New Group Calls For Boycotting Systemd

[energyman]

It's amazing how much time people can spend on bashing something that they don't like or understand.

For me it's simple - I don't like german cars, I don't own any and I don't give a shit about them. I don't waste my life on whining on german motorization

unlike systemd, German cars aren't forced upon you.

How would you feel, if you HAD to drive a German car and every other car would be banned and/or blocked from using roads and parking spaces? Hm?

[robclark]

unlike systemd, German cars aren't forced upon you.

How would you feel, if you HAD to drive a German car and every other car would be banned and/or blocked from using roads and parking spaces? Hm?

except that is not what is happening here.. to continue your analogy, non-german cars are not banned/blocked/etc. You still have all the same software freedoms you had before. It is all open src, you are in no way blocked from doing whatever you want. But non german cars are becoming less popular, so you might have to search harder to find a non german car dealer to buy a new non-german car. Or you might have to maintain your old non-german car if you don't feel like putting in the effort to obtain a new non-german car.

To flip this around, if 90% of drivers preferred non german cars, would you block them to appease the 10%? *That* would be the injustice that people should scream about.

[interested]

No if there's an even amount then exactly half of them will be above average and exactly half of them will be below average, if an odd amount then 1 will be average, half of the remaining above average, and the other half below average, because that's how averages work.

It was a allusion to the fictional place Lake Wobegon, from the radio show "A Prairie Home Companion", where "all the children are above average."

http://en.wikipedia.org/wiki/Lake_Wobegon

I could say that for the set of all Mensa members half of them are below average, while correct it would be an utterly idiotic thing to say, because it literally tells me nothing about them, while implying that they're stupid, which by definition they're not.

Yes, you get it. It is all about the selected data subset. In this case "all programmers" includes people with even rudimentary Bash/VB skills, so the selected data subset is quite wide.

Personally I think plotting highly complex skills like IQ (for which no real definition exist) or programming skills on a Bell curve is a dubious thing to do since it reduces something really complex into a single number.

Quote Originally Posted by interested:
But as I explained, there is nothing wrong with being below average in a problem domain, we all are in different domains.

Enough with this game, and pussyfooting about what you want to say. Just say it: "Sys Admins generally suck at programming", there I said it for you. Was that so hard to drop the pretension and say what you mean?

It isn't pussy footing, it is called a nuanced insight into human behaviour. It is exactly the opposite as seeing things in black and white, so that if people have below average programming skill, they must therefore "suck" at it.
Even a little Bash programming skill for making simple task automated, is a great skill to have. And knowing your limits with whatever you do, will keep you out of trouble too. Many SA's are competent programmers for what they do, but often they spend most of their time on something else, so often they don't have the same skills and experience as full time programmers.

So let me briefly return to the original topic and the context; SysVinit may force SA's to debug and make code (which may not be their main skill), in a programming language (Shell/usually Bash) they may not use very often (some use perl for almost everything etc). This is a flaw in the SysVinit design and a violation of the KISS principle.

[Isedonde]

No if there's an even amount then exactly half of them will be above average and exactly half of them will be below average, if an odd amount then 1 will be average, half of the remaining above average, and the other half below average, [..]

That's not true. If you're talking about the median type of average, it's sometimes true. Usually, "average" refers to the arithmetic mean though.

[2,10,10,10,10]

Arithmetic mean "average": 42/5 = 8.4
1 is below average, 4 are above average

Median "average": 10
1 is below average, 4 are average


It's true only for special cases, like this median average:
[2,3,4,5,6]
Median "average": 4
2 are below average, 1 is average, 2 are above average

[?] because that's how averages work.

No.

[interested]

You two should just stop using Linux and go with Windows instead. You really do not have to touch any of the scripts when it is too complicated for you. Clearly, the freedom given with UNIX and Linux is wasted on you.

The point is exactly that badly designed init systems like SysVinit may _force_ people to deal with extremely complex, badly written spaghetti code.

This is why e.g. the Debian Technical Committee all preferred any other init-system than SysVinit for future Debain releases. Even members highly sceptical about systemd, preferred it as an alternative to SysVinit.
SysVinit scripts are basically seen as hard to maintain and error prone, and it doesn't help that they often differ from distro to distro, so people can't even collaborate on them.

Executable config files where code and program settings are mixed, is simply a bad idea, which is why SysVinit have been on the kill list for decades.

It is SysVinit's many shortcomings and the many advantages of systemd, that have steam rolled SysVinit out off existence on all major Linux distros.

[doom_Oo7]

At least with plain-text, if something gets overwritten, you can still see everything else because there's no format the file has to be in.

If you corrupt a random char in a journald log I think you can still read every non-corrupted part with

journalctl --file

.

[Luke_Wolf]

It was a allusion to the fictional place Lake Wobegon, from the radio show "A Prairie Home Companion", where "all the children are above average."

http://en.wikipedia.org/wiki/Lake_Wobegon

ah

Yes, you get it. It is all about the selected data subset. In this case "all programmers" includes people with even rudimentary Bash/VB skills, so the selected data subset is quite wide.

Personally I think plotting highly complex skills like IQ (for which no real definition exist) or programming skills on a Bell curve is a dubious thing to do since it reduces something really complex into a single number.

It isn't pussy footing, it is called a nuanced insight into human behaviour. It is exactly the opposite as seeing things in black and white, so that if people have below average programming skill, they must therefore "suck" at it.
Even a little Bash programming skill for making simple task automated, is a great skill to have. And knowing your limits with whatever you do, will keep you out of trouble too. Many SA's are competent programmers for what they do, but often they spend most of their time on something else, so often they don't have the same skills and experience as full time programmers.

While I would agree that plotting such things into single numbers is dubious, sorting people into the box of "Below Average" only leaves one or arguably 2 boxes open (Above average, and Average), which tells us exactly nothing, is setting things into black and white, and this talk of limitations only reinforces that idea.

A much more useful measure would be to state the problem domain, such as web programming, database work, desktop applications, automation, etc... and then using a graduated scale that explains how capable an individual is at a task starting with "doesn't know anything about it" and ending with "Master Programmer, can handle anything you can throw at them in this domain"

Of course then you'd need actual statistics for skill level as opposed to only being able to rely upon statistical definitions.

[erendorn]

That's not true. If you're talking about the median type of average, it's sometimes true. Usually, "average" refers to the arithmetic mean though.

Exactly.
And as programming skill is probably like a power law, or a truncated gaussian, the median is likely to be below the average (very high performers skewing the mean but not the median), so I'd say it's true that at least half of programmers are below average

[Luke_Wolf]

That's not true. If you're talking about the median type of average, it's sometimes true. Usually, "average" refers to the arithmetic mean though.

[2,10,10,10,10]

Arithmetic mean "average": 42/5 = 8.4
1 is below average, 4 are above average

Median "average": 10
1 is below average, 4 are average


It's true only for special cases, like this median average:
[2,3,4,5,6]
Median "average": 4
2 are below average, 1 is average, 2 are above average


No.

If someone tells you that half of X are below average that automatically means he's talking about the median, with the additional assumption that the area where it splits is a set of unique numbers (which is altogether reasonable when talking people). Why? Well, quite frankly unless you've got a set that grows in a linear manner then the median and then mean won't be the same thing, which is required for "half of X" to be below average. In your own example 1/5 of the numbers are below average, and 4/5ths above average

[sdack]

The point is exactly that badly designed init systems like SysVinit may _force_ people to deal with extremely complex, badly written spaghetti code.

This is why e.g. the Debian Technical Committee all preferred any other init-system than SysVinit for future Debain releases. Even members highly sceptical about systemd, preferred it as an alternative to SysVinit.
SysVinit scripts are basically seen as hard to maintain and error prone, and it doesn't help that they often differ from distro to distro, so people can't even collaborate on them.

Executable config files where code and program settings are mixed, is simply a bad idea, which is why SysVinit have been on the kill list for decades.

It is SysVinit's many shortcomings and the many advantages of systemd, that have steam rolled SysVinit out off existence on all major Linux distros.

No. I have already explained where the problems come from, but since you have already admitted you are not good at these things do I not expect you to understand them, but to take my word for it. Debian still uses much of the sysv init process to boot up and systemd is also doing far more than just replacing the init process. You cannot use your half-knowledge to declare everybody else's software as "spaghetti" code. You believe systemd is all right and conclude everything else must therefore be wrong. This however does not give you competence, nor does looking towards others give you this. You need to start looking into it to see what it does.