Timezone should not require superuser on laptop

[gpk]

Binary package hint: gnome-system-tools

When you carry a laptop from one timezone to another, it's nice to be able to set the time correctly, even if you don't have administrative rights on the laptop.

ProblemType: Bug
DistroRelease: Ubuntu 10.10
Package: gnome-system-tools 2.32.0-0ubuntu1.1
ProcVersionSignature: Ubuntu 2.6.35-24.42-generic 2.6.35.8
Uname: Linux 2.6.35-24-generic x86_64
NonfreeKernelModules: nvidia
Architecture: amd64
Date: Fri Dec 31 18:30:08 2010
EcryptfsInUse: Yes
InstallationMedia: Ubuntu 9.10 "Karmic Koala" - Release amd64 (20091027)
ProcEnviron:
 SHELL=/bin/bash
 PATH=(custom, user)
 LANG=en_GB.utf8
SourcePackage: gnome-system-tools

[Milan Bouchet-Valat (nalimilan)]

Everybody shouldn't be allowed to set the system timezone by default, as it affects all users on the system. If you don't have administrator rights on your laptop, you can ask the administrator to grant you the rights for the org.gnome.clock.mechanism.settimezone PolicyKit action, but this decision can't be made by Ubuntu.

Note that if you travel a lot, you may find convenient to define locations in the preferences of the panel's clock. This way, you can see the time in all the places you're interested in. Maybe the clock should support showing the time according to a location without changing the system timezone, but that would be another bug.

[gpk (gpk-kochanski)]

Oh, that's silly!

It's a laptop, darn it! The other users will (almost certainly) be on the console, and they'll be in the same time zone your are in. (Unless they have *very* long arms.)

What you are proposing is not the normal use case for a laptop -- you seem to be thinking that I carry my laptop around so that people can ssh into it. That's quite rare.

[gpk (gpk-kochanski)]

Sorry. Let me be a bit more formal.

I suggest that laptops are either:
1) single user
in which case changing the time zone won't hurt other users,
2) physically shared amongst several people

[gpk (gpk-kochanski)]

Sorry. Let me be a bit more formal.

I suggest that laptops are either:
1) single user
    * in which case changing the time zone won't hurt other users,
or
2) physically shared amongst several people
    * in which case, they are in the same time zone, and if one user sets the zone, that is a positive benefit to other users,

Occasionally, they might be
3) multi-user, shared amongst a small cooperating group of users. In which case if someone fools with the time zone, you send them a nasty e-mail.

And they will very rarelyl be used as servers occupied by untrusted users out to cause trouble.

In use cases 1 and 2, changing time zones is a real benefit. In 3, it is largely harmless.

[Charlie Kravetz (cjkgeek)]

the changes you are requesting aren't really a bug and require more discussion, which should be done on an appropriate mailing list or forum. http://www.ubuntu.com/support/community/mailinglists might be a good start for determining which mailing list to use.

 As stated above, the answer for an individual without admin rights is :

If you don't have administrator rights on your laptop, you can ask the administrator to grant you the rights for the org.gnome.clock.mechanism.settimezone PolicyKit action, but this decision can't be made by Ubuntu.

[Milan Bouchet-Valat (nalimilan)]

We can't change PolicyKit permissions depending on whether the system is a laptop or not. So we'd have to make the change for all systems, which is not necessarily what people expect. And even laptops can be used e.g. in kiosks or schools, where you don't want people to play with system settings.

I think the whole point will become obsolete anyway when GNOME 3 introduces automatic detection of location based on network. See http://live.gnome.org/Design/SystemSettings/DateAndTime for a rough mockup, which is partly implemented upstream.

[Vish (vish)]

Alan , we dont assign bugs to the team. Usually bug is assigned to folks *fixing* the bug, not to folks triaging the bug.
for more info : https://wiki.ubuntu.com/Bugs/Assignment

[gpk (gpk-kochanski)]

OK. I agree it would not make sense as the desktop default.

On 01/01/11 14:53, Milan Bouchet-Valat wrote:
> We can't change PolicyKit permissions depending on whether the system is
> a laptop or not. So we'd have to make the change for all systems, which
> is not necessarily what people expect. And even laptops can be used e.g.
> in kiosks or schools, where you don't want people to play with system
> settings.
>
> I think the whole point will become obsolete anyway when GNOME 3
> introduces automatic detection of location based on network. See
> http://live.gnome.org/Design/SystemSettings/DateAndTime for a rough
> mockup, which is partly implemented upstream.
>

[Alan Burgess (alan-burgess)]

Yes, I figured as much.
I didn't intend to do that, but then I couldn't figure out how to un-assign
it.

Sorry, still learning how the bugtracker works.

If I see a bug that I think I can help with, is it ok for me to leave
comments in the bugtracker?

On this particular bug, I only brought it up in ubuntu-bugs irc. this
morning, some people had been working on it.

On Sat, Jan 1, 2011 at 9:16 AM, Vish <email address hidden> wrote:

> Alan , we dont assign bugs to the team. Usually bug is assigned to folks
> *fixing* the bug, not to folks triaging the bug.
> for more info : https://wiki.ubuntu.com/Bugs/Assignment
>
> ** Changed in: gnome-system-tools (Ubuntu)
> Assignee: Ubuntu BugSquad Mentorship group α
> (bugsquad-mentorship-group-alpha) => (unassigned)
>
> --
> You received this bug notification because you are a direct subscriber
> of the bug.
> https://bugs.launchpad.net/bugs/696115
>
> Title:
> Timezone should not require superuser on laptop
>
> Status in “gnome-system-tools” package in Ubuntu:
> Invalid
>
> Bug description:
> Binary package hint: gnome-system-tools
>
> When you carry a laptop from one timezone to another, it's nice to be able
> to set the time correctly, even if you don't have administrative rights on
> the laptop.
>
> ProblemType: Bug
> DistroRelease: Ubuntu 10.10
> Package: gnome-system-tools 2.32.0-0ubuntu1.1
> ProcVersionSignature: Ubuntu 2.6.35-24.42-generic 2.6.35.8
> Uname: Linux 2.6.35-24-generic x86_64
> NonfreeKernelModules: nvidia
> Architecture: amd64
> Date: Fri Dec 31 18:30:08 2010
> EcryptfsInUse: Yes
> InstallationMedia: Ubuntu 9.10 "Karmic Koala" - Release amd64 (20091027)
> ProcEnviron:
> SHELL=/bin/bash
> PATH=(custom, user)
> LANG=en_GB.utf8
> SourcePackage: gnome-system-tools
>
> To unsubscribe from this bug, go to:
>
> https://bugs.launchpad.net/ubuntu/+source/gnome-system-tools/+bug/696115/+subscribe
>

--
Do not meddle in the affairs of wizards, for they are subtle and quick to
anger.

[gpk (gpk-kochanski)]

Actually, on consideration, I don't think that automatic detection of
the timezone
is a viable solution. It suffers from all kinds of potential problems
like
"What does it do if it's connected to a VPN?". If I connect via VPN
to my
office in the UK when I'm in France, I'd get the incorrect time.

Or, if I'm in the airport and I'm not yet connected to a network in my
new location,
it will be silently giving me incorrect timezone information. It'll be
telling me
what time it is in the other country. That's acceptable if you know
you have to
set it manually, but people will depend on an automatic system, and miss
their
connecting flights.

The trouble this will cause is a classic computer problem: it will be a
system that is
smart enough to be right most of the time. Being mostly right, people
trust it
and depend on it, and then get surprised in those cases where it is wrong.

Besides, there is a laptop-detect package in Ubuntu:

> $ aptitude show laptop-detect
> Package: laptop-detect
> State: installed
> Automatically installed: no
> Version: 0.13.7ubuntu2
> Priority: optional
> Section: utils
> Maintainer: Ubuntu Core Developers <email address hidden>
> Uncompressed Size: 57.3k
> Depends: dmidecode (> 2.8-2)
> Description: attempt to detect a laptop
> laptop-detect attempts to determine whether it is being run on a
> laptop or a
> desktop and appraises its caller of this.
>
So, it is certainly technically possible to tell if you're on a
laptop. Therefore gnome can make
decisions based on that fact.

On 01/01/11 15:48, gpk wrote:
> OK. I agree it would not make sense as the desktop default.
>
> On 01/01/11 14:53, Milan Bouchet-Valat wrote:
>> We can't change PolicyKit permissions depending on whether the system is
>> a laptop or not. So we'd have to make the change for all systems, which
>> is not necessarily what people expect. And even laptops can be used e.g.
>> in kiosks or schools, where you don't want people to play with system
>> settings.
>>
>> I think the whole point will become obsolete anyway when GNOME 3
>> introduces automatic detection of location based on network. See
>> http://live.gnome.org/Design/SystemSettings/DateAndTime for a rough
>> mockup, which is partly implemented upstream.
>>

[Milan Bouchet-Valat (nalimilan)]

Le mardi 04 janvier 2011 à 15:51 +0000, gpk a écrit :
> Actually, on consideration, I don't think that automatic detection of
> the timezone
> is a viable solution. It suffers from all kinds of potential problems
> like
> "What does it do if it's connected to a VPN?". If I connect via VPN
> to my
> office in the UK when I'm in France, I'd get the incorrect time.
People using VPNs are not the majority, and we can disable automated detection
of timezone if we're connected over a VPN. Plus, we can also use GPS devices
when present.

> Or, if I'm in the airport and I'm not yet connected to a network in my
> new location,
> it will be silently giving me incorrect timezone information. It'll be
> telling me
> what time it is in the other country. That's acceptable if you know
> you have to
> set it manually, but people will depend on an automatic system, and miss
> their
> connecting flights.
If people are smart enough to change their timezone manually, they are smart
enough to understand their laptop still uses the previous timezone
because it has no way of finding out you've moved.

> The trouble this will cause is a classic computer problem: it will be a
> system that is
> smart enough to be right most of the time. Being mostly right, people
> trust it
> and depend on it, and then get surprised in those cases where it is wrong.
So you'd suggest computers should always wait for users to do things manually,
just to avoid being smarter than them? The automated solution is better
than nothing. Anyway, discussing this here is pointless since it will be
implemented in GNOME and I'm not involved at all in this work.

> Besides, there is a laptop-detect package in Ubuntu:
>
> > $ aptitude show laptop-detect
> > Package: laptop-detect
> > State: installed
> > Automatically installed: no
> > Version: 0.13.7ubuntu2
> > Priority: optional
> > Section: utils
> > Maintainer: Ubuntu Core Developers <email address hidden>
> > Uncompressed Size: 57.3k
> > Depends: dmidecode (> 2.8-2)
> > Description: attempt to detect a laptop
> > laptop-detect attempts to determine whether it is being run on a
> > laptop or a
> > desktop and appraises its caller of this.
> >
> So, it is certainly technically possible to tell if you're on a
> laptop. Therefore gnome can make
> decisions based on that fact.
We can, but PolicyKit really isn't designed around the idea of changing
policies depending on that kind of setting. And the fact that your
computer is a laptop doesn't mean you want to grant this kind of
permissions to all users on the system.

[gpk (gpk-kochanski)]

On 04/01/11 20:52, Milan Bouchet-Valat wrote:
> Le mardi 04 janvier 2011 à 15:51 +0000, gpk a écrit :
>> Actually, on consideration, I don't think that automatic detection of
>> the timezone is a viable solution. It suffers from all kinds of potential problems
>> like "What does it do if it's connected to a VPN?". If I connect via VPN
>> to my office in the UK when I'm in France, I'd get the incorrect time.
> People using VPNs are not the majority, and we can disable automated detection
> of timezone if we're connected over a VPN. Plus, we can also use GPS devices
> when present.

OK. Kind of OK, anyway. But that means a traveller has to know
and remember the rule
about VPNs. If they don't, they will swear at you. You will get
people saying "I know I
connected to the network, why didn't the damn thing set the clock properly?"

And, doubtless, developers will respond "RTFM". But RTFM is a
hopeless task these days!
My Ubuntu system has tens of thousands of pages of documentation. You
have to design a
system that will not mislead people in important ways, even if they have
had better things
to do than read the Gnome documentation.

>> Or, if I'm in the airport and I'm not yet connected to a network in my
>> new location, it will be silently giving me incorrect timezone information.
>> ... That's acceptable if you know you have to
>> set it manually, but people will depend on an automatic system, and miss
>> their connecting flights.
> If people are smart enough to change their timezone manually, they are smart
> enough to understand their laptop still uses the previous timezone
> because it has no way of finding out you've moved.
>

Wrong. It's a very different process in the two cases.

Manual case: you just have to remember "did I set it or not"? *You*
are the only actor.

Automatic-but-imperfect case: You have to understand how the automatic
time-zone setting logic works.
1) You have to know that it gets its time from looking up its IP address
in a table.
2) You have to keep track of whether the computer has connected to a
network.
   2a) It may connect without your knowledge -- if you go somewhere you
have been before, it
          may detect up a WiFi network where it has the password and
autoconnect.
   2b) It may fail to connect somewhere you expect a connection.
Hardware could be down,
          or overloaded, or the signal could be too weak, or a password
expired...
3) Even if it does connect, do you know that it has a global IP address
in the right country?
     3a) There's the VPN problem.
     3b) There's the possibility that you have a IP address behind a
firewall. You may be getting
            data, but it might not be able to deduce where it is.
     3c) You may be getting your IP address from a cellullar base
station on the other side of the border.
4) The software is depending on a map from IP addresses to geographical
location.
      4a) How accurate is that map?
      4b) How rapidly is it updated if some big ISP moves a block of IP
addresses from one time zone
            to another?

In principle, you need to run through that entire checklist each time
you've l...

[Milan Bouchet-Valat (nalimilan)]

Le mercredi 05 janvier 2011 à 07:49 +0000, gpk a écrit :
> Essentially, the trade-off is a minor convenience 99.999% of the time,
> versus a modest (or major) disaster 0.001% of the time.
That's also exactly how I'd describe the idea of allowing all users to set
timezone without authentication on laptops. :-p

Anyway, it's not the place to discuss the choice of automated timezone
detection: go to the gnome-control-center list if you really to want to
argue this.

[gpk (gpk-kochanski)]

On 05/01/11 12:34, Milan Bouchet-Valat wrote:
> Le mercredi 05 janvier 2011 à 07:49 +0000, gpk a écrit :
>> Essentially, the trade-off is a minor convenience 99.999% of the time,
>> versus a modest (or major) disaster 0.001% of the time.
> That's also exactly how I'd describe the idea of allowing all users to set
> timezone without authentication on laptops. :-p

Well, then the easy answer is to do neither.

> Anyway, it's not the place to discuss the choice of automated timezone
> detection: go to the gnome-control-center list if you really to want to
> argue this.
>
No thanks. People love to add features. You can't stop someone who
has a pet idea that they haven't
thought through very carefully. They just produce self-justifying
reasons and don't listen.

Me, I'll just turn it off when it appears. Then in 5 years, either it
will become useful,
or the distros will cause it to default to "off".

[Milan Bouchet-Valat (nalimilan)]

Le mercredi 05 janvier 2011 à 20:28 +0000, gpk a écrit :
> No thanks. People love to add features. You can't stop someone who
> has a pet idea that they haven't
> thought through very carefully. They just produce self-justifying
> reasons and don't listen.
That's an idea that came from designers, so they first thought about the
practical interest of this feature before their personal interest the
developers have in implementing it.

> Me, I'll just turn it off when it appears. Then in 5 years, either it
> will become useful, or the distros will cause it to default to "off".
I'm not sure at all it will default to ON.