Apple Tries to Patent Fast User Switching

Ashcrow writes "An article from The Register points out Apple's attempt to patent fast user switching. It seems that Steve Jobs admits that Microsoft beat them to the punch but believes Panther's implementation is superior."

Comments

Direct link to feature description on Apple

[$$$$$exyGal]

Here's the Direct Link.

Re:When was this last on Slashdot?

[nacs]

You've probably just seen it on other sites since the news has been out for a while now.

Can someone please read the article

[Anonymous Coward]

because it appears that nobody here is able to. They are not patenting fast user switching per-se but changing application settings and preferences on the fly, such as what location manager is doing.

Re:Actually unix beat them both

[madcow_ucsb]

It's definitely possible to run multiple X sessions on different virtual consoles. Sucks a bit of ram, but you can definitely do it.

Re:Actually unix beat them both

[gantrep]

Sure, multipse X sessions are easy! Here's a very nice tutorial on how to do that.

That's like saying...

[itistoday]

...that UNIX beats itself. OS X is UNIX, and user switching is really just as fast as on any other *NIX based system, the only delay is Apple's insistance on having a perty Quartz graphics transition between it.

Re:That's like saying...

[Anonymous Coward]

OS X is not UNIX. Unix is a trademark of the Open Group. They say it is not a Unix. Stop trying to dilute their trademark.

Repeating the same lie over and over doesn't make it true.

Legit

[alset_tech]

If you read the article, you will see that Apple has filed a CONTINUATION of a patent involving user-switching, originally granted in 1995. This is a completely legitimate move on their part. It is arguable that they have owned this technology for years. As for using SU for user switching, this would be true as an earlier form, however the patent applies to the GUI. Dan

Important Quote

[jmt9581]

Straight from The Register story (the The Register story?):

At WWDC, Jobs admitted that Microsoft had beaten Apple to market by offering such a feature in Windows XP, but he claimed Apple's implementation was the better of the two.

That would imply, surely, that Microsoft has a solid prior art claim?

No. The current application, filed last November and updated this past June, turns out to be a continuation of a patent, number 6,512,525 filed in August 1995, long before Windows XP arrived, and finally granted in January 2003 with the same title. That patent is also assigned to Apple.

:)

Re:SU

[The Turd Report]

Actually it is a means by which anyone can become anyone else, with the correct password.

Actually, no

[cscx]

su actually stands for "Switch User" I believe, and you can in fact switch to any user:

SU(1) OpenBSD Reference Manual SU(1)

NAME
su - substitute user identity

SYNOPSIS
su [-fKLlm] [-a auth-type] [-c login-class] [login [shell arguments]]

DESCRIPTION
su requests the Kerberos password for login (or for ``login.root'', if no
login is provided), and switches to that user and group ID after obtain-
ing a Kerberos ticket granting access. A shell is then executed, and any
additional shell arguments after the login name are passed to the shell.
su will resort to the local password file to find the password for login
if there is a Kerberos error or if Kerberos is not installed. If su is
executed by root, no password is requested and a shell with the appropri-
ate user ID is executed; no additional Kerberos tickets are obtained.

startx -- :1

[Erisian Pope]

On linux (and probably any other system with XFree86 ) To get to the first virtual console Use: CTL+ALT+F1. Then login and type:

someone@server someone]$ startx -- :1

X windows starts using the next available console. To switch between X sessions use CTL+ALT+F7 and CTL+ALT+F8. To start more sessions use :2, :3 etc. This has been available (but not well documented) for many years. Have Fun!

Re:Actually unix beat them both

[mcgroarty]

It's definitely possible to run multiple X sessions on different virtual consoles. Sucks a bit of ram, but you can definitely do it.

It's not only possible, but some of us have been doing this for over ten years. If Apple pursues this patent, I'm definitely calling prior art on this.

Re:SU

[optikSmoke]

Of course, some environment variables are still based off of the original login, so if you have trouble there then you need to logout and login. In most cases, though, su works just fine.

You can simply include the "-" arg if you want to fix all the env vars and get a proper login shell: "su -" will give you a root shell, "su - joe" will give you a shell for joe, etc -- all variables, etc properly set.

Re:Skins == Trademark

[tupps]

The case against UI skins maker was the use of the Apple Logo and calling it Aqua. Most of the people removed the apple logo and changed the name and everything was ok.

Re:Actually unix beat them both

[zenyu]

Anyone know if KDE/Gnome or even Xfree is planning something like this? I heard talk about multiple X servers, but its not out of the box simple use, of even possible.

If you use gdm to login, add the line "1=Standard" after "0=Standard" in your gdm.conf. If you use kdm I think you just add the line ":1 local@tty1 /usr/X11R6/bin/X vt8" after the line ":0 local@tty1 /usr/X11R6/bin/X vt7", but I don'y use kdm, so your milage may vary. (xdm is more complicated, so google if ya use that). As suggested by the kdm config to start a new X server on another virtual terminal just specify the vt you want to use. I think this has been around since shortly after XFree86 was first ported to Linux, maybe earlier on the BSD's. Recently it's been possible to program virtual terminal switching to keys other than the basic F1-F12, so easy switching isn't limited to just 12 users anymore. I never understood why multiple X servers haven't been used in the Linux distro's, at least on a "allocate one X terminal per 256 MB of RAM the computer has" basis. My desktop has had a gigabyte or more of RAM for years, I'm not really concerned about a few extra buffers eating up a tiny bit of memory. Even my laptops with 256MB-512MB in the last 5 years can handle an additional X server without batting an eye.

You can also give the different servers different configurations, which is the traditional use for this. But by default the X server started by kdm/gdm requires a login and uses the same config, so it is exactly what you want. BTW if you want to be able to login with the same user twice you will have to enable that, by default it is not permitted to prevent remote users from starting lots of X servers and consuming all your resources... (though this is also limited by the number of virtual terminals you allow.)

I don't think so

[2nd Post!]

Unless you're familiar with Location Manager and can disagree with that implementation, I don't believe there *is* prior art.

Or rather, that Mac OS (Classic, not X) *is* the prior art, and that *Apple* owns the original patent, of which this is merely a continuation if you read the article.

Location Manager allows a single user to change multiple settings on a computer with a single selection:

Wireless Plugged
Wireless Unplugged
Netless Plugged
Netless Unplugged
Wired Plugged
Wired Unplugged
At Home
At Work
Roaming Unplugged

So that with a single selection the user can change:
Bandwidth settings on Qucktime and the network
Power/Battery/Energy saver settings
Screen Saver settings
Printer settings
Network settings (DHCP and Proxies)
Browser settings ... AppleScripts to manage everything else

If you look at 'Multiple locations - User' :: 'Multiple users - Machine' then it does appear valid that Fast User Switching is a continuation of the original patent Apple holds regarding the technology that is Location Manager.

su is *not* fast user switching, it is just changing the user locally in a terminal; and it doesn't change the settings of the machine, applications, or even the environment *outside* the terminal, unlike FUS or Location Manager.

Re:I don't think so

[2nd Post!]

Okay, so the connection between fast user switching at location manager?

However, I will answer an unasked question, where you question the value of location manager.

My laptop starts at home, plugged into power but with airport. I use that setting.

I go to work, where I have wired ethernet and power. Mail settings change, as do proxies.

I then go to a meeting, where I have no wireless or wired, and no power. Battery settings correspondingly change.

A second laptop user wants to create an adhoc network to transfer files after the meeting. I switch to the wireless unplugged.

Go back to my desk, and it's back to 'work' settings

I then go home, and it's back to wireless plugged. But perhaps a friend calls, and I drop by. I am now roaming unplugged to check my email, and then I put my system to sleep while we watch a DVD over a beer and snacks.

I then go back home, and back to wireless plugged.

You can question the value all you want, but this is how I live with my laptop ^^

Now this isn't possible under XP: I have friends with XP laptops, and they have to shut down and restart when they change from mode to mode: Wired, wireless, unwired. The corresponding proxy changes, DHCP changes, etc, also require fiddling with the network control panel. And any corresponding changes to bandwidth uses also require fiddling in the respective programs, and printing preferences, etc. My printers at work and at home were automatically configured via Location Manager. Heck, even the display calibration of the monitor at work and the monitor at home.

So now you ask about FUS?

Okay, so in *my* situation, I'd have multiple logins for *myself*.

An admin login, with a bright red desktop where I do all my admin stuff.
My 'personal' login where I have my email, IM, and browsing apps
My 'work' login where I might use Photoshop, iDVD, iMovie, and other stuff.
My 'developer' login where I would have XCode and a handful of terminals, X11, and some man pages loaded up.

Just for ME

And I might live *mostly* in the personal login, but I would switch once in a while to admin to install or maintain stuff, I would switchinto work to do worklike stuff a little more often, and developer whenever I felt the need to, say, build mozilla, or code something.

If my friend came over and wanted to use the web, I would switch to a 'guest' login, or request that they do so.

This is not something I would do on my PowerBook, but rather my PowerMac.

How does this relate to Location Manager? Because it is me, one user, with multiple use models, first handled by location manager for my laptop to handle certain configurations, and then handled by fast user login to handle certain usage patterns. One requires mobility, the other requires a lot of CPU and ram (Hmm, like a G5)

is this useful for you?

Possibly not. You think location manager is of dubious use to you, but I know I loved it.

I also know with FUS that my dad would have one account for him to handle scanning and photos, since that's what he does, and I would give my friends a guest account for web, email, and IM, while I would use 3 or 4 accounts myself.

Again, that is the value to ME and that is why *I* think it's worth existing and possibly even a patent. Just because it's worthless to you, or anyone else reading, does not deny the value of FUS or LM.

Re:I don't think so

[steeviant]

All of the things you mention sound vaguely useful, for someone clueless about the actual workings of their machine.

Hmm, so if I were 'clueful', rather than moving to the top of my screen, I'd much rather click on the terminal and type in a line to set my new network address, a line to delete the default route, a line to add a new default route, and then edit the /etc/resolv.conf line and change the nameservers, and then go to the preference pane for internet settings and change my cache settings every single time I move my laptop between the office and home.

Not every network in the world has DHCP on it, the location manager bundles all of this stuff up in a nice, easily accessible script. Thank you for accusing me of being clueless you gormless twat.

under OS-X, 90% of the "useful" progs run setuid root anyway, so again, you basically have every user a superuser. Why switch between equally useless (or open to abuse, depending on your perspective) accounts?

I also take issue with the idea that most useful programs run with root priviliges, this statement makes me wonder if you've ever laid eyes upon an OS X machine in your life, but more disturbingly, it makes me wonder what you actually meant by your statement.

Is there some new found study linking the percentage of programs running as root with security?

I certainly haven't found anything that allows me to do things I shouldn't be able to do, and programs that require root priviliges usually seem only to want the priviliges to accomplish the task they require before dropping back to normal user. The fact that OS X has a system wide consistent interface for doing this inside a GUI makes it MORE secure than most other Unix systems, which just choke when this happens, thus requiring you to run the program with root priviliges, making your system insecure if you have to leave it unattended for any length of time.

In case you were wondering, my OS X machine shows 70% of it's tasks running with non-root privilige, while my OpenBSD machine shows only 43% of it's tasks running with non-root privilige, therefore I conclude that Mac OS X is more secure than OpenBSD. Logic supplied by pla.

Has anyone here actually read the pattent?

[RalphBNumbers]

Has anyone here actually read the patent that the Register article links to?
It covers "Multiple personas for mobile devices".

It's a hell of a stretch to go from that patent to fast user switching. The Register even admits it's a inaccurate description of user switching, although they underplay it.

That patent sounds like it would more accurately describe a handheld device that could serve multiple roles (like a mp3 player, a movie player, a camera, a phone, etc) and could rapidly reconfigure it's GUI to accommodate whichever 'persona' the user wanted.

I'd say this is just the Register blaring sensationalist bullshit to get attention (and succeeding wildly since they have a front page /. story now), with only their wild guess as to what Apple is actually patenting.

Re:Actually unix beat them both

[asteinberg]

Surprised I haven't seen any posts specifically describing this yet, but in KDE 3.1 (at least the version in Debian, but I think it's distro-independent) there's a "Start new session" option in the k-menu which opens up kdm in a new x session. You can also access this when the screen is locked - there's a button in the password prompt to start a new session.

Re:SU

[CatOne]

su is not the same as fast user switching. I don't know whether you've used "fast user switching" on windows or OS X, but they are not at all the same. su allows you to become another user (really, in the shell). fast user switching lets you switch to a desktop which is for a specific user -- like a normal windows or OS X login. But you can have many of these running concurrently -- say one user has a window with their browser running, and one with a mail client runnning, etc. Fast user switch to another user and do whatever you want... the other users desktop is unmodified and runs just fine in the background. Of course this is a resource hog... if you have 200 MB of GUI apps running for 3 or 4 users, they're all taking up their own process space and RAM.

Re:patents and implementation

[Jokkey]

Microsoft uses a thin local Terminal Server that's running in the background to achieve this. Quite elegant actually. The only problem with this is that it adds another layer between user & hardware, so there's a performance hit.

I was really impressed with how WinXP handled itself in a little test I devised: Start installing MS Office in 1 account, switch to user2 and start surfing the web, fetching e-mail, etc. Switch back and hey presto! office was installed. I say: sweet. that's a good implementation. I don't know anyone who would want to do that on a daily basis, but hey, it's proof of concept.

Except that Microsoft's fast user switching doesn't work if your computer is joined to a Windows domain. Which makes it pretty useless for a lot of settings. Example: Fast user switching would be very convenient for my wife and I, but since my home PC is joined our the campus domain, it's disabled.

Re:Has anyone here actually read the patent?

[Aapje]

Has anyone here actually read the patent that the Register article links to?
It covers "Multiple personas for mobile devices". That patent sounds like it would more accurately describe a handheld device that could serve multiple roles (like a mp3 player, a movie player, a camera, a phone, etc)

A laptop is a mobile device too. In fact, the patent explicitly makes a distinction between a hand-held computer and a portable computing device (see claims 34/35, 43/44 and 50/51 from the patent). In the desciption it states that: "Computers are becoming increasingly [...] portable. [...] Laptop, notebook, and sub-notebook computers are virtually as powerful as their desktop counterparts." In other words, the class of portable computing devices referenced in the patent includes laptops.

It covers "Multiple personas for mobile devices". It's a hell of a stretch to go from that patent to fast user switching. The Register even admits it's a inaccurate description of user switching, although they underplay it.

The term 'user' as used in Unix is really just a persona. It's certainly not equal to a person since the root and guest accounts are not directly related to a person. They are more like roles, aka personas. Multiple personas seems to be an valid description of (fast) user switching.

I'd say this is just the Register blaring sensationalist bullshit to get attention (and succeeding wildly since they have a front page /. story now), with only their wild guess as to what Apple is actually patenting.

Unfortunately, I haven't got the time to examine the entire patent (it's very long), but your criticisms don't seem to hold water. So for the time, I'll have to give The Register the benefit of the doubt.

PS. Quotes were edited for readability.