can_suspend_to_{x}

[David Zeuthen]

On Sat, 2006-05-06 at 13:41 +0200, Holger Macht wrote:
> I don't think that they should be called Hibernate() and Suspend() in Hal
> at all. The first time I looked at gpm and hal, I was completely impartial
> in respect to any other OS, and thus, I had absolutely no idea what
> Hibernate or Suspend does. I had to look up the code to get what
> hibernating does. I really don't care about what the GUI shows to the user
> in the end, that's up to the project/desktop to decide, it's a plain
> useability issue. The Hal methods should exactly describe what they
> actually do in a technical manner. So for me, it would also be ok to call
> the HAL methods ACPISleep_S3/S4 or the like. 

This sounds really stupid in my opinion. The whole point of HAL is to
abstract the system (because the kernel is doing a really poor job in
that regard) so desktops developers have a simple API to use. Desktop
developers simply don't care if ACPI, PMU or APM is used, if they do
care then they have a bigger problem than what the methods are called.

> But not Hibernate/Suspend. A
> developer which writes a application which uses these methods knows the
> technical meaning of suspend to disk, but he may not know what hibernating
> does. The frontend can show whatever it likes.

He should know because it's precisely defined here

  http://lists.freedesktop.org/archives/hal/2006-May/005190.html

if not in the spec then at least in that mail. Someone should patch the
hal spec so it says so too (please Richard!).

> In a technical discussion, I don't like Hibernate/Suspend for suspend to
> disk/ram at all, because these are suitable inventions for the desktop
> user, but not for a developer.

As long as it has well-defined meaning it's more than OK with me to use
these terms so people don't get lost in acronyms.

    David