From Wikipedia, the free encyclopedia
Jump to: navigation, search
WikiProject Computing / Software (Rated B-class, Mid-importance)
WikiProject icon This article is within the scope of WikiProject Computing, a collaborative effort to improve the coverage of computers, computing, and information technology on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.
B-Class article B  This article has been rated as B-Class on the project's quality scale.
 Mid  This article has been rated as Mid-importance on the project's importance scale.
Taskforce icon
This article is supported by WikiProject Software (marked as Mid-importance).


I wrote this article, adapting from something written a while back, because, to me, cron is the main topic, not crontab. I am sketchy on what came before V7 and "Kernighan's cron" as we called it at Purdue and also a bit sketchy on the subsequent implementations. I hope others can round this out. I didn't really know that my cron got into "the real Unix" until I took a job at SGI in 1990 and looked at the source code and instantly recognized the awful coding style (no spaces, lots of asterisks in comments) I used as a grad student. I think Keith wrote the hardest part: figuring out the NEXT minute in the future that any particular crontab entry should be run, with out walking forward minute-by-minute asking the binary question, "now?" It took a while to debug that.

A few years back I tried to convince the Purdue Research Foundation that they should claim title to the Unix cron but we didn't save notebooks or even e-mails from those days...this was 1979!! So they chose to just let it fall by. I suspect a lot of code in Unix was written by grad students like me and Keith. Another one of those "warm up project" I designed was given to Chris Kent/Kantarjiev and the "pg" command in Unix came out of it. But everyone uses "more" now...though I believe pg was the first paginator that would allow you to go backwards, even if you piped output to it!

I apologize for any deviations from standards and style and decorum and trust the community will correct these or instruct me on how to do so...or both!

Rlb408 02:11, 9 February 2007 (UTC)

Not the crontab article[edit]

Unsure why someone added the description of the crontab fields at the top of the article. This topic is cron, not crontab. The format of cron's spec file is incidental to the topic and I think distracts. Since there is a separate article on crontab and this artile is about the history and algorithm, can we just delete the crontab format description from here? As an analogy, if there were an article about compiler optimization algorithms, would you put a description of the syntax of the language itself in that article? How very odd.

Rlb408 07:32, 17 February 2007 (UTC)

cron and DST changes[edit]

Is Vixie cron the only cron that can not tolerate changes to daylight savings time rules while running? (See If so, maybe a note belongs there. If it affects many crons, then a note should go here.--Theodore Kloba 13:59, 26 March 2007 (UTC)

apparent crontab inconsistency[edit]

I will leave to others the editorial decision as to whether the crontab format is appropriate to include in the cron entry. However the documented range of 0-6 and the statement that either 0 or 6 conveys a Sunday for DOW (day of week) appears to be contradicted by the statement that both 0 and 7 are considered Sunday, although 7 is considered trayf* on AIX manpages.

I find it hard to believe that 0, 6 and 7 are all intended. A Unix guru might be able to verify this point in relation to difrferent Unix flavors.

--Ed Smiley

"... on a 1 MIP system ..."[edit]

What does MIP mean in the text? It's wikilink goes to a disambiguation page whose entries don't make it obvious (for me). Could someone fix the wikilink so it goes to a page describing the "MIP" being talked about? --Zabadab (Talk) @ 12:37, 27 May 2007 (UTC)

Fixed it. IPS stands per instructions per second, and the M stands for million. Sometimes people get confused and think it's 2 MIPS but 1 MIP, but, if you think about it, that's a wee mistake. Nicol (Talk) 14:01, 30 May 2007 (UTC)

Acronym (1)[edit]

"CRON" is an acronym that stands for "Command Run ON"

That's probably not right. Sounds too much like "Unix System Resources", "Extended Tool Chest", and other such poor attempts to explain Unix terms. -- 14:24, 26 July 2007 (UTC)

See also the discussion below. — Xavier, 13:15, 12 June 2015 (UTC)

Code Examples[edit]

It would be great if someone could include what different commands you can set a cron job up to. I had to spend too much time trying to figure out how to run a php script once a day...

I will come up with a couple of examples on things you can do with cron. -- dat789 — Preceding unsigned comment added by Dat789 (talkcontribs) 10:31, 19 February 2015 (UTC)


The following discussion is closed. Please do not modify it. Subsequent comments should be made in a new section.

There's no reason for the current split. The articles are inextricably linked and short enough that a merge would be a good idea. Chris Cunningham 17:56, 29 August 2007 (UTC)

Oppose: While they surely originated together, CRON expressions have come to be much more widely used than in CRON. They have become the de facto standard for creating scheduling definitions. Thus, the link goes only one way, that is, from this article to CRON expressions, and not the other way round. Also, the article on the secondary topic (CRON expression) is currently larger than the would-be main one (cron), which I think makes it not a good idea. --uKER (talk) 17:15, 2 April 2010 (UTC)
Support per Chris. To most readers this distinction will be hair-splitting. That the format of a crontab file may now be used elsewhere is not sufficient to warrant separation. (In any case CRON expression is an abnormally cased article name - cron is not an acronym.) Globbet (talk) 21:37, 8 June 2010 (UTC)
Support you can still link to the page section as such: cron#cron expression. People are unlikely to search for "CRON expression." I am yet to be provided with any references that suggest this expression is used outside of interacting with the cron command. Even if it was used elsewhere, it still doesn't mean it should get its own article.--Bxj (talk) 10:53, 27 July 2010 (UTC)
Support CRON expression is not notable except as a part of cron, and the article itself seems little more than how-to content. me_and 18:17, 18 October 2010 (UTC)

Merge done as per foregoing. -- P 1 9 9 • TALK 16:57, 3 March 2011 (UTC)

The above discussion is closed. Please do not modify it. Subsequent comments should be made in a new section.

In "See also"...[edit]

Might and Magic II: Gates to Another World, which is set in a land called Cron

...does this really belong here? If anywhere, it should be at the top of the page, or in a disambiguation page. Warriorness (talk) 07:58, 19 December 2007 (UTC)

Nope. Moved. Chris Cunningham (talk) 13:48, 19 December 2007 (UTC)

Sketch for rework[edit]

Note to self or others: here is what I'd expect the article to say, if I directed one of my curious co-workers to it:

  • What it is in simple terms (lets you schedule commands to run periodically, without being logged in).
  • That it is enabled by default on most or all Unices.
  • What system administrators use it for (backups, cleanup, indexing ...)
  • What normal users use it for, and that they may not be allowed to.
  • The user interface (crontab(1), the file format, and the command's output mailed to the owner).
  • What the equivalent on Windows (if any) is called.
  • A reference to at/atd (Oh, it's there. Good.)
  • The history/algorithm stuff as-is, but not at the top of the article.
  • A pronunciation it kron (like the beginning of chronological), or cr??n (as in rhymes with pawn), or something else entirely? — Preceding unsigned comment added by (talk) 13:29, 22 December 2011 (UTC)

JöG (talk) 22:33, 28 February 2008 (UTC)

Wikipedia is not a Unix reference manual[edit]

Just a matter of opinion, but the value of wikipedia is that it captures material that is not collected uniformly elsewhere. Based on this thinking, I prefer that we not turn it into a Unix or Linux reference manual; that will diminish its unique value. If I want to learn how to use cron or write a crontab file, I can look it up in any of the hundreds of online reference manuals or simply type "man crontab" in a shell window.

I would like this article and others like it to tell the back story. You're not going to find this in any reference manual. The article should address questions such as

  • where did this come from?
  • what are its inner workings?
  • what is the technical basis underlying it?
  • what, if any, controversy is there?

Imaging your son or daughter or niece or nephew was to write an essay on "how did such-and-such come about" - would a reference manual page help? Not much. Similarly, I think it is important that new developers in the present day fantastic flurry of open source understand that even something as widely used as the Unix cron was started by one grad student, joined by another, in the 70's no less.

So, to me, the value of wikipedia, specifically concerning cron, is that the distinctions of the various crons are not widely known and that the Unix version is algorithmly distinct from all of the others. Do anyone care about that? Very few people, I suspect, but that applies to so many wikipedia articles.

Let's leave the reference manual content in the reference manual or some "wikiUnixReference" site.

Rlb408 (talk) 19:39, 13 April 2008 (UTC)

I agree with the above user. I clicked on this link randomly from an external site, and I honestly cannot understand most of what this article is about. I have no background in Unix or programming, but I shouldn't have to in order to understand a Wikipedia page. This page needs a major overhaul.§

*/31 example[edit]

I'm not an expert in these things, but there is an example of */31 in the minute field. Wouldn't that run twice, since 0 mod 31 = 0 and 31 mod 31 = 0? —Preceding unsigned comment added by (talk) 05:18, 9 August 2009 (UTC)

You can ask technical questions like this on forums. Here, we should provide a general working examples that can be reworked into a solution. Dat789 10:52, 19 February 2015 (UTC) — Preceding unsigned comment added by Dat789 (talkcontribs)

CRON expressions[edit]

For the second time, I have removed [information] which for some reason made a sad attempt at explaining CRON expressions. Anyone reading this article could probably be assumed to be familiar with CRON expressions, but in case they weren't they, can find an more than appropriate explanation on them in the appropriate article. Unless a merge was performed and expressions were then properly covered in this article, we're better off without that info given how terribly incomplete it is. --uKER (talk) 16:47, 2 April 2010 (UTC)

However, that article, CRON expression, describes an expression that includes seconds and year. That doesn't match the cron article. — Wdfarmer (talk) 17:51, 27 December 2010 (UTC)

The example documentation on the Quartz Scheduler site also makes use of a "seconds" field. —Preceding unsigned comment added by Charlieholder (talkcontribs) 15:28, 16 May 2011 (UTC)


Is "Command Run ON" is an acronym or a fan's backronym? Lacrymocéphale 12:21, 13 October 2010 (UTC)

I've never heard of it called that. The article has the name coming from chronos, which makes much more sense to me. me_and 12:24, 13 October 2010 (UTC)
Me either. I asked because I'd like to understand why it is on fr:Cron. It can be found but in poor references. Lacrymocéphale 06:00, 19 May 2011 (UTC)
As a long term Unix user, Chronos makes much less sense to me than any acronym. That's why I asked for a ref and some kind user added it (thanks!). But that ref does not seem to be very reliable to me, though. A reliable ref would be a Unix book from 1970s or 1980s, especially if it is written by one of the Unix "gods".
I went to Google books and found nothing for that years. Prior to 2000, I found only _one_ 1998 book backing the chronos etymology. On the other hand, I also found "Commands Run Over Nigh" (one ref from 1997) and "Command Run On Notice" (one ref from 1994). The latter seems the more likely to me, being from a 1994 Unix conference held for Unix specialists.
Hope this helps. I added the refs to the article. — Xavier,
It is interesting, but perhaps not authoritative (unless you demonstrate that the quote was by one of the original developers of the cron program — unlikely due to the 15-20 year gap). The relevant manual pages would normally hint at something like that. For instance I have at hand manual pages from SunOS4 which deal only with the relationship to time: crontab - table of times to run periodic jobs (6 October 1988), cron - clock daemon (6 December 1988). You might find something to verify that here. Looking at source code, I see that cron (and crontab files) were present in 3BSD (1979), using similar wording (and no attempt to make it an acronym). The wording agrees with more accessible/later documentation such as Tru64, HPUX, IRIX, etc. As your comment hints, other explanations than those based on time are rare, and presenting them as equal in the lede of the topic makes it unbalanced (see WP:FRINGE). TEDickey (talk) 21:45, 12 June 2015 (UTC)
Well, I'm neither trying to prove nor claiming it is an acronym. My point, as summed up in my edit is that we do not know the exact origin of the name, and that it _may_ be an acronym (or acronym-like, like tar or sed). Of course, being a scheduler, the cron service is related to time. But 1. that does not prove that "the name cron comes from the Greek word for time" as stated originally in the summary, and 2. that does not preclude the acronym-like origin, or any other origin (the name of a dog?).
The fact is that nothing is said about the origin of the name in the original Version 7 manpage (would it be described as a "chronological something" rather than a "clock daemon", then I wouldn't argue here). And we both fail to find any authoritative source telling us the actual origin of the word. What we both have found are only credible explanations, decades after the word cron was coined, and from source that are only loosely related to the origin (Unix expert at a conference, manpages from another Unix, etc). For that reason, if we only mention chronos here, we are giving undue weight to one probable (yet unattested) origin among others. — Xavier, 00:09, 16 June 2015 (UTC)
However, the topic as written has in the lede information which makes it appear that cron was intended as an acronym (with some dispute over exactly what), and there is no followup in the body of the topic (simply because there are not enough sources for that). Trimming the entire comment would improve the topic. As it is, it reads like original research TEDickey (talk) 00:41, 16 June 2015 (UTC)
The average reader who is interested in cron is likely able to understand the relationship with the Greek root without the necessity of confusing them by making it appear that the authors of cron had instead a sort of joke in mind rather than the obvious. TEDickey (talk) 00:56, 16 June 2015 (UTC)
Alright TEDickey, you are certainly right. Make any change/revert you deem necessary. — Xavier, 23:27, 19 June 2015 (UTC)

The alternative names are fantasy. But they've appeared in print. On 22 December 2015 Doug McIlroy updated the page based on private communications from Ken Thompson, but it was backed out because there was no reference. This is a clear indication of the limitations of WP:RS. Watch this space Groogle (talk) 02:06, 23 December 2015 (UTC)

A quick check also fails to find any WP:RS relating Douglas McIlroy with the editor who made the change mentioned. Talk page comments are, as usual, not a WP:RS TEDickey (talk) 08:57, 23 December 2015 (UTC)
Likewise, quora as a source is no improvement over talkpage comments (it's not a reliable source, if one chooses to read the guideline. TEDickey (talk) 00:01, 24 December 2015 (UTC)
It seems that we have three issues here:
  • The changes do not satisfy WP:RS.
  • I haven't demonstrated any relationship with Doug.
  • You, and you alone, are objecting to these changes.
But the references to the other fantasy expansions also don't fulfil WP:RS. They're random quotes in documents without any kind of substantiation. We have them on the one hand and verbatim comments from the author on the other hand (did you check the quora article?). And what should my relationship with Doug have to to with the matter? You backed out his change too. Half the Unix history community is up in arms about this matter. What outcome do you want, given that there are no reliable secondary sources? Have you considered WP:UCS and WP:IAR? Do you doubt the veracity of Ken's statement, or are you just sticking to bureacracy? Groogle (talk) 23:02, 24 December 2015 (UTC)
You apparently do not read the change history. Start with this, then this and this, and perhaps have something to comment on. I make a practice of verifying sources on edits; something that you might also try. TEDickey (talk) 02:06, 25 December 2015 (UTC)
As I noted: edits on quora are no closer to being a WP:RS than your comments on this talk page. In context, a suitable source might be the ostensible author's personal website. That's not quora. TEDickey (talk) 02:08, 25 December 2015 (UTC)
To review: quora (like Wikipedia) is a community-edited resource, with (despite the implication of names) is largely by anonymous people, with/without aliases. None of those satisfy the guideline for reliable sources, which says many things (all relevant), including the section on self-published sources. TEDickey (talk) 10:45, 25 December 2015 (UTC)
I did, indeed, read the change history. I referred to all of your referred changes. I commented on them and asked questions which you did not answer. I verified the sources both of the former claims and found them unsubstantiated. Did you? Just putting something in print doesn't make it a reliable source. Communication from the author generally is reliable, even if it doesn't fit the standard Wikipedia policy. But there's a solution for that, WP:IAR. I referenced that, but you chose to ignore it.
What do you mean by ostensible author? The author of cron, who is without a shadow of a doubt Ken Thompson, or the author of the article? The latter doesn't satisfy WP:UGC.
Can you come up with somebody else who agrees with your interpreation of the Wikipedia policy in this context? Groogle (talk) 22:38, 25 December 2015 (UTC)
The author(s) of the edits (who may/may not be the individual posting on quora). There's no reliable source of information so far to discuss, so I don't have any idea why you continue to make comments. TEDickey (talk) 12:57, 26 December 2015 (UTC)
By the way, you appear to have a misconception about the amount of attention any given page on Wikipedia may draw. Starting with the page-view statistics, not many people were reading it until this discussion came up; and (as usual) a large fraction of the edits have been solely for the purpose of adding unsourced/poorly-sourced information. Your implication that there is a large well-informed community of editors devoted to making this a good topic falls short of the observable facts. TEDickey (talk) 13:09, 26 December 2015 (UTC)


I tested it on my ubuntu, and it fails.

24 12 * * 4#2  echo "notOK" >> /tmp/test.txt
24 12 * * 4#3  echo "OK" >> /tmp/test.txt

my /tmp/test.txt showed both "notOK" and "OK".

Need source on that section!! —Preceding unsigned comment added by Fridim (talkcontribs) 10:38, 19 May 2011 (UTC)

Which implementations use the question mark?[edit]

Debian and Ubuntu GNU/Linux don't accept the question mark. Which cron implementations do? Gronky (talk) 15:27, 22 January 2012 (UTC)

Specifying both DoM and DoW[edit]

At the end of the overview section it says

While normally the job is executed when the time/date specification fields all match the current time and date, there is one exception: if both "day of month" and "day of week" are restricted (not "*"), then either the "day of month" field (3) or the "day of week" field (5) must match the current day.

The emphasis on "either" and "or" makes it sound like if both match it doesn't run. This doesn't seem right. That's not how it works right? I recommend this wording instead:

While normally the job is executed when the time/date specification fields all match the current time and date, there is one exception: if both "day of month" and "day of week" are restricted (not "*"), then one or both must match the current day. — Preceding unsigned comment added by (talk) 00:18, 12 July 2013 (UTC)

Done. — Xavier, 00:20, 16 June 2015 (UTC)

"CRON expression" section[edit]

Except for its "Nonstandard characters" subsection, most of the information in said section is redundant to information that is mentioned earlier in the article. What should I do? --SoledadKabocha (talk) 06:39, 17 August 2016 (UTC)