflow.com
stackoverflow.com
sign in/sign out is clearest to the average (nontechnical, mass consumer) user...
...but the problem is, so is "sign up." so when in conflict with a third "sign _" action, the preferred and easily recognized choices are "log in," "log out," and "sign up."
this is a generalization, but "log on" really only resonates with people who had modems.
there you go.
If they were unrelated, it would be a problem to have similar naming, but in fact they are all related actions.
Sign up is required to join the website, sign in is required to access the website later, and sign out is the opposite of sign in.
That said these questions are always difficult because differing cultures/usages/dialects, even within one language, can dramatically affect comprehension. So a site really has to be tailored to the particular audience it seeks - a site in the US, the UK or India might use very different terms depending on the particular content it covers. But if you're targeting native speakers of English in most idioms I'm familiar with I think the terms above are pretty clear, and made clearer by using the same verb; other choices make them sound like unrelated actions, which they are not.
These composed verbs, whose meaning depend on the modifiers, are definitely confusing.
You can go a long way without mastering them, and their meaning (assuming there's only one) isn't always easy to infer from the roots. Think of getting off, for example.
Some of the combinations are still not part of my passive vocabulary[1], and most are not part of my active vocabulary, because I can usually get away with synonyms.
For this scenario, "log in" and "register" offer the best syntactic contrast, and should be favored, IMO. "Sign in" and "sign up" require more thinking. Pet peeve of mine, actually.
[1] There was a vocabulary depth study posted here on HN a while ago, I was at the 95th percentile of non-native speakers, equivalent to the median of natives of my age.
... in English-speaking countries.
> [M]ost people would only have come across log in within a computer related context.
Most of the time, non-native speakers come across English in computer-related contexts, i.e. on the Web.
This group of customers may be larger than the group of native speakers depending on your target audience. So optimizing for non-native English speakers living in non-English speaking countries is something which should be considered.
Getting particles (eg. "in" in "sign in") and prepositions right is one of the hardest parts of learning a new language. (Word gender for languages with that feature is another hard part.) You can understand almost everything of a website just from understanding verbs and nouns.
Use: "Login" for login. "Logout" for logout. "Sign up" or "Register" for registration.
You don't want any ambiguity in your buttons. "Login" is internationally understood, "Sign in" not so much (and takes more effort to visually distinguish from "Sign up").
Users want to provide their identity credentials to access your site and verify who they are. They do that via an electronic signature which is their username and password.
Therefore they want to "Sign In" with their signature.
The fact that you're 'logging' their signature is secondary to their needs. And as far as making a compound word out of "log" and "in"... well that's stupid (and wow... "Logout" is a new verb created just so you can leave a website securely? Try explaining what that is to someone who has never used the internet... 'no it's not street slang for going to the toilet').
As for your 'more effort to visually distinguish' between "Sign In" and "Sign up", you already solved that problem. "Sign in" and "Register" are the winners.
And we don't have to involve any lumps of wood (or new verbs that suggest bowel movements).
Once you've registered your signature with the site, the next time you arrive, you merely provide your signature. Register and then Sign In.
seconded
I also agree that "register" is a great alternate to "sign up."
My original intent in responding to this thread is to help educate HN readers between data and opinion when it comes to "good UX." As a designer I've had opinions about these things for years, as we all have. But not all of us have had the opportunity to test our ui lexicon in front of millions of users. So while we might think "oh, register makes more sense," I am simply offering that no, in fact, it did not. "Sign up" was the most commonly-expected term for participating in the system as a non-user. "Register" was viewed as more committal, technical, involved.
active: tested lexicon on focus groups, asked what users expected actions to do, testing hypothetical affordances, etc.
However, none of this is relevant to the "Login" vs "Sign in" debate. If that button influences anything then it would be in the retention, and I doubt the impact can be measured in insulation either way. It's merely one of the tiny details that make a difference in aggregate.
And with Old Media, who routinely use it to mean "visit a Web site." As in, "Log-on to msnbc.com to get all the details"
This clearly isn't a programming question, it's a UX or language question. Appropriately enough, a very similar question was previously posted on UX.StackExchange.com: http://ux.stackexchange.com/questions/1080/using-sign-in-vs-...
When the site was younger there were lots of questions that were only relevant to programmers, not actually about programming. In small numbers these weren't bad, but they became increasingly popular compared with actual programming questions. That was not a healthy trend -- communities needs to keep their focus -- and the stricter moderation that Stack Overflow has now was partially a response so that.
If somebody wants to contribute an answer, they should post it on the on-topic UX.SE question. Closing the Stack Overflow post shouldn't affect anyone. (Deleting is another matter, and I'll oppose deleting this question if it's suggested.)
Please downvote me if I missed the meat of what this was really supposed to be about. I think maybe I just don't understand the HN community most of the time.
Also "sign" vs "log" and "in" vs "on" are A/B testable.
No, no at all.
From a back end, sure, log your access in or on and out of off. But from a user perspective, why are we logging anything. Why do we need to sign a metaphorical log sheet every time we come and go?
Is there no better way to do it? Are we destined to spend our lives and future signing the log sheet at the metaphorical security desk for the rest of human existence?
To play devil's advocate: to me, "logging out" and "exiting" are not the same thing. Particularly, the first is just a change of state--I'm still in your app, but now I'm no longer signed in. On the other hand, "exiting" is more like a motion; I expect to actually leave your app and not just be logged out. In UI terms, I see "exit" as a synonym of "quit".
Now, perhaps this makes sense for your app. However, if it makes sense to interact with it without being logged in, I think that is the wrong terminology. So I would definitely not use it for something like HN because browsing stories works regardless of whether I'm logged in or not.
I'm sure you thought about this far more when you were making your decision, so this is just my impression. You can always try to test the alternative empirically once you have a bunch of active users :).
We are looking into a few other alternatives including the most obvious where we convey that they are indeed leaving the app but in the next screen let the user know that their google account is still signed in and provide an option if they still want to sign out. Thankfully we have quite a while to launch, so I have time to run a couple other iterations and pick the one which seems most clear.
From what I understand it entered into English as a stage direction, but by the 16th century it had become a noun referring to a doorway for leaving.
People agree on the meaning of symbols/conventions, and then dictionaries draw targets around the arrows.
But now, forget Facebook, think SMS and IM. In 50 years, everyone under 75 will be familiar with its vocabulary, and will have used more for personal communication than the "correct" language. I predict that, school be damned, it will replace the current spelling in everyday life for most people.
Note that I love beautiful language and its historical roots. At the gut level, I find it painful to watch it being torn down. But I'm too young to be a curmudgeon...
I already do not understand SMS spelling: in 50 years, probably a 30 years old person will not understand things written by a 15 years old person.
Conservatism in schools has the advantage that we can still read Shakespeare (or Dante) without translation. Languages that change too quickly defy their own purpose (i.e. being a common ground for people to communicate).
Conservatism isn't inherently bad, and innovation for its own sake is idiotic.
Note that, for Shakespeare at least, even though one can still understand the main meaning of the text, a lot of puns and cultural references are lost to the modern reader (lest he's a specialized literature scholar)...
Therefore, I simply don't think about it and typically just use "login". It's recognizable, it simple, and users know what to do with it. Everything else is simply keeping you away from the more important parts of your project. Just one man's opinion.
- website vs web site - setup vs set up
That is just two examples that come to mind in as many seconds.
The point is that the actions do not cease to exist simply because they are abstractions over a lower level of operation. Users don't need to know what sessions are and it detracts from the user experience, so its not a good idea to expose them to the details of that level of abstraction.
Finally, I don't see any clear distinction between session/create and session/new. I'm confused about which one corresponds to creating an account vs logging in. I'd say confusing terminology like that is far worse for user experience than any of the log/sign options.
How about "authenticate"?