LWN: Comments on "printk() and KERN_CONT"

That will take some getting used to...

nix — Mon, 11 Sep 2017 11:52:28 +0000

I think it's mostly because Jon *is* in that restricted group (which is probably why he uses the style I'm familiar with).

(Also, you don't have to be *in* the inner group. That's the beauty of work in the open: you just have to read the same mailing lists!)

That will take some getting used to...

mcortese — Mon, 11 Sep 2017 07:02:59 +0000

I think you can't apply the same rules to the inner group of co-workers who have daily interactions and the broader audience of a newsletter. I've never thought of LWN as targeting the same people who is involved in the kernel development and is thus familiar with its jargon, nicknames, and abbreviations. In that restricted group it might be OK to use first names, as it's also OK to use well-known acronyms like BKL. For the rest of us, LWN provides a great service defining them on first use and applying a consistent scheme.

That will take some getting used to...

nix — Sun, 10 Sep 2017 13:07:58 +0000

Alternatively, just using first names (the way people in non-insanely-formal workplaces actually do) would mean we wouldn't have to dash around in this ridiculous mapping chaos at all. I use first names and sometimes initials or self-chosen identifier handles to identify people ('rms', not 'Stallman'; 'Linus', not 'Torvalds'; 'mkp', not 'Petersen') simply because those are the identifiers by which they identify themselves to other people.

If these people were in the habit of introducing themselves to other people by their surname, using surnames as identifiers in news reporting might make sense. (For, e.g. political reporting, this is often true). But until Linus is is the habit of calling himself 'Torvalds' on the linux-kernel list, and other people call him that, I think it's the wrong identifier to use, and needlessly confusing. It seems rather disrespectful, too. Why call someone 'Machek' when he's made it clear in every post he makes that he wants to be called Pavel?

(This may be a generation-gap thing, too. A lifetime ago, perhaps before WWII and definitely before WWI, it was disrespectful to not call your elders and betters by their surnames whether or not they called each other by anything else, but in my view this has completely inverted by now.)

That will take some getting used to...

mcortese — Sun, 10 Sep 2017 08:24:10 +0000

It's not journalistic style, it's how we actually speak: you fully define the persons the first time you mention them, then you refer to them with a placeholder. "I met my younger brother at the pub. He was drinking a beer".

LWN's style is the same: full name and surname (and possibly role in the kernel community) at first mention, unadorned surname afterwards.

There's no need for your cheat sheet, if you don't remember who is who, all you have to do is go back some paragraphs.

That will take some getting used to...

nix — Fri, 08 Sep 2017 21:03:42 +0000

Nope! Jon uses first names everywhere, just like, er, everyone who's actually involved in kernel development. Using surnames is, as noted, weirdly distancing.

That will take some getting used to...

Wol — Fri, 08 Sep 2017 08:45:01 +0000

No doubt because there are (afaik) several Linus involved in kernel development. I'm sure I've seen the odd post/message/article where they were both there at the same time ... :-)

And the same almost certainly holds true for other names!

Cheers,
Wol

That will take some getting used to...

rvfh — Mon, 04 Sep 2017 16:11:25 +0000

At LWN, I have always seen the name printed in full first (e.g. "Linus Torvalds"), then only the surname used (e.g. "Torvalds"). It is the case in this article.

That will take some getting used to...

nettings — Mon, 04 Sep 2017 15:49:08 +0000

Granted, it's widely accepted proper journalistic style, but having to mentally map arbitrary identifiers to values that actually make sense to the way I think about the kernel community is a strain on my wetware at times. If you have the same problem, and in the open source spirit, here's my cheat sheet: { torvalds: "linus", machek: "pavel"};
;)

printk() and KERN_CONT

grawity — Thu, 31 Aug 2017 11:45:23 +0000

He's defending it now, but I remember when the new /dev/kmsg system got merged and broke implicit (non-KERN_CONT) continuations for the first time. Linus wasn't amused.

(Well, it's fine. As a userspace person, I'd much rather have /dev/kmsg with strict records than the old crusty interface with continuations...)

printk() and KERN_CONT

abelloni — Thu, 31 Aug 2017 07:17:25 +0000

I had déjà vu and I was wondering were I saw this. I was actually greatly reported in https://lwn.net/Articles/703110/. I guess Pavel needs an LWN subscription ;)