Re: kernel-headers rpm ?

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Richard Troy wrote:

...In the fall of 1977, if I recall correctly, I got my first PC - it was
a Commodore PET, 32Kb. (That's Personal Electronics Translator, for those

<snip>
I got a late start, born to late I guess. It ws '86 and the C+4 for me.
And I still remember lots of 6502 code.

I should briefly add that mine has always been a role of de-mystifying,
information sharing and enabling to the benefit of non-wizards or would-be
wizards such that I have long held a keen interest in the concerns of the
newly indoctrinated, the newbies, if you will, of our craft.

I to continue to do that. I use linux because I want to understand the system, not have Gate's ware hold my hand constantly. I try to inform people why, not just give them a step-by-styep sequence to follow. Learning the step for a given situation is great for that situation, but doesn't help with another situation. Learning why those steps are done, allow that knowledge to be applied later to other situations. This is true of most things in life, be it computers, cars, or physics.

met, the more likely they are to use the system in question. Thus, it
directly follows that if in fact Riku's suggestion eases the path to
utility, it should be supported.

Still a proper solution is important. Adding another copy of the kernel headers somplace else won't solve the problem.

FOR YOUR ENEMY, MICROSOFT! ...That is to say, each of the divergent lines
of Linux are a curse on the whole effort. They need to remain as common as
possible where it counts most: kernel compatibility and installation
strategy.

So we should have a single version of linux, run by a corporation that dictates the level of compatibility everyone wants?
The diversity is one of the reasons I like linux.

Note that it's _incompatibility_ that kills things. A system that can
tolerate a dozen different ways to install things is OK, so long as
"the right thing," whatever that is, is done every time.... And that's
kinda hard to do with the Unix/Linux view, at least as it seems to me
today.

So who decides what "the right thing "is? Red Hat wants config files in one place, and Suse in another while the developer put them somewhere else. Who's right? Who is responsible for the installers know all those different places?

Look at an example: Windows lets you install a new network driver by going
to the network control panel, selecting the adaptor panel and hitting
"new", or, if you wish, you can run the software's setup.exe, which might
run "installsheild" or any of a half-dozen other installation middleware
solutions. Or, you can put a card in, boot the system and have it prompt
you! "Diversity" you might say - they all work in harmony. The user can
(since, what, W98?) delete any software package easily, too.

All of which change the same, M$ dictated configuration information. M$ decided where the driver files go, and what tells the system to use them. All the different methods modify the same files. In linux, each distributor decided where they wanted the files. So now the installer have to know many places to put the files instead of one. It's an apples to oranges comparison, 1 OS (M$) compared to many OSes that happen to use similar parts.

Think automobiles again. Take Ford. Many different modeles share parts, but each has differences too. Know where everything is inside the Tarus, doesn't mean you can work on a Mustang. Now, look at the GM lines. They use very similar parts, and accomplish the same task, but in slightly different ways. But the devil's in the details once again. Sure a Mechanic can use experience and figure out most of the differences, but imaging writing a program to do it. That what is being asked for. Or for Ford and GM to agree to make everything except the outer shell the same.

The different distributions came about because somone thought they could do something better. Like Mandrake. It started out as Red Hat with beter KDE support. They have since diverged, each doing what they though was best. You coud even do they same.

When Mr Dodds remarks, "Windows is definately not as easy as you make it
sound", he's really missing the mark by a wide margin. The first point is
that there are by now a lot of Windows versions out there and some are
much better than others, and secondly, yes, Windows really is that much

The same is true of linux. Red Hat 8.0 is much improved of 5.x or 6.x in many ways.

Worse than thinking Windows is still as bad as it was in 1995, we get
comments like this, "Then she didn't READ the requirements before she
downloaded the file." Well hell, I'm what you might call one of the most
experienced people you'll ever find in computer science and _I_ don't want
to have to read the GDamed file either! The Goddamed thing should check
for whatever disaster might befall me and PREVENT IT on my behalf and then

OK color me stupid. I actually read the documentation. Before I download or by something I read the requirements, and the instriuctions. I also read the manual for my DVD-LD combo, thmy DVD-VCR combo, the TV, the satellite reciver, and the reciever/amp/decoder in my living room.

maybe TELL ME what and why and suggest a remedy! READMEFIRST files are for
lazy programmers and/or thoughtless managers who are either too stupid or
don't know any better, or for people who do things for themselves - they
can be forgiven, I suppose. -sigh-

Then so are the "requirements" listed on the side of any comercial software package. If it says Windows XP is required and refuses to install on you Windows 2000 system whose fault is it?

All that, or, _my_ perceptions could be just some dream I had one evening.
You'll have to figure that one out yourself. -shrug-

All this remind me of my first VCR. It was a top loading one, with a wired remote. It wasn't "cable ready", it had like 10 channels, each of which had to be tuned to the desired channel, be it VHF low(2-6), VHF high(7-13), or UHF (14+). To program it for timmed recording, you had to use buttons on the front panel. My mother never could program it. Now VCRs are cable ready, will automatically detect what channels are available, and use menus for programming. But people still have trouble programming them (my mother included). VCR+ was an attempt to fix that, you look at the schedule and enter the code for the show you want, and it knows the date, time, and channel. But it only works if you have the codes, and don't use satellite TV.

As I mentioned, the kernel has has the option to load drivers, but the distibutions choose not to use that option for support reasons. They also choose not to support add on drivers.

Perhaps you, Riku, and Florin should start you own distribution. Disable the module versioning and offer support for 3rd part drivers like VMware, NVIDIA, and such. Let me know how it goes. My guess is support issues will quickly change you mind. It has for most others. Else you'll have to charge a lot for it, like M$, Apple, Sun, IBM, and HP do.

-Thomas




_______________________________________________
Redhat-devel-list mailing list
Redhat-devel-list@redhat.com
https://listman.redhat.com/mailman/listinfo/redhat-devel-list

[Index of Archives]     [Kernel Newbies]     [Red Hat General]     [Fedora]     [Red Hat Install]     [Linux Kernel Development]     [Yosemite News]

  Powered by Linux