Welcome, Guest. Please login or register.

Author Topic: Linux: The way forward?  (Read 820 times)

Description:

0 Members and 2 Guests are viewing this topic.

Offline stefcep2Topic starter

  • Hero Member
  • *****
  • Join Date: Sep 2007
  • Posts: 1467
    • Show only replies by stefcep2
Linux: The way forward?
« on: February 02, 2008, 09:56:50 AM »
da900 wrote: "
However, as far as you Linux comments, I'm sorry to have to say the below (before I start, I've been using Linux since the day I was told I could play Doom over ethernet, that's around 1995, have followed kernel development, done little kernel programming, and lots in various open source projects from XMMS to X11, etc. Aka, I'm no newbie to Linux):

1) You clearly have no concept of how scalable sytems work. Linux and Unix is the embodiment of a scalable architecture (due to their history and birthplace in highly-networked and distribute environments from AT&T/Bell Labs to academia, DARPA projects, etc), which is an architecture vastly more complicated and compartamentalized in order to overcome the problems with closed systems, which would have "neater layouts" (not only at the filesystem level but also internal module/API/ABI levels) and also less updates (many updates are a necessity for highly modularized and interdependent systems, which is due to optimal usage of resources, such as libraries of code, and data. This is known as re-use or code-reuse and is a staple of less buggy software. If you're not familiar, just think of this: what are the chances of writing the same sentence on the blackboard and making more mistakes: 10 times or 1000 times? By re-using already written code, you minimize, statistically speaking, error propagation). This allows for far faster growth, something you're not aware of, as you state that multi-tasking is poor under Linux.


No it doesn't its: its a two step forward one step back result, because every time they fix two things one will break something else.  Where this is a problem is when it results in a non-bootable system, just because you installed for example a crappy 500k tetris clone (via the repos) that installs a 100k library update to your gui which doesn't work unless you have installed 40 MB of gui updates, which don't get installed automatically, unless you "mark all upgrades", which if you do tells you that have 550 MB of other updates...I think I'll give tetris a miss then..



You have no concept of the pre-emptive kernel work, low-latency kernel work, RT Linux work, etc. Yes, those are not things you click and get, that's only on a Mac. These you have to either compile or get the right distribution for,


And that will appeal to the average user?  Sure if you want to keep linux in the closet so that only people who can do this can have a responsive OS, then thats fine, thats where it will stay, in the closet. I was referring to the current momentum that Linux has as a desktop replacement for Win or OSX.  Ever heard of Executive, a shareware Amiga task scheduler?  It can do all of the of the stuff Linux is trying to do, and all the average Joe has to do is double click" install executive" and he can chose from more than half a dozen task schedulers depending on his user-profile.  Oh and you don't need to recompile anything and you don't even have to reboot just because you changed the type of scheduler you use, and its was written 10 years ago!!  Linux is still a multitasking elephant, its noticeably worse than XP, and this running on on a 3ghx duel-core processor.


but there are few modern OSs (sorry, AmigaOS doesn't count given it's current featureset) other than RT OSs which guarantee sub-millisecond latency response.



I was describing my experiences with Linux which incidentally are on the more positive end of the scale compared to what people on the forums go through.  I DO NOT need to know any of the technicalities about "scalable" systems to confirm that the USER EXPERIENCE needs alot of work with Linux, anymore than I need to know how the computer management system in my care works so that I can drive it.


2) Your comments are clearly biased against Linux due to your tainted experience.

No bias whatsoever.  I tell it like it is.  As I said my experiences are on the more positive end of the scale, many users-experiences are MUCH worse than mine.



But you've missed the point: Linux, being an open system allows people to do whatever they want. You went asking the PCLinux or whatever the name was, people to cut down on their methodology (frequent updates), but that's obviously not their "mantra", so their best answer was what you got: you want to look at another Linux that doesn't believe in that.



Yeah thats fine if it weren't for the fact that 80% of the support questions on their forums occurred BECAUSE of this mantra.  Its also a problem when you have multiple users saying "OK if you are going to be dogmatic about it and not listen to the users then at least make it a "sticky" in the forums for newbies to read:  Say for example "If you don't want to constantly upgrade EVERYTHING that you have installed with our distro" then DO NOT install anything new, ever."  But you know what, they refused to do this, because then no new user will ever install that distro on their hard drive.  



Although not the only one, and not recommended for newbies, I often use Slackware which is much "slower" at updates, and serves me better.


And thats why we have a gazillion distros at there, all re-inventing the wheel.  Its "Well if you don't like it go away and try another distro or make your own. Ad-Infinitum"


The whole point is: Linux isn't one entity, and neither does it owe you anything just because it's free.


Yes I've read this elsewhere.  I am not a Windows user who demands that Linux issues be fixed to my liking, for nothing


What it is, is a system which you can modify (if you so wish badly enough, and if not try to find others who feel the same way and wish it more badly than you, so that they have actually taken the time and put the effort to do so), so that it does what you want it.


The whole problem is that people stuff around creating new distros that address largely minor or irrelevant issues whilst the important stuff is left to far too few.  Is it too much to ask that I can drag a window without my sound breaking up when playing an mp3 on a Athlon X2 6000+?  Oops, sorry they fixed this now, welcome to 1992..Is it too much to ask to have the worlds most popular distro Ubuntu have a properly working dial-up tool?

3) I also don't like many aspects of Linux and Unix, but I understand that complex and advanced systems aren't built overnight and because of that they carry a lot of history. In your particular case, "/etc" is confusing to you (much like the SYS:L would be an Amiga newbie), since you've no concept of Unix's history and the "baggage" that Linux got due to its ideological roots (not talking about code here) being in the Unix world. Apple for example, despite deriving part of their Mac OS X from FreeBSD, have chosen to change the historical conventions. Linux folks don't see it necessary. Don't like it? Goes back to the previous point, make your own distro and change the naming (but be warned: it'll be a pain, because you're going against historical trends, once again, which have been deemed by those working on Linux as HELPFUL to them, unlike what you think, and thus fully ingrained into their Linux work). If that sounds like too much pain, then Linux or your current distros aren't for you. Now, what people mean when they say "Linux is the future", is not that it's perfect, but that it's a platform embodying the perfect mentality for massive-scale parallel software developement. Aka, it's Open Source.


I am fortunate enough to have experienced OS's which simply "do it better".  Consequently, I, like many users know where the problems are.  Most Linux "guru's" are so caught up in their world that they do not take negative feedback at all well especially when its fully justified.  Your system multitasks badly? hell the kernal developer can't reproduce it on his quad-core 3 ghz 4 Gig Ram system, must be something you're doing wrong...Again thats OK but don't in the same breath claim that its a user-friendly, interactive beast of an alternative to what else is out there..Its NOT.

And to end on a positive note: I'm certain there'll be a Linux version/distro in the future that foregoes the old Unix baggage in many ways (X11 included, and I for one would like to see something like Apple's PDF based Quartz-like architecture for GUI - DRI doesn't suck completely, but it's not really such a good architecture, nor very clean - too long a discussion for now) and evolves into the future of Linux.

For this to happen in a reasonable time frame, there needs to be a co-ordinated centralized effort, which goes against the grain of what many Linux developers believe Linux should be about.  Ofcourse some would say its already happened, but you have to pay for it and you need propriatory HW:  Its called OSX.


Sorry for the off-topic guys. I saw a box of soap and stepped on it... Feel free to ignore it, but don't be fooled into thinking that ignorance is bliss. It's mental poverty.
   Reply
 

Offline Hammer

  • Hero Member
  • *****
  • Join Date: Mar 2002
  • Posts: 1996
  • Country: 00
    • Show only replies by Hammer
Re: Linux: The way forward?
« Reply #1 on: February 02, 2008, 10:00:33 AM »
Example of Windows platform scaling via clustered computing methods are Windows PCs infected with Zombie/BotNets viruses.  
Amiga 1200 PiStorm32-Emu68-RPI 4B 4GB.
Ryzen 9 7900X, DDR5-6000 64 GB, RTX 4080 16 GB PC.
 

Offline stefcep2Topic starter

  • Hero Member
  • *****
  • Join Date: Sep 2007
  • Posts: 1467
    • Show only replies by stefcep2
Re: Linux: The way forward?
« Reply #2 on: February 02, 2008, 10:08:47 AM »
So this "scaling" business hales from server-type set ups.  If so Linux is good at this, but the desktop is a different matter.