Re: Linux, the final decision
From: The Ghost In The Machine (ewill_at_sirius.athghost7038suus.net)
Date: 02/02/04
- Previous message: Pureheart: "Re: How to truncate log files."
- In reply to: Kadaitcha Man: "Re: Linux, the final decision"
- Next in thread: Kadaitcha Man: "Re: Linux, the final decision"
- Reply: Kadaitcha Man: "Re: Linux, the final decision"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: Mon, 02 Feb 2004 01:00:20 GMT
In comp.os.linux.advocacy, Kadaitcha Man
<nospam@rainx.cjb.net>
wrote
on Sat, 31 Jan 2004 18:34:10 -0800 (PST)
<NhcdqXtqaIP9E5BEF635zpeoFjUbNZNf@internet.hapra.at>:
> The Ghost In The Machine wrote:
>> In comp.os.linux.advocacy, Kadaitcha Man
>> <nospam@rainx.cjb.net>
>> wrote
>> on Fri, 30 Jan 2004 22:29:56 -0800 (PST)
>> <klPHUr2xDNtv7E285182pk77cDrx1G8h@nsa.silentsky.com>:
>>
>>>The Ghost In The Machine wrote:
>>>
>>>>In comp.os.linux.advocacy, Kadaitcha Man
>>>
>>>>Actually, a GUI restart is worse for the workstation sitter.
>>>>With a reboot, it's simple: CTRL-ALT-DEL and that's it.
>>>>A GUI restart can be initiated on Linux by one of the following.
>>>>
>>>>[1] The X server crashes on its own. This is somewhat akin to
>>>> a Windows spontaneous reboot except that the rest of the
>>>> server doesn't notice. Under normal circumstances the user
>>>> will just see his screen vanish, to be replaced by an
>>>> authentication widget, unless he used 'startx' in which case
>>>> he might get a message. The apps will die. If the session
>>>> manager's smart enough they might restart when he next logs in
>>>> but I've discovered that, at least for xterms, they're not too
>>>> bright about it; if you have a chance try setting up a
>>>> multiworkspace Gnome/sawfish desktop, fire up a bunch of
>>>> of Xterms, then log out and log back in. The results are amusing
>>>> to a person of a certain mindbent but hardly useful.
>>>>
>>>> For its part Galeon fares somewhat better but one has to be careful
>>>> when restarting it to start it from the same workspace/window where
>>>> it died or its session was saved; otherwise it shifts.
>>>
>>>SHOCK! HORROR! And to listen to the vast majority of the cola posters,
>>>that is an out and out lie that simply does not happen in linux.
>>
>>
>> Well, it may depend on what one defines as "Linux"; kernel crashes
>> are rather rare.
>
> Indeed, but being as I'm a *** troll, I won't let that distinction
> get in the way of healthy debate.
OK.
>
>>>BWAHAHAHAHAH!!! I tried it whilst reading that. I lost the lot - no
>>>prompts to save my work, nothing -- POOF! BWAHAHAHAHAHA!! Thank you for
>>>that *eg*
>
> [snip]
>
>> That most editors don't bother indicates that we have a challenge
>> in the Linux community -- a challenge that should be met if
>> we ever are going to compete with a buggy but omnipresent
>> competitor.
>
> Indeed, but windows does go POOF! without warning. That cannot be
> denied. Of course, the scenario we are discussing is avoidable.
>
>> (HINT HINT, guys. I know you're out there. :-) )
>
> LOL
>
> [snip]
>
>>>Yes, I understood most of that already.
>>
>>
>> Good. :-)
>
> Well, if I'm going to succeed at being a linux troll instead of a
> wintroll, I'm going to have to know my stuff.
>
> [snip]
>
>> Welcome to Fundamental Philosophical Incompatibility Land. :-)
>>
>> As you might have noticed, there are (at least) two general
>> possibilities for design that are possible.
>>
>> [1] The system pretends the crash didn't happen and waits for
>> the remote box to restart, then continues where it left off.
>>
>> [2] The system notifies the app that something went wrong, which
>> the app then decides to notify the user, confusing him
>> to some extent, then the app terminates.
>>
>> With NFS, one can get either option. :-) (I'd have to look to see
>> how Windows deals with an analogous situation. One advantage
>> of course is that Windows drivers have access to the GUI, which
>> might be of some assistance for notification if a remote node
>> decides to go down.)
>
> That drivers have access to the gui is not a hard and fast rule. See the
> common object model.
No, but many drivers presumably do. Ideally, they all would, if
only through some sort of interface that would prompt the user
and ask "Hey, I'm having a problem here; what should I do?"
Or at least indicate through the application instigating an
action that there's an issue.
>
>>>I didn't mean first time at that age, but you're right. I was going to
>>>mention that I cannot give linux to my kids because the software they
>>>use, and you mention Jumpstart, which the 3yr old uses here, is simply
>>>not available. The older kids are into very heavy graphics games and
>>>they will not play under XO.
>>
>>
>> Correct, although it may depend on the engine; both Quake and
>> Unreal engines can run under Linux. (Of course it depends on
>> the mods the gamedevelopers did on the games, and there are
>> many many other game engines -- Build comes to mind [Duke Nukem,
>> Max Payne, etc.] -- which haven't been ported yet.)
>
> Hopefully the developers will get on the bandwagon when XOO/Wine et al
> are stable enough and start to build apps that support that environment.
Or simply port from Windows to Linux using XOO/Wine. In fact,
they may have to do absolutely *nothing*, in an ideal sense;
Unreal (an old copy) for instance works fine under Linux *as is*.
(Linux + Wine, that is.)
The main problem appears to be the installer.
>
> [snip]
>
>> Linux does *not* work out of the box. You take a Linux kernel
>> (any kernel), compressed, uncompiled, and make it work without
>> any utilities.
>
> There's the kernel distinction again. So, what is the correct term for
> what I'm talking about? Is the whole package, including the GUIs, the OS
> or not? Is it something else? What is the collective term for all the
> things I have installed that are currently called Xandros?
A distribution. Or one can refer to a specific distro:
Debian, RedHat, Fedora, Gentoo, Xandros, Mandrake, Thiz (or ThizLinux),
etc.
There's a lot of them out there; www.distrowatch.com may be
of some interest, listing over a hundred.
(And people wonder why Linux appears fragmented, disjointed,
and confused.)
>
> [snip]
>
>> I'd have to look up the details.)
>
> No need.
>
>> And Windows has exactly the same problem, although Windows
>> sidesteps it through the Magic Of Microsoft Marketing(tm).
>> Simply put, no one has to install Windows because the
>> vendors of prebuilt x86 machines have done all of the
>> schlogging for you; just plug it in and go.
>
> Not true in most cases. You have over-generalised.
That I have, but the naive user will want to purchase
something he can *use*, not something he has to fuss with
and install 2,714 drivers, utilities, and patches to get
it to work the way he -- or at least the way the vendor --
wants it.
Hence, bundling.
There are Linux distros out there selling preinstalled
equipment (ThizLinux comes to mind, although I forget
who's selling it; I've seen it in Fry's Electronics, though,
on equipment running for display), but they're few.
Of course, you're a power user, who will probably customize
the hell out of a box. (*I* would, too. :-) )
>
>> If one does have to install Windows from scratch (a
>> full XP install, that is), one has a little work ahead,
>> judging from the posts I've seen thus far, with perhaps
>> three or four reboots -- including the one that starts
>> reading from the CD. (I suspect most of these are't all
>> that necessary, though.)
>
> I can get my machine installed with XP, all upto date drivers
> and SP1 in two reboots.
I'm assuming that's not including the power-on slap-in-CD one?
That's sort of a reboot. If I don't count that most Linux
distros can do it in one, and it sounds like XP can, too,
under many conditions.
> It's a matter of knowing WTF you're doing.
That *always* helps. :-)
> I could do it in one, but it's risky because the SP might have
> unmet driver dependencies and could be waiting for the drivers to
> be updated by the reboot. But windows actually tells you that you
> cannot proceed without a reboot when the dependency issue occurs.
>
> [major snippage]
>
>>>>I take it you've read the Deuglification Howto?
>>>
>>>The what? oh - de-uglification. No. But I just found it
>>>on google. Thank you. At least I'm not the only one
>>>complaining about linux' font rendering.
>>
>>
>> You're hardly the first. :-)
>
> The problem ended up being mozilla. TTfont support is good.
> It worked out of the box in Xandros. All I had to do was
> install the fonts. Even OpenOffice picked them up to use.
> It turns out mozilla is a piece of ***. It wants to control
> the fonts and refuses to recognise your system settings.
> I'm a happy linux*** at the moment.
Should we worry? :-)
>
>>>>If you're a masochist you can of course design your own fonts... :-)
>>>>I might consider such at some point (for other reasons).
>>>
>>>I used to do that many years ago as a postscript programmer. Oh joy!
>>
>> Like I said...if you're a masochist. :-)
>
> I installed linux, didn't I?
That you did, confusing the issue quite a bit. :-)
>
>> Personally, I'd try to do something with BDF, were I all that
>> interested in the problem. Of course, BDF is ancient tech
>> nowadays; I probably should find info on TrueType.
>
> It's worth the effort.
Indeed, especially since I have the Three Magic Books on PostScript.
(Red, green, blue). Of course, actually finding them in my
demesnes might be an interesting problem.
(Organized? What's that?)
[.sigsnip]
-- #191, ewill3@earthlink.net It's still legal to go .sigless.
- Previous message: Pureheart: "Re: How to truncate log files."
- In reply to: Kadaitcha Man: "Re: Linux, the final decision"
- Next in thread: Kadaitcha Man: "Re: Linux, the final decision"
- Reply: Kadaitcha Man: "Re: Linux, the final decision"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]