LWN: Comments on "The great leap backward"

The great leap backward

sdalley — Thu, 11 May 2017 14:35:21 +0000

Not to mention the Muslims and the Atheists...

The great leap backward

Darkstar — Wed, 10 May 2017 14:03:20 +0000

Oh boy... there's a cross in there. We can't have crosses now, can we? Think of all the Christians who might be offended by that

;-)

Sybase ASE

sbishop — Sun, 07 May 2017 03:22:46 +0000

I'm surprised that no one has mentioned that Sybase bumped the version number of their "Adaptive Server Enterprise" product a couple of years ago from 12 to 15 for the very same reasons.

The great leap backward

muwlgr — Sat, 06 May 2017 18:55:10 +0000

"alternative incrementing" is a gem clearly modeled after "alternative facts" (or earlier version, "alternatively gifted").
liked it very much

The great leap backward

Wol — Thu, 04 May 2017 13:59:43 +0000

Kernel 1.3.100

Or was that so long ago the developers have forgotten how to count since then?

Mind you, according to the website I got that from, the latest kernel is 2.5.3 :-)

Cheers,
Wol

The great leap backward

thoeme — Mon, 01 May 2017 09:01:46 +0000

Hi Jon,
This article was a nice start into monday morning :-) Besides, if the OpenSuSE community (which I am a member of) has nothing else to worry about, all is great !
- Thomas

The great leap backward

lab — Sun, 30 Apr 2017 17:40:27 +0000

Oh Jon, you gave me a pretty good chuckle there. We humans really are quite silly sometimes :-)

The great leap backward

excors — Sun, 30 Apr 2017 09:33:39 +0000

No, I think you mean http://rebeccapurple.bikeshed.com/

The great leap backward

tedd — Sun, 30 Apr 2017 00:24:53 +0000

Wrong address. Should be http://antiquewhite.bikeshed.com

The great leap backward

tdz — Sat, 29 Apr 2017 18:02:01 +0000

Thanks a lot for this humorous and entertaining article.

The great leap backward

storner — Sat, 29 Apr 2017 11:16:53 +0000

This is pure bike-shedding. http://www.bikeshed.com/

Quality article

rahvin — Fri, 28 Apr 2017 22:28:13 +0000

You left out the Islamic Calendar based on Lunar cycles independent of Solar cycles. Now that would be interesting as I don't think anyone outside the middle east would even be able to figure out how to convert the Islamic dates without some sort of calculator that did it for them. Anyway, this is the problem with any version number, everyone around the world has superstitions about different numbers "luck" and by the time you throw them all out you've got to jump around on numbering constantly.

For those that don't know: https://en.wikipedia.org/wiki/Islamic_calendar

The great leap backward

nix — Fri, 28 Apr 2017 18:58:41 +0000

That's probably because putting random chunks of what looks like rock in food is often dangerous, so triggers a disgust response. (An even more extreme example: almost nobody from any culture would be happy about putting a piece of iron shaped like a turd into their cooking. This might be irrational, but it is definitely sensible, because most things shaped like that are really really dangerous to put in cooking.)

The great leap backward

nix — Fri, 28 Apr 2017 18:55:12 +0000

Yeah... if you do that you probably want to spell his surname right. :)

The great leap backward

nix — Fri, 28 Apr 2017 18:28:24 +0000

There were definitely people who were very offended about it being called Demon.

The great leap backward

AdamW — Thu, 27 Apr 2017 17:54:19 +0000

I was particularly fond of the one where someone tried to work around some bug caused by the space by wrapping the whole thing in quotes. Unfortunately, they used *single* quotes. So whatever the thing was (I forget) wound up with the codename as "Schrödinger" and some "'s cat" lying around making the place all messy...

The great leap backward

danielkza — Thu, 27 Apr 2017 13:34:28 +0000

I would bet you there are thousands of scripts of all kinds out there that rely on the Debian version number, even if they are not part of the distribution itself. One just has to browse the most popular configuration management modules for software such as Puppet, Chef and Ansible to find checks that use comparisons such as `distribution_version > 7` to determine if a Debian system should have a particular package available, or whether systemd or SysV init scripts should be installed.

The great leap backward

lkundrak — Thu, 27 Apr 2017 12:48:50 +0000

This is easily the most entertaining LWN article ever. Was it written in a smoke-filled room?

Thanks for the laughs.

The great leap backward

lacos — Thu, 27 Apr 2017 12:14:29 +0000

> The "lucky iron fish" is supposed to be used when cooking so the iron will leach into the water in the pot and correct for iron deficiency in the users' diets. Iron does leach into water, so what's superstitious about that?

The fact that the ingot had to be cast into fish form, otherwise people wouldn't put it in their pots? See e.g. the attempt with the lotus flower shape:

> The research group distributed an iron disc to women in the village, asking them to place the disc in their pot while making soup or boiling water. The women were reluctant to use the chunk of iron while cooking, and "almost no one used it". [...] Charles and others distributed iron ingots in the shape of a lotus flower, but that was also rejected by the villagers. During discussions with village elders, Charles learned about a fish species deemed a symbol of good luck, health, and happiness in local folklore. The group created fish-shaped iron ingots, which were received more positively by the villagers and led to immediate increases in blood iron levels amongst the villagers, and anemia was virtually eliminated. Charles would later state that "You can have the best treatment in the world, but if people won’t use it, it won't matter."

yyyymmdd.hhmmss

dmoreno — Thu, 27 Apr 2017 11:06:53 +0000

Or just YY.MM as ubuntu does.

The great leap backward

MarcB — Thu, 27 Apr 2017 10:59:18 +0000

The sentence "Somewhere between 2038 and..." makes no sense to me.

I tried hard, but was not able to come up with a date after January, 19th 2038 in my head. Obviously, there is nothing after 2038.

The great leap backward

linuxrocks123 — Thu, 27 Apr 2017 10:40:53 +0000

The "lucky iron fish" is supposed to be used when cooking so the iron will leach into the water in the pot and correct for iron deficiency in the users' diets. Iron does leach into water, so what's superstitious about that?

The great leap backward

zenaan — Thu, 27 Apr 2017 09:48:26 +0000

Oh, come on! Seriously, what's wrong with completely random major version numbers anyway? (Or minor version numbers for that matter..)

yyyymmdd.hhmmss

sysrich — Thu, 27 Apr 2017 09:26:33 +0000

We do YYYYMMDD for Tumbleweed, openSUSE's rolling release. Because there it makes perfect sense to do date based versioning.

But in the case of Leap there is meaning behind the madness

Like in SLE, a major version is meant to be somewhat compatible through it's lifespan. That means several years where users can expect the operating systems behaviour to be mostly the same, no major architectural changes.

Minor updates of that major version (or Service Packs in SLE terminology) may introduce more than just new bug fixes. New features and versions are certainly possible and do happen with surprising frequency, but should never compromise the general premise than the major version is consistent throughout it's lifespan and major 'workflow breaking' changes should wait until the next major release.

In a situation like that, a major/minor model like we now have with Leap 15 (to be followed by 15.1, 15.2, etc, until Leap 16) is very useful, whereas date based systems would not be reflective of what people can expect in each version.

The great leap backward

aggelos — Thu, 27 Apr 2017 08:55:06 +0000

I find it ironic that the article goes into such length for what it claims is a totally trivial issue, not really worth any discussion, why are people still talking about it, the stakes are low I tell you, wait now it's made the press too, wth! Etc.
Personally, I appreciate the occasional humorous sentence in a technical article more than I did this article - to me it came off as kind of forced. Please never drop the humor though :-)

Quality article

sysrich — Thu, 27 Apr 2017 08:51:06 +0000

Oh we have a reasonable understanding of why SUSE chose to do what they did, it just isn't relevant to the community or the discussion we needed to have about openSUSE's versions.

SUSE chose to skip 13 and 14 because of marketing/sales concerns with 13 being considered unlucky and any number including 4 being related to 'death' in eastern cultures.

Obviously, those concerns are something which the openSUSE Project does not share - we already had openSUSE 13.x, and the plan was to have 4x.y as our versioning for the next few decades with Leap.

But that's one of the joys of openSUSE being a very independent community from it's partner and sponsor in SUSE. SUSE made their decision about SUSE Linux Enterprise and figured the community was free to do whatever the heck we wanted.

We could have ignored SUSE's decision entirely, there was no suggestion from SUSE management or anything like that about changing anything.

But given the whole unique point of openSUSE Leap is that it is, by design, a community built derivative of SUSE Linux Enterprise, yeah, the Board and openSUSE's release team are somewhat fond of reflecting that closeness in version numbers. The community at large had expressed similar thoughts in the very, very, very long mailinglist thread we had a few years ago when we chose 42.x in the first place.

Heck, there was even suggestions back then that we should try to encorage SUSE to skip a SLE version or two so they could sync up, but we dismissed that idea as madness because we assumed SLE would NEVER skip a version ;)

And so SUSE's jump (leap?) from 12 to 15 gave us an opportunity to sync up the version numbers. We took it. It isn't going to be pain free, but it's going to be a lot less painful now than doing it in the future. If long mailinglist decisions are the worst fallout of this decision, I can live with that.

Going forward we have a much more sustainable, rational, and easy to maintain structure for Leap versions, which is a good thing, and SUSE and openSUSE can collectively screw up our version numbers of our regular release distributions together in the future, which isn't the worst risk in the world to have.

The great leap backward

lacos — Thu, 27 Apr 2017 08:43:41 +0000

(1) This article is pure gold. Thank you. It has so many shining nuggets that it's hard to pick examples.

> leapfrogs Leap

Beautiful.

> 19 is suspiciously prime

Oh man. The tears.

(2) The incredible backwardness and savagery encoded in various cultures are mind-boggling. Man is man's biggest enemy. Here's another example: <https://en.wikipedia.org/wiki/Lucky_iron_fish>.

NB I'm not picking on the so-called developing parts of the world; the so-called developed world is terribly superstitious too. Anti-vaxxign? Homeopathy? Remote healing?

(3) Regarding the less than obvious "suse_version" values, I guess the distro is called *tumble*weed for a reason :)

The great leap backward

sysrich — Thu, 27 Apr 2017 08:33:28 +0000

Jonathan, I think I can wholeheartedly say that this article, and the laughs it gave me, are quite possibly the highlight of the whole endeavour. Thank you.

It's not going to be as controversial as changing version numbers but I've half a mind to suggest to our release team we add a codename Leap 15 and call it "Corbert's Crisis" in honour of this article.

Actually, on second thought..maybe it's best we stick without a codename..

Regards,

- Richard Brown

The great leap backward

niner — Thu, 27 Apr 2017 07:21:56 +0000

You found out what horrors code names can bring, openSUSE is testing the waters with creative numbering. That's teamwork! Would be rather sad if we only had a single distribution that got all the excitement ;)

The great leap backward

AdamW — Thu, 27 Apr 2017 06:38:51 +0000

SUSE. Seriously. Take it from us, the people who once code named a distribution "Schrödinger's Cat", then found that the number of bugs that can be caused by a codename with a non-ASCII character, a space *and* an apostrophe in it was so damn high we *stopped codenaming releases*. Take it from us, and do the simplest goddamn thing possible.

Yours truly, the Fedora team and a large bottle of gin.

The great leap backward

_xhr_ — Thu, 27 Apr 2017 06:22:30 +0000

Great article. I like it when our beloved editor writes a not-so-serious article!

The great leap backward

eru — Thu, 27 Apr 2017 06:02:39 +0000

Unsurprisingly, some community members are horrified by the idea of a software release having a lower version number than its predecessor

Indeed, having version numbers increment really is the other important piece of information they convey, the other being uniqueness. Otherwise they could proceed according to the Fibonacci series, or be based on Julian day numbers, or whatever.

Quality article

mcatanzaro — Thu, 27 Apr 2017 03:59:54 +0000

The article was good, but best part is that nobody seems to know why SUSE decided to skip 13 and 14. Incredible community management there.

The great leap backward

JdGordy — Thu, 27 Apr 2017 03:02:33 +0000

Indeed.

@Jon, how about just dumping the security page of the newsletter and do a lighthearted article every week? (I mean, for those who cringe at the security state of the world instead of laughing maniacally?)

The great leap backward

jkowing — Wed, 26 Apr 2017 23:03:28 +0000

No, no, it is clear that +40 is the only acceptable response!
Thanks Jonathan for a much needed hearty laugh!

yyyymmdd.hhmmss

tnoo — Wed, 26 Apr 2017 22:58:43 +0000

>Whatever happened to the old practice of using dates for version numbers?

like, mmmh, 95, 98, 2000, ..., 7, 8, 8.1, 10?

yyyymmdd.hhmmss

felixfix — Wed, 26 Apr 2017 22:29:50 +0000

There are many variations. Try ddd.hhmmss, with ddd being days since initial release.

As for what calendar, that's easy -- the project owners decide. I already said versions are valid only for a specific project and it's meaningless to compare versions in different projects. This should be obvious, as should its corollary, that multiple versions for the same thing just make a mess for no gain.

If project owners want to use French revolutionary decimal time, or Star Trek time, more power to them.

If project owners want to use some bizarro negative time, or choose random numbers, again more power to them. But then won't make their would-be customers happy.

Knuth uses digits of pi and e, IIRC. Each version appends the next version.

Whatever floats your boat. If you want multiple versions based on different calendars, start your own project. But if you lead from too far out front, you may find you have no followers.

The great leap backward

Cyberax — Wed, 26 Apr 2017 22:16:23 +0000

They should switch to Chinese numerals to avoid conflicts: OpenSUSE 四十三 (43).

The great leap backward

neilbrown — Wed, 26 Apr 2017 22:01:40 +0000

It may well be 100 years until SLE-42 is expected, but I think there is an earlier concern that is much more significant for some of us. Somewhere between 2038 and SLE42 we can expect to see the release of Linux kernel v19.19. As it is clear that kernel developers cannot count beyond 19, I'm wonder if that might be the end of the line.

The great leap backward

excors — Wed, 26 Apr 2017 21:43:02 +0000

I wonder if many people complained to Demon Internet about how all their dial-up phone numbers ended with 666.