Lennart Formally Introduces "mkosi" Tool

[computerquip]

Because then things are bundled into a monolithic blob with dependencies on each other.
So then GNOME depends on systemd and systemd depends on all these other kinds of things.

Then instead of things getting used by its own merit, they are used because Lennart put them as a dependency of systemd.

I don't want to give Lennart any ideas, but think if here were to bundle like a mail client, irc client and web browser and media player with systemd too. It would be ridiculous.

This is not how anything works. Two things can be related to each other while being completely independent of each other. It's called loosely coupled design.
What do you actually gain from spouting this non-sense?

[caligula]

Well, if I want to pull down the source code for this tool, I don't want to pull down systemd and a thousand other projects. That's just sheer stupidity. I mean look at AOSP compile process - I downloaded several GBs worth of source code to build a 300 MB zip file.

The 300 MB zip probably contains over 1 GB of decompressed data. Not really surprising those numbers.

[caligula]

Take a look of the Armbian distribution creation shell scripts. It is very easy and fast to edit them to your liking. What redhat and Lennart is doing is reinventing the wheel and causing anxiety to users with a monolithic, buggy and slow software.

The shell script approach works for a single distro, but what about a selection of multiple distros? Much easier to depend on an infrastructure.

[ermo]

Maybe Lennart & Co.'s productivity and willingness to question the status quo and come up with workable alternatives based on clearly articulated goals and motivations tends to make people feel threatened?

I'm not sure I completely understand why. I mean, I get why people can be passionate about their tools, but perhaps the flip side of that coin is that one can become so attached to ones preferred tools that the attachment itself hinders ones ability to take a step back and look at the tools and workflows from a detached, objective and rational point of view.

Perhaps it has something to do with the reptilian part of the brain being activated when one encounters perceived threats to ones self and/or domain? I believe it is generally accepted by now that once this biological fight-or-flight response is activated, the part of the brain that handles civilized discourse and rational arguments shuts down?

YMMV as always.

P.S. I don't think it is correct to call people 'stupid' for defending their perceived domain. Biologically, the response makes perfect sense. It's just that our biology works against us whenever we are confronted with changes in the domain we have adapted to. And in IT, these changes can (and routinely do) happen quite rapidly.

[rtfazeberdee]

Maybe Lennart & Co.'s productivity and willingness to question the status quo and come up with workable alternatives based on clearly articulated goals and motivations tends to make people feel threatened?

I'm not sure I completely understand why. I mean, I get why people can be passionate about their tools, but perhaps the flip side of that coin is that one can become so attached to ones preferred tools that the attachment itself hinders ones ability to take a step back and look at the tools and workflows from a detached, objective and rational point of view.

Perhaps it has something to do with the reptilian part of the brain being activated when one encounters perceived threats to ones self and/or domain? I believe it is generally accepted by now that once this biological fight-or-flight response is activated, the part of the brain that handles civilized discourse and rational arguments shuts down?

YMMV as always.

P.S. I don't think it is correct to call people 'stupid' for defending their perceived domain. Biologically, the response makes perfect sense. It's just that our biology works against us whenever we are confronted with changes in the domain we have adapted to.

Its things like "old dogs not wanting to learn new tricks", "out of their comfort zone", "professional jealousy", "not understanding" spring to mind...

[starshipeleven]

I mean, I get why people can be passionate about their tools, but perhaps the flip side of that coin is that one can become so attached to ones preferred tools that the attachment itself hinders ones ability to take a step back and look at the tools and workflows from a detached, objective and rational point of view.

Human beings capable of meta-thinking, abstraction and relatively complex thought required to see one's mistakes are less common than they should be. It's not passion, it's tunnel vision.

Perhaps it has something to do with the reptilian part of the brain being activated when one encounters perceived threats to ones self and/or domain?

The "reptilian part of the brain" can be suppressed just like any other reflex or instinct. It requires a strong enough pre-frontal lobe though and as said above not enough people can do it properly.

Really there are 2 "command centers" for decisions, the older one tends to give you crap answers that would have been good enough for an animal dealing with a relatively simple natural environment with limited or 0 social interactions, as it simplifies waay too much and has no gray area between extremes, and the newer one that at least gives a good shot at giving you a decent answer due to higher granularity of input.

The fun in all this is that it is still the old command center that is in control of gratification and pain, so your life goals, what you actually like and don't like, is all decided by the old crappy one while the newer one is only hitching the ride. It can only orient the older one but not suppress it permanently and take the helm.
That's why psychology is a pain in the ass, you have to deal with this duality, and try to communicate with a fucking reptile.

Deceiving people instead is significantly easier, as you can usually count on most people reacting according to their reptile's biological programming.

P.S. I don't think it is correct to call people 'stupid' for defending their perceived domain. Biologically, the response makes perfect sense. It's just that our biology works against us whenever we are confronted with changes in the domain we have adapted to.

Intelligence is the ability to perform correctly outside of known situations. Lack of this ability is called stupidity.

Also, don't use biology as a moral scale. There is a large amount of stuff that is simply there because it is legacy and an event that disabled it properly didn't yet happen.

[sdack]

"Lennert Poettering", it's a name to fear like "Justin Bieber" or "The Kardashians". I could see his name turn into a meme...

[srakitnican]

Well, if I want to pull down the source code for this tool, I don't want to pull down systemd and a thousand other projects. That's just sheer stupidity. I mean look at AOSP compile process - I downloaded several GBs worth of source code to build a 300 MB zip file.

It is located as a separate project: https://github.com/systemd/mkosi

[moilami]

I wait for Lennart to formally announce development of replacement kernel

[rmoog]

I'd like to see this ported to C. Portage is written entirely in Python and it causes a few problems.
What I would NOT like to see, is having systemd as a hard dependency for this tool. This is great and for embedded device firmwares, uselessd (if it was maintained) and OpenRC are good enough.