36

Many of the research papers that I have read are not dated in terms of publication date. By dating I mean including at least the publication year. The papers I refer to are mostly free PDFs from the internet on various topics, usually affiliated with some academic institution (mostly universities).

One can then only guess the publication date from the dates of newest references. Does this weird trend have some reasoning? If I wrote a paper or even its draft and made it public I would clearly date it.

Edit:

It seems I have missed the option of checking the PDF document properties which show a creation date. But this still does not answer the question of not including the date within the body of the document itself.

Kozuch
  • 485
  • 4
  • 5
  • 5
    What is your field? In my areas of interest (CFD, finite elements, etc), submission or acceptance dates and publication dates are very common. Also, there's almost always a copyright statement on the article that includes the year. – Bill Barth May 05 '15 at 12:21
  • Well I am a computer science student. The papers I reffer to are mostly free PDFs from the internet on various topics... I can not really understand someone writing a 5+ page paper and not dating it... – Kozuch May 05 '15 at 12:24
  • 5
    @Kozuch Is is possible that you are asking why the PDF of the paper doesn't have the date printed on it? On the publishers' websites, you typically find these dates. The PDFs on authors' homepages are typically self-compiled using the standard style used for reviewing, which simply does not feature the dates. Authors are often required to use a self-compiled version. Also, only few people see this is as a problem, as for a citation, only the year is mandatory to be included, and the authors' homepages, which link to the PDFs, have the years on them in most cases. – DCTLib May 05 '15 at 12:25
  • 9
    Yeah, if you're not looking at the officially published version on the journal or conference page, then you're not looking at a "published" article but a preprint or a draft. You need to get to the original source which will be the publisher's page. – Bill Barth May 05 '15 at 12:27
  • 3
    I am really not sure what you are talking about. Many journal styles have a submission / acceptance / publication date on the PDF. Even for the ones that don't (typically workshops and conferences), finding the year of the conference only is a trivial Google search. And if you can't find anything about the paper either in IEEEXplorer, ACM, DBLP, etc. - well, chances are it's not reliable anyway. – xLeitix May 05 '15 at 12:35
  • I don't have Acrobat Reader readily available, but on my pdf reader (evince, part of Gnome) I can access a creation date and a modification date in the menu entry file->properties. In case you are wondering, they are different from the file creation/modification date recorded by the filesystem. – Federico Poloni May 05 '15 at 12:43
  • 4
    Are you talking about the random "free tutorial" someone writes up (which rarely has a publication date), or are you actually reading research/peer-review-quality papers? – apnorton May 05 '15 at 15:11
  • This sounds very odd. I can't recall ever seeing a paper that doesn't have at least a date of submission or publication somewhere in the text. Can you give some examples? – Superbest May 05 '15 at 18:44
  • 1
    I don't think you should trust a computer science paper which in a DOC. – tomasz May 05 '15 at 22:41
  • A PDF file normally contains the date of its creation in its Properties (which can be accessed, e.g., using Ctrl+D in SumatraPDF on Windows, or using "pdftk [name of PDF] dump_data" on Linux). This is at least an upper bound on when it was actually written (but so are publication dates). – darij grinberg May 06 '15 at 03:25
  • PDF creation date is easily changed, and may be inaccurate in case of re-created PDFs. – Raphael May 06 '15 at 06:57
  • Because the papers you are talking about are not papers / publications. A random PDF which is not per-reviewed, not even on a preprint server, is not full scale publication in the academic sense. – Greg Apr 22 '21 at 19:25

3 Answers3

31

"The papers I refer to are mostly free PDFs from the internet on various topics" – then it's most likely one or both of:

  • It is not a peer-reviewed paper published by a reputable journal. Be careful what you read, there are plenty of quack theories out there
  • You are viewing a preprint, and by finding the article on the journal's website you will find the date. Also, the final version of the article will probably be better formatted, and may be better written and have mistakes corrected (this depends upon the stage of publication at which the preprint was circulated).

Many preprints are circulated prior to publication, so you may not find a better version yet. It's still really poor form for an author not to date a preprint. If a preprint is a few years old and has not yet been published, you have to wonder why.

I can not recall ever seeing a journal that does not put the journal name, date, and issue number on at least the first page. It would be very poor practice not to do this, because taken on its own it is impossible (without further research) to tell where the article came from. It's common even to date every page with the journal name, date, and issue number.

I suspect that if you're seeing this trend in most of the "papers" you read, that you really aren't reading research articles written by legitimate researchers. Ask your advisor or lecturer for the names of the most relevant journals in the field, and start from there.

Moriarty
  • 15,035
  • 4
  • 45
  • 76
  • 4
    In my area (pure maths), many preprints aren’t dated, and even for published papers, the preprint is often the easiest version to find. So while most of the papers I read have a date in principle, the specific pdf I read them from is very often undated. – PLL May 05 '15 at 12:56
  • 5
    @PLL, but when you cite them, you go find the publication date, right? – Bill Barth May 05 '15 at 13:10
  • 3
    @BillBarth: of course, yes — one cites the journal version, not the preprint. – PLL May 05 '15 at 14:34
  • Since the OP is a computer science student, one of my CS professors told my class that most of the top tier journals in our field are published by the ACM and the IEEE organizations. Searching Google Scholar for the ACM and the IEEE brings up plenty of high-quality peer-reviewed papers. – Kevin May 05 '15 at 17:20
  • 1
    Most PDFs you find are pre-prints, and the standard workflow of many journals and conferences does not include dates in preprints. – jakebeal May 05 '15 at 19:51
  • 1
    @PLL: In my experience, preprints are usually ciriculated using arXiv, and you can see the submission dates there and every paper (and every version of paper) is automatically stamped with submission date. – tomasz May 05 '15 at 22:37
  • @jakebeal "Most PDFs you find are pre-prints" – the only preprints you read should be the ones that are not yet published or that you don't otherwise have access to. I still find it hard to believe that it is not possible to find the date of a journal or conference preprint – if not in the PDF, then obviously available on the website. What's next, omitting the author's names? – Moriarty May 05 '15 at 22:47
  • 7
    @Moriarty I highly disagree: when I search online and Google Scholar pops up with a PDF, it's probably a pre-print, not the paywalled final version. That's the one I'm going to read, even when I have journal access, because it's one click rather than a bunch of fiddling with library portals. And those preprints are almost always missing the final date. The date is then easy to find online, it's just not in the PDF. – jakebeal May 06 '15 at 11:55
  • @jakebeal A "bunch of fiddling" with library proxies can be reduced to a browser bookmark, if all you have to do is append a proxy address to the URL: javascript:void(location.hostname=location.hostname+'.yourproxyhere'). I don't know how it is in CS, but preprints are not always of the peer-reviewed version. So maybe I'm being rather conservative, but it's not a practice I'd actively encourage. Perhaps also I'm unusually lucky as an astronomer that the published article is always only 2 clicks away from the arXiv page... – Moriarty May 06 '15 at 12:52
17

Most (CS) papers that you find on the websites of the publishers have the bibliographical information in the PDF, typically on the first page of the paper. This typically includes the year. These papers are however also typically behind paywalls, meaning that you won't get access to these unless you (a) buy a copy of the article, or (b) have access through your institutional subscription.

Many research paper PDFs that you find freely on the internet are self-archived versions of the papers. These are PDFs provided by the authors on their personal or institutional web pages and not prepared by the publisher. While these do not constitute the official versions of the papers, authors normally do not modify the content of the PDF so that the official and unofficial versions of the papers get out of sync. Now it happens to be that most paper style files provided to the authors for writing their papers with for a specific venue do not feature a field for the bibliographical information. Rather, the information is later added by the publisher. Thus, the information is missing on the PDF made by the authors themselves.

It should also be noted that only few people see this as a problem. The bibliographical information is contained on the authors' webpages from which you often download the papers. Also, you mainly need the bibliographical information for citing the paper, and for that, you can pretty much always download the whole bibliographical information entry for a paper from the publisher's website. Just type the paper title into your favorite search machine and click onto the respective result. For computer science, most papers are in DBLP anyway, which also gives you a complete bibliography entry at the expense of a mouse click.

DCTLib
  • 15,060
  • 46
  • 66
  • +1 for bib data from the publisher even if you get the paper direct from the researcher. You can also quickly compare the abstract in case your search engine took you to a similarly-named paper from the same authors (e.g. you forgot to put quotes round the title). – Chris H May 05 '15 at 14:19
1

I agree that it is extremely frustrating to not have a date on a technical paper. It is often impossible to tell the currency and hence whether any conclusions represent the latest thinking. The only reason I can think of is that most technical papers are published by technical associations, who then generate revenue by selling access to papers. By not including a date, researchers are unable to identify currency and are thereby forced to purchase more papers than they need, hence generating more revenue for the publishing house or society.

Peter Jackson
  • 11
  • 1
  • 1
    unable to identify currency and are thereby forced to purchase more papers than they need --- I'm confused. Don't most published papers include a received date, a date that (in my experience) is nearly always provided to those (such as myself) who do not have access to papers behind a paywall? "randomly chosen" example 1 (gives received, accepted, published dates at bottom of the web page) and "randomly chosen" example 2 (also gives received, accepted, published dates) – Dave L Renfro Apr 22 '21 at 09:34
  • @DaveLRenfro You’re right. However, the paper itself, e.g. the PDF, does not indicate the date it was published. It is part of the metadata that you see in a citation. – Asim Jalis Nov 08 '23 at 23:16