Daniel Lemire's blog

, 8 min read

“I have read all of your papers”

8 thoughts on ““I have read all of your papers””

  1. Greg Linden says:

    Really? There’s at least a couple people I follow so closely I can actually say this, I think. I have alerts on a few, for example, and see every paper they publish. Yes, they are that amazing, at least in my opinion.

    But sure, in general, yes, even people I follow closely and admire deeply, there’s only some of their work I will have actually read.

    1. I should clarify that my post was not about the fact that nobody is interested in research papers or that people should stop writing research papers, or that there are too many research papers. I meant none of these things. Just in case that’s what you are implicitly replying to.

      Back to my statement…

      It is like saying that nobody reads EULAs. Of course, it is true. But then you will find a few people who actually read it all and review them. I am sure that there are bloggers who specialize in dissecting EULAs.

      I still think it would be fair to say that nobody reads EULAs. It is true to such a good approximation that it is “true enough”.

      Greg, I know that you are one of the few people (outside academia) who actually read research papers. I have attended local (Montreal) meetups where engineers gather to discuss research papers, so there are definitive more. And people who follow my blog know that I read quite a few research papers.

      I also follow a few people, more than a couple, and I see all of their papers. Here is one of them : Peter Turney. But, to take Peter’s example, he writes relatively few papers. I attended a talk where he described his yearly output, and it was something like 5 outputs a year, with not all outputs being research papers. (I think he would count a piece of software as an output.) Another one I follow is Harvard’s George Church. And so forth.

      However, to say that, for example, I have read all of Peter Turney’s papers would be a lie. And I don’t think that even George Church has read everything George Church has co-authored.

  2. Perhaps it was something one could have said in bygone times, where publication was not so prolific, or meaningless.

  3. Martin Cohen says:

    Re: Google. What about journals fhat have no online presence?

    1. What about journals that have no online presence?

      Please see my reply to Greg. There are, of course, journals that Google has not indexed… but it is true enough to say that Google has indexed them all. It is a very good approximation.

  4. I tried to read one of yours but it was written in French

    1. Must have been a paper by a student I supervised.

  5. Andrew Dalke says:

    “bygone times, where publication was not so prolific”?

    When do you think that was? Leibniz and Gauss each published many volumes. Even by the 1940s, a hot topic was in how to deal with the publication crisis caused by the torrents of new publications.

    As to the main topic, I am an amateur scholar of Calvin Mooers, an information retrieval pioneer, and have read about 3,000 pages of his publications, letters, and notes. That said, I haven’t read all of his works, like his 1947 book “Electronics: What Everyone Should Know”, which is available online so Google knows about it.

    On the other hand, Google hasn’t read some of his publications which I have. He published through his company’s house organ. While many people have cited some of his publications, the only copies I could find were in Mooers’ own papers archived at the Charles Babbage Institute. His papers include records submitted to the USPTO which demonstrate three uses of superimposed coding before his invention of random superimposed coding. (Knuth in TAOCP doesn’t know about these earlier uses.)

    I also have a 1956 internal publication from the US National Bureau of Standards which is not online nor available from NIST (the successor to the NSB). I bought it from the used book market. And I have read MSc. and PhD theses from the 1960s and 1970s which are not scanned and only available by visiting the library in person. Some of the authors are still active in the field. The MSc theses never will be scanned for archival purposes because of copyright law, and because the library is unlikely to keep them until copyright expires.

    These are rare edge cases meant to test the hypothesis. They all far pre-date Google, and are part of the grey literature which is not controlled by commercial publishers and is much harder to archive and digitize.

    To a very good approximation, Google has read all the papers I reference.