SI
SI
discoversearch

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor.  We ask that you disable ad blocking while on Silicon Investor in the best interests of our community.  If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.
Technology Stocks : How high will Microsoft fly?
MSFT 476.080.0%Dec 17 3:59 PM EST

 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
To: David Howe who wrote (62931)11/13/2001 5:27:12 AM
From: dybdahl  Read Replies (2) of 74651
 
The computer I use right now is a 450MHz Pentium III. I have no intention to upgrade. But it's true, that the number of CPU cycles required to do certain tasks is the same on Windows and Linux.

When people talk about that slow PCs do Linux fine, it's because many servers don't need extremely high CPU clock cycles. One of our customers still has a Pentium 100MHz server with 16MB RAM running. It's a webserver, DNS-server, file and printer server, on a remote location with three client PCs attached, and there are absolutely no plans to replace it.

You can also use Linux for different purposes than running applications. If you got yourself a 486-66MHz, it will do as an X-Windows thin client, which would make it possible to get than 2GHz application speed from the server, just with a slower screen update.

The smallest instalation I ever made with Linux was one diskette. That was fairly easy. The more applications you want, the more harddisk it takes. The biggest installation you can make directory from a Red Hat Linux 7.2, is 3 Gigabytes. But then you also have installed chinese translations for all your software. If one guy has said 9GByte, then he probably installed SuSE Linux - they include so much software, you cannot believe it. I think they distribute 7 CDs.

The guy you know that tries "latest versions" is experimenting with his Linux. Having the source code for all basic software, you don't need to be binary backwards compatible. KDE 1 and KDE 2 libraries are not binary compatible, and KDE 3 won't be binary backwards compatible either. Instead, you install a compatibility library to support old programs, and all programs that want to use the new libraries, have to be recompiled.

The benefit of this approach is, that you can get rid of old stuff in the APIs, you can compile to a newer processor etc. There is also the big discussion of C++ application binary interfaces (ABI). ANSI has standardized the C++ ABI, which means that shared objects (equivalent to DLLs) can now expose C++ classes, instead of just C functions as in the early days. Red Hat made a controversial decision when it uses a compiler for Red Hat 7.0, that tried to implement the ANSI ABI, but wasn't finished, yet (gcc 2.96RH). This broke the compatibility with previous and later versions of that compiler, but the benefit was that Red Hat could include applications i Red Hat 7.0, that exploited the new C++ ABI, while using the same compiler for the rest of the applications. Red Hat has decided to use the same compiler for Red Hat 7.1 and Red Hat 7.2, in order to maintain binary compatibility. Red Hat 8.0 will include a new compiler, gcc 3.x, that will conform to the ANSI standard, and thereby break compatibility. Other distributions do not use gcc 2.96RH and binaries for those distributions are therefore not compatible with Red Hat 7.x.

The reasons why it was so controversal is that Red Hat made many things much more complicated than it needed to be, and the people who primarily will experience problems, are those who try out new alfa and beta-version of different kinds of software. He is very right in his criticism.

But then - how many people are unsatisfied with WordPad in Windows XP and would like to download a later, untested version? Not many. The same applies to Red Hat Linux. Not many people download later versions of the installed software.

When you talk about stability on Linux and Windows, people normally talk about different things:

- The Linux kernel is definitely more stable than the Windows kernel. If you need to reboot a Linux computer, your hardware has a malfunction. It isn't the Linux kernel that makes you reboot.
- Some X-Windows drivers are unstable. Don't use them. I don't. My NVIDIA X-Windows on my primary Linux has never crashed. The NVIDIA graphics drivers often crash my Windows 2000.
- Unstable programs on Windows often cannot be minimized or closed. On Linux that is less a problem.
- Windows installations typically need to be supported during their lifetime. Linux installations can often run unserviced from their installation until their hardware is replaced.

It is true that there are more unstable Linux software around than Windows software. The reason typically is that software is released before it is 100% bugfree. If you buy a Red Hat 6.0 and use the GUI programs in that, you will see quite a lot of it, but it improves all the time, and in my Red Hat 7.2, I only have found one unstable program (Brahms), that works until you close it, then it crashes.

These kinds of errors are less than ever before, and also raises the requirements for when distributions want to include software. Red Hat did not include KOffice until it was stable and had a usable functionality, which is a change in politics.

When people start to use Linux, they will immediately find all the things that are worse and they will immediately miss all the functionality they know so well and cannot fint in Linux. This is very natural and almost given by nature.

Most people, I know, didn't switch to Linux until they found functionality that didn't exist in Windows. When my old mother asked me about using Linux, I originally declined, until she mentioned a few things I had mentioned, that she wanted but couldn't get on Windows. So we made an analysis of what she needs and what she can get, and then she was sure that she wanted Linux.

The first experience she had with Red Hat 7.1 was the same as usual - she didn't know how to do this and that, but quickly learned to do most stuff. Now, a couple of months later, she has upgraded to Red Hat 7.2, and everything still looks the same but all software has improved. She likes that a lot. She is also now used to a lot of functionality that she didn't have on Windows:

- Localized, GUI secure FTP for uploading homepages to my server, which doesn't have FTP for security reasons.
- Login as another user to see her websites offline, breaking file:// links that point to her own directories.
- Login locally and from two Windows computers with X-Windows so that the grandchildren can play Linux computer games on her computer from three screens at the same time (they play more via Linux than locally on the 1.2GHz Windows 2000)
- She is absolutely virus and worm safe, without having slowing scanners and firewalls installed.
- E-mail program has encryption and signature built-in. We don't use that much, but it happens that we send passwords or messages that shouldn't be read by random 13 year old hackers.
- Backup is made every day, compressed, encrypted and uploaded to a server on the internet, 100% automatically.
- My father couldn't find his HP CD burner software, so we had to move his CD-Writer to my mother's computer, where we use KreateCD. The alternative was to use illegal copies of software or to pay for new software.
- My mother knows that her current settings in all the programs will survive for many, many years to come, surviving many versions of Red Hat Linux, and surviving many generations of PC hardware.

Personally, I have a lot of other features that I miss on Windows, the list is too long to be listed here.

Until you have understood what makes Linux one of the worlds most widely used operating systems, you haven't understood Linux.

Lars.
Report TOU ViolationShare This Post
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext