[mythtv-users] OT: A major difference between Widows and Linux

[Mike Perkins]

On 12/05/16 14:58, Stephen P. Villano wrote:
>
>
> On 5/12/16 8:41 AM, Damian wrote:
>> On 12/05/2016 11:02, Simon Hobson wrote:
>>> Damian <myth at surr.co.uk> wrote:
>>>
>>>> A fundamental difference between Windows and Linux
>>> Couldn't you have raised a less contentious topic like politics or
>>> religion :D
>>>
>>>> it seems to me, is ...
>>>> In Linux, if pretty much anything goes wrong, the user needs to
>>>> enter some commands into the terminal in order to get to the bottom
>>>> of the problem.
>>>> In Windows, no matter what goes wrong, there is almost never a
>>>> 'need' to go to the command line.
>>> Go back a year or three and it's more like :
>>> In Linux, everything is possible (even if a bit arcane and tricky to
>>> configure), while in Windows the simple is easy and the difficult is
>>> impossible (ie if there's no GUI button to click).
>>> In Linux, when something goes wrong, it'll be isolated to something
>>> that's visible and fixable, while in Windows it's often easiest to
>>> just re-install the OS.
>>>
>>> On both, times they are a changin'
>>>
>>> With SystemD, Linux is rapidly moving over to the Windows ground of
>>> big binary blobs of obfuscated functionality and many unrelated
>>> functions rolled up into the one black box - with restricted
>>> configurability and obfuscated logs that need a special program to
>>> access. Windows on the other hand, is moving the other way, with
>>> PowerShell and even Bash now !
>>>
>>> As someone else put it, debugging with SystemD is like fault finding
>>> on electronics with nowhere to stick the voltmeter.
>>>
>>> It was often said long ago that "DOS is like Unix with the useful
>>> stuff taken out" - with PowerShell and Bash, they're finally fixing
>>> that. A few decades late, but ...
>> Thanks everyone,
>>
>> I'll not respond to any individual point as I know the topic could
>> snowball and be an annoying distraction from MythTV.
>>
>> Thanks for those of you who responded. Very interesting to read your
>> thoughts.
>>
>> Cheers,
>> Damian
> I'm still trying to wrap my head around the thinking that one requires
> the GUI to administer windows systems. I've rarely used the GUI to work
> on and fix windows systems, I'd psexec a command prompt, examine and fix
> what was broken, exit and call it a day. I had scripts set to psexec in
> as well, they'd run in the background, fix a common thing that was
> broken, exit and call it a day (wsus breaking comes to mind under XP Pro).
> Linux doesn't have a GUI by default and design, the GUI is added by the
> distro as a default GUI, which can be changed to an assortment of window
> managers on top of the X windows system or other interface (kiosk
> machines come to mind, I used one kiosk to check in for a lab test
> yesterday, a whopping single button to push to start the process),
> execute program, accomplish the goal of the task.
> Windows originated in a DOS world, initially, networking didn't even
> exist and networking was shoved under the hood, barely fitting
> initially. Adding high memory helped, as the drivers initially left
> little usable memory available, mouse, NIC driver, etc using the first
> 640k of available memory!
> Yeah, I used to have Windows 1.0 discs around, lost them when my home
> was emptied out during a long deployment...
> Windows 95 tightened the GUI and DOS linkage, but didn't eliminate the
> bolted on effect. I won't go into the ping of death or Microsoft lying
> about its existence or worse, legal threats made by Microsoft against
> those reporting the bug... Their decision to switch to Sun Solaris
> servers, when their NT4 boxes couldn't stay operating, due to "network
> load issues" (they were being crashed by tens of thousands of scripted
> ping of death attacks). Windows 2000 finally started cementing the GUI
> on, with a finer layer of shellac on top of the varnish, covering the
> cracks.
>
> So, where are we today?
> We've had compromises of Windows hosts on our network, the compromises
> were all command line attacks, although a couple of successful
> compromises started at the CLI and ended in an RDP session for the
> convenience of the attacker. Have the command line, own the host.
> Which is precisely where we started.
>
Having written for both Windows and Linux, and before that for mainframes and for single chip 
processors like the 8080 and Z80, I'll leave the fine detail to others. I can make this simple 
distinction between Windows and Linux/Unix:

With Linux, it is assumed that everythings starts off as a process which is forked from init or a 
substitute and even if it creates a GUI the same basic underlying program functions the same way 
that all the others do. All GUI interactions for all programs are handled by the X server or an 
equivalent, which is just another user program.

With Windows, every single program starts from the assumption that it has a Main Window, even if the 
user doesn't see it. ALL interactions with each program are as the result of interrupt functions to 
that main window, dispatched by the OS scheduler. Sometimes these interrupts are handed off to child 
windows, but the 'window' is the basic unit of Windows programming.

That is the fundamental difference between Windows programming and Unix programming.

-- 

Mike Perkins