Login Screen - My final design



About a week ago, I posted some login designs for haiku. Based on the feedback, the designs where not accepted well. So I decided to buckle down, and put my mind to work. I worked out details on paper, and then came up with this:

Tell me what you think of it, I think it stays pretty real to the classic “haiku design”


Simplicity is elegants!


no, it doesnt need to look “classic”. Anyway, haiku doesnt have multi user support… and also will not for a long time. So it’s senseless to be consistent with the current design.
Multiuser support witll be somewhere… much later (after R1,… or R2, who knows). Till then I guess also the GUI of haiku will improve a lot, having hw accelerated appserver… with transparency… and all that stuff.

Of course somebody could make such a log-in app, to have haiku password-protected… I people would use such an app.

If you want to be really useful… than think about a great user interface, for example for my DocumentViewer. I asked already people to think about a user interface that is more modern, and usefull… but i didnt get good feedback.



i found some good idears.

I like:

  1. http://3.bp.blogspot.com/_wiFcIJb18iU/Swftb7lmzJI/AAAAAAAAACc/rijW6cNtN64/s1600/Screen+shot+2009-11-21+at+2.01.59+PM.png



i vote for a mix 1 + 2 (3) and 7 !



Hey, so I took a look at your DocumentViewer v0.3.2 and I must say, I was impressed! The layout is logical, easy to use, and this is quite a many piece of software. I installed it, and I was pleased with the results! Great job, I don’t know why you didn’t get good feedback, I quite enjoyed using it…


ok we have winner

simplicity and usable and beautifull design… for me is haiku about simplicity… this can be good:

but when you want graphical efects i wote for something like this



I think it looks nice. I appeals to me since I am a BeOS-person…
But I think the user:password-design is, in general, outdated.
It shows usernames - bad idea. 50% of the intruder’s job is done. If username is YOUR name, then it’s 75% (it makes guessing password much easier).

I propose:
On login screen there is bunch of images (or avatars) from which to choose.
On login screen there is a classic password prompt.
On login screen there is a numeric keyboard (like the one you have under numlock).
All visible, all the time.

To log in user must choose a correct combination of those. Only he knows if it is image+password or image+numpad or image+image and so on, maybe even image+image+numpad or maybe just a password.
This makes entry much faster for real users, yet intruder has no clue from where to start and in what sequence to go by. Is it password and image? image and image, then which ones? just numpad series? password and numpad? etc…

On less vulnerable systems (at home) one could just use an image, on laptop one would use sequence of all three, maybe.

On the system side: the login image can be made to a hash that is the “real” username or password - a series of numbers and letters.
Two of the three can be combined, so on sytem side there still remains just the classical username:password-combination.

Why the numpad option?
It’s very fast data entry point and one can make “patterns” on it (Android has that?) if he forgets numbers easily. And if Haiku goes mobile, well we are set then. :wink:

Just some thoughts, I’m sorry I don’t have time to present them in a more comprehendable manner.

PS. Why a window? Will You have other windows on login screen?


yes for me that is not “good feedback”. Good feedback for me means, criticism, telling me how to improve it. I encouraged people for “gui mock-ups”, but in fact nobody came with an idea (except karl, who showed me a few videos of a viewer on mac).

Related to your login-screen. It’s very useless, for a developer. Not only because it’s not needed by now, but because it is very trivial. Something like you “designed” is very trivial, and every beginner in haiku-programming that is also artistically handicaped would have come with something similar or better.
That doesnt mean, it should be done like that, it means just, that your “design” is not a help for a developer since anyone that wants to make such a login-screen , would be doing something like that or similar or even better.

So you are just wasting your time and the time of other people. And I guess that’s also your intention to waste time to get a little attention with less work.
If you really want to help haiku then try to invest many month of work into a mockup… or whatever. But real effort, not a quick and small “hack”.

please dont take my words as an insult. It’s just the impression that i have and of course i can be wrong.


I am just waiting for a new Haiku Release…

I dont care much about design until it will become beta…

I am just happy if there is some progress on Haiku…

No feedback is good feedback sometimes… I watched and installed your document viewer and this app is running… design will be easy to change in the future…

once the os is working… design and apps will follow…

Maybe I am wrong, but I dont want to spent much time on thinking about design… more over good programming skills are more important for an OS…

I just wait and dont write and spent much time on thinking about design and other things like “haiku only apps” but I still follow Haiku developement by reading the progress it is doing so far…


These mock-ups would be better reposted on the design mailing list( http://www.freelists.org/list/haiku-cdt) where developers can interact, but I don’t see the problem with people posting their mock-ups on this forum; people have been doing this for years.

As for the uselessness of the endeavour, some of the tasks for the Google Code-In programme for teenagers that Haiku participates in has been to create mock-ups for things such as the package manager. The purpose of this was not to create a finalised design but as a means to create discussion among the developers and to aid the young person in dealing with the feed back that is found when creating such commissioned works.

Further as Ryan notes:

The multi user part is indeed years away, but nonetheless the login screen can serve a purpose before then. Depending on the the difficulty, a task or multiple tasks could be created for this years Google Code-In. This would be to create a optional login screen for security conscious user who wishs to not have haiku simply auto load the desktop of Baron (its still called Baron right?). This mock-up could then serve as general guide to how it could look.


I think a log on screen is very important; if only to prevent my kids from logging in to my account and mucking with things! Even if it doesn’t actually provide a secure file system and even without allowing multiple users per se …

So, yes, a simple application that prevents loading the desktop until the password is used would be very beneficial indeed!

And, if I can be so bold as to also suggest a basis for the UI … it should use the desktop background and only show the system user’s name, a password field, and a login button (for those that don’t press the enter key -there are always some!), and not much else. Maybe put the os name and version very small in the corner…

here’s why :

The Haiku’s OS is all about the user. It’s his (or hers). It’s simple to use. It gets out of the way so the user can be productive right from the start…

Just my 2 cents.


Currently, a login screen can be achieved using:
Is also customizable… This is the mine:

Lock Haiku

Lock Intruder

Click on thumbnail to full view


That’s a great idea. All you need is a screen locking application. I was going to suggest one, but was surprised that I couldn’t find something like that on Haikuware. :frowning:


You should read my post above… :wink: LockWorkstation can also act as Screen Locking… Ok it’s manual, but you can also put LockWorkstation inside BootScript and also calling it on demand!


Gesture login? What’s next, whistle recognition? Drum breaks through a MIDI controller? As cb88 points out, any such novelty login system will have to either demand enough precision that it would be too annoying to use, or be not even as secure as a simple weak password login. You might as well not have it at all and just boot straight to desktop - which is what Haiku already does.


I hate ALL login screens. I want something completely original and no screen where there is a user’s icon, photo, or avatar. I want a Mouse Gesture Login Authentication. But that’s me.

Have a look at http://graphem.berlios.de for a general idea.

There are many other projects in the works that are looking to redefine user interactum. Here for one, http://youtu.be/zWz1KbknIZk is pretty stunning!


It seem to me that gesture based strategies could be easily broken if you had a way of faking a mouse (so anybody with an arduino?). I’m thinking that would be true since there are a the limit on the gestures is really how complex a gesture can the software accurately recognise without missrecognising and also how complex the gesture the person chooses is…

Say each movement would be worth less than a single charachter since I doubt any software would be more complex than that (22.5 degrees between each possibility of motion from the starting point). Even if it were better than that its still not nearly as good as even n alpha numeric input which would be over 36 possibilities without any symbols at all!

That said it would probably be possible to make take longer to test each possiblity… at least 3-4 seconds to input it. and 30sec if you fail or something like that maybe 5min if you fail multiple times but that would work for any password shcheme. That said gesture based is VERY nice on mobile devices… dispite the possiblity that it would be easier to figure out your gesture than a password.


I must concede to your correctness on certain points, however in this case we’re speaking about Haiku which has NO protection from loading the profile. It’s just an idea that makes sense to me. Maybe I don’t make a lot of sense? laughs at self

Future proofing Haiku is very a good idea since there is quite of bit of stir porting it to ARM, I see this as a great opportunity for prempting a move to an even smaller platform like a 7" tab or even 4.6" phone (yes I know the phone portion wouldn’t work anymore).


Oy…like a bad penny, this “tablet/phone Haiku” notion just keeps turning up. Whatever your feelings on tablets, the thing is that they require, by their very design, a radically different approach to UI than desktop OSes. (Which is why nobody bought XP tablets back when they were a thing.) Haiku isn’t that, and you’d have to change it beyond recognition to make it suitable for a tablet, and then you’d be left with the fact that you’d have even less usable software than you do now, because all Haiku software is designed for the desktop.

Tablet Haiku isn’t “future-proofing,” it’s a dead end. There’s absolutely no point to it, and if anybody were seriously pursuing it instead of just discussing it hypothetically in the forums, it’d be a massive waste of effort.


I can understand that you are fond of the current design of Haiku, and that you opine so strongly about almost everything that you reply to, but be honest at least, the very reason BeOS came to be (no pun) is because of radical thinking by devs that wanted to make someting that was interesting and powerful.

Just because you could port Haiku to a tablet in a year from now, doesn’t mean that it would turn Haiku into an invalid or hurt it in any way! You never know until you really use it in such a manner. Adaption is key to the success of an OS. I do not mean commercial success, I just mean a viable alternative to accomplish something very cool. I can perhaps come up with 10 projects that I could use Haiku as a base for because of its speed and usability with limited resources that you would normally find on tablets.

Let me list a few, if not all 10:

  • High End Tablet Remote to control an autonomous home.

  • High End WiFi Audio Player using 3D Mix in conjuction with a Music Track-by-Track download portal that would allow you to mix other peoples’ music tracks, just the ones you want to add into the song.

  • A dedicated remote-desktop light server in an arduino sized package.

  • An embeded PBX Controller.

  • Live Multi-Track Audio Recorder.

  • WiFi Control Surface for just about any application:
    – lighting schemes for concerts
    – portable radio station for interviews in the field that are live-to-air
    – in dash car HUD

Jeez, I have to stop myself now because I can just totally see the use of this OS well beyond the desktop… all because of it’s basic desgin.

P.S. - I checked out your site, and I love your choice of music and books. You’re alright in my books Cdre-John. :wink: But please don’t hold back in reply because I paid you a compliment.


Heh, I won’t :smiley:

See, the thing is that what you’re saying is what a lot of tablet advocates have been saying ever since they started being a “thing” a couple years ago - namely, responding to the general criticism by pointing out X, Y, and Z situation where someone’s using them for something. But just because you can use a tablet for some task doesn’t make it necessarily a good fit for the job; more often than not it’s a kluge just to get it to work, when a real general-purpose computer would fit the bill much better.

Of your examples, the only ones where it would even be as good a fit as a proper desktop are the remote and the wifi control surface - and even those are extreme corner-cases. (Autonomous homes have been perpetually “just around the corner” for about fifty or sixty years now, and there’s only so many control setups that are equally well-represented by touchscreens as they are by discrete physical controls and readouts.)

So arguing that tablets overall are comparable to a real computer by way of those examples is like saying that a sharpened rock is as good as a modern box of tools because it can fill most of those roles with enough creative problem-solving and elbow grease. Technically it might be true, but it’s missing the point that the toolbox is a massively more versatile and capable solution.

All of which is only halfway relevant anyway, as my point wasn’t that tablets are useless (they are, but that wasn’t the point I was trying to make,) but that Haiku (like any other desktop OS) is just plain not well-suited to the UI constraints of a touchscreen-only device, and trying to contort it to fit those constraints is a waste of effort because then you’d have an OS/shell that worked well on a tablet from which you could run…a bunch of applications that still didn’t.

(You could maybe reuse the Haiku internals to build a separate tablet/phone OS around, as Google did with Linux and Android, but you’ll note that even Android didn’t really hit it big until it had a decent base of Android-specific software to rely on. An OS without applications is just a nice car with nowhere to drive to.)

You’re right that BeOS got to be great by deciding what it wanted to do and then doing it really well, without bothering itself about how the major contenders did things. That’s exactly why trying to turn Haiku into some kind of nebulous all-things-to-all-users God-OS is a terrible idea; Haiku is derived from BeOS, which was designed specifically to be the best damn desktop computer OS out there. You’ll no more get good results trying to turn it into something other than a desktop OS than you would trying to hammer nails with a pocket watch.