I'm not sure that's really an argument for compiling one's own kernel though? That's about 15 seconds a day saved for a fair bit of work identifying exactly which things need enabling or not in the kernel. It might work out to save some time but a few minutes a week which could be spent making a coffee or whatever doesn't seem to me like a reason to compile one's own kernel.

Curiosity can be a motivating factor also, I know I'm weird but I always have sort of fun when I discover some kernel features and read the help texts at the configuration step, I have a lot of respect to all who put the efforts on this

Talking of coffe i switch my machine on in the morning go and make a cuppa and its ready by the time my coffe is, somseems like a good deal to me and as my hardware rarely changes i use the same config file for the new kernel and the buikd is scripted so no hassle.

"I use the kernel shipped with my distribution" more often or "I compile my own kernel (official tree)" if the hardware is old and slow so clicked "Other", "Vote Now".

I Debianizing the packages using my own & share some to public.
I'm not a self-compile makeinstall individualism.
I've fun of turn lead into gold on Debian Jessie.

Compile is a gathering dependencies to make the software runnable.
You can't make a sword(package) without minerals(source).

When I started with SLS, I always compiled my kernel. I used a i386 w/16mb of memory. God it took a while. From there to an early version (4 CDs) of slackware. Again I compiled kernel. With Redhat-8, I both compiled and used distros kernel. I finally went back to Slackware 14.1/2/current and always compile a kernel. I make sure I have a few modules that Slack doesn't distribute with Stock Distro.

What on earth are you doing on this forum as a newbie? You must predate all of us. I've never met anyone before who had actually used SLS.

Were those early kernels monolithic? If they were, then you had to build them yourself to fit them to your hardware without making them too big. So it isn't really a fair comparison with the present modular stock kernels.

There are a number of us that started with SLS (1.2/1.3 for me). It felt VERY familar if you had used an AT&T System V release or a SunOS distribution before (but had a lot more software available).

Modules didn't exist for the kernel until after about 2.5 kernel. As I recall, that evolved from experiments with microkernel version of Linux (which had too much overhead), but my memory may be a bit faulty.

Up until the in-kernel module loader existed it was .. tricky. Kernels were configured for the drivers that were available - distribution kernels (like the ones Slackware used at the time) had all disk drivers built-in (it was what made the kernel "big"), but even then, it was still under 800K. You were expected to trim the kernel after installation as the systems only had 4 to 8 MB. My 486 only had 4MB, but could run X if there was enough swap. 8MB was much faster... custom compiles could take a couple of hours (less after you find the "-j 2" option worked - though the disk got a lot busier). I seem to remember getting the kernel down to about 400K (but didn't look all that closely). I didn't compile with X running though... Too much swap activity. Backups to ftape (floppy tape) were a pain - it would take all day to backup a 50MB partition.

Oh - one more tidbit. The install system was two floppies. The kernel was on one - and would pause after initial boot telling you to put the root floppy in. Slackware had two kernel floppies - one for IDE controllers, and another if you had SCSI. A combined floppy was bigger than 1MB making it unusable for an install boot.

What? I didn't realize this was a newbie forum... I'm 57 years old and, yes, I apparently didn't read the docs closely enough about LQ. He he he (cough cough). Well I think I had better read the other posts, Hazel (right name?). I image they will be interesting.

No, this is a general forum. But you are labelled as a newbie, which I find rather comical in the circumstances. Of course you are not responsible for the label, I realise that now. The forum software does it, depending on how many posts you have made.

Oh right. I see the little "label" under my handle.
Well, now I have to remember the question. Compiling own kernel I believe. You know, I don't mind the compiling and I have some old hardware that I'm to cheap the replace. Like my Radeon Graphics card. Great card. Not quite up to speed like some Gforce Cards. The builtins are minimal at best. I do like the Intel 810 that was in an old Dell but still put in the Radeon. Anyway, with Linux, I will start a kernel compile and just ALT-Fn to another tty term (CTRL-ALT-Fn while in X; You know, I didn't find out about that <CTRL>+<ALT>+Fn until a couple of years ago. I just either started an xterm or usually just bomb out of X.

LONG live Linux; Down with Windows!!!

I am probably labelled as a newbie as well.
I have never had to compile any drivers for Linux in my 45 years of using computers. I have had to get a few? packages changed either to fix errors or to support functions that were new or omitted from the original.

I have done enough support in my working life; I just want to sit and enjoy using my PCs.
It doesn't matter much whether it is Linux or Windows as far as I'm concerned. They are both just tools to be used.

I always compile my own. The reason is that I am upgrading to a later kernel. On one of my older laptops it is necessary to get the dynamic frequency scaling to work due to a broken BIOS. It has to use the deprecated speedstep-centrino module to work. So compiling a kernel is needed to get that to work. Maybe someday I'll figure out how to write a proper DSDT that actually works and use that for an override.

After everything's works, I update the kernel.

Started playin' with the kernel back with Slackware 4....