share interactive transcript request transcript/captions live captions
download
|
MyPlaylist
SPEAKER 1: It is the greatest privilege and pleasure to introduce my dear research colleague and friend Professor Roald Hoffmann. As you will all know, we are very, very fortunate at Cornell to have him here. He shared in the 1981 Nobel Prize in chemistry. And if that tells you nothing else, it should tell you that he knows a lot about the subject.
And he's going to be talking today about what finally is our, so to speak, stock in trade, research papers. And how do they get published? And what is the right strategy? And what is the right form of procedures and so forth? And by the way, he should know. He's got just a shade under 600 publications to his name, so he's well-practiced in this business.
[LAUGHTER]
Now, what he's going to do today is outline the problem before us all. We just want to get our fundamental research published, in front of our colleagues, and even, hopefully, gain the acceptance of our colleagues. And that means that the documents have to be exceedingly carefully and thoughtfully prepared. And that's what Professor Hoffmann is going to be telling us about today.
So the title of his talk is Publishing Scientific Papers please join me in welcoming Professor Hoffmann.
[APPLAUSE]
ROALD HOFFMANN: Yes. So there are many things to talk about and thank you all for coming and eating standing up.
So first, the good old days. Just for nostalgia. So what you see at left is, this is a paper from 1981. And what you see at left is where the word cut and paste comes from, because this is-- you can see the lines of it-- a cut and paste from the pieces of actual paper, which we-- we actually use different colors of paper even, and this is the writing of Kaz Tatsumi, a distinguished Japanese chemist, who was a postdoc here of mine. And his writing, It's really interspersed with mine. There are two patches of his writing at the very bottom and near the top, but the rest is my writing.
The drawings were done here sketched by hand. Then they were done on paper sketched out a little more carefully. Than they were given to Jane Jorgenson, whose kids I put through college with these, and were done in India ink on tracing paper with a Leroy lettering set. These are all words that hearken to a distant past, which you can find in antique stores along with circular slide rules and other things like that. And eventually they find their place in the paper. So the drawing you see here, you see at upper right in there.
But the process, actually, of getting the paper published has not changed. Now what does it look like today? So this is the way we submitted a paper, Yuta Tsuji and I, in 2012. And we actually were obliged by the Journal of the American Chemical Society to submit it in a template that they provide. So at this point it actually looks pretty much the way it looks when it will be published. There isn't that much difference in it, with one thing which I will, however, describe in a moment.
So now the paper prepared, the drawings are done by computer, no longer done by hand and then by a draftman or woman. Not only did I employ Jane Jorgenson, but I also employed the children of my colleagues to provide gainful employment in making these drawings in their time. And so the Chester children worked for me at one time, two of them, at least, in that capacity.
But now let's talk about the steps in the process. So there are lots of interesting questions. Some of these would take a whole lecture by themselves. One is to do the research, how one does research. Another thing is about writing the paper. And there all kinds of questions that come up. When is it time to write the paper? Who writes it? What is the order of the authors? Who are the authors of the paper? How to credit previous work. There are questions here of ethics that come in. And which work to cite?
And I will not talk very much about these, though I am very open and I have written about some of them. And I'm actually writing a paper now about citation ethics, which is another story. Then there's a question of what journal to submit it to. Now if you are a graduate student or a postdoc, the decisions on these things will probably be made by your research advisor, with some consultation with you. But eventually in your life, you will be making those decisions yourself. And there are interesting questions around all of them.
But let's go through the process. You have done the research. You have written the paper. You're the authors. So the first thing is submission. And here is a sort of line of all the steps, just so you see. This is not meant to discourage you. Just to tell you about the steps that are involved. So you submit it to a journal, which means online submission today. And it means submission to the editor-in-chief, the chief office of that journal.
That editor-in-chief has a consulting technical editor, who will do a preliminary screening, whether you fulfilled some of the basic criteria. So if the communication is only 2,000 words and if you insist on sending in something with 3,000 words it'll get stopped right here, of course. So there are other things about the technical editor which I'll tell you. Technical editors are basically subhuman.
[LAUGHTER]
We'll talk about this. Whereas, editors are human.
The editor-in-chief, usually for a larger journal, has several editors. For instance, Journal of American Chemical Society of Phys Rev, this would be true. They would send it to a subeditor, based on the subject of the paper plus the workload that the subeditor is handling. And the subeditor is the one who decides to send it out to reviewers. How many? One, two, three, four? I'll show you in a moment. In the first time around, these are called also referees. Their reports in time trickle back to the subeditor, who will make a decision, an interim or full decision on the paper, and who will send that to the authors along with the anonymous reviewers' comments.
There is a whole other story of people have wondered about this review process and whether there could be alternatives. The authors revise the paper, who sends it to the subeditor. Now there is a big parenthesis because, what if that revision, what if the revision had to be so substantial that the paper had to be sent back to the reviewers again? So the cycle goes on. Or, conceivably, they could be sent to other reviewers.
But eventually, let us say that it is successful. And it is accepted. It could be rejected. You'll see some things in a moment that I'll show you. When it's accepted it gets sent to another technical editor, who manhandles or "womanhandles" the manuscript in various ways with technical points in mind. And then eventually proof and galleys are set. This today, because it is computer setting with modern layout programs, is a relatively small perturbation and not many mistakes will intervene between what you send-- and this is why the technical editors in the beginning are so concerned about what is there.
Not too many mistakes are made here. There are still layout problems, because layout, ultimately have to be done by humans. Layout means how the figures are cited relative to each other, where the tables and the figures occur, and so on. And that gives you proof and galleries in which you comment which go back to the technical editor, and finally the paper is published.
The process looks complicated, but it just flows naturally, in these things. Now I want to show you this through a case study. I'm a strong believer in case studies in these things. I want to show you this paper, which a coworker of mine, who's in the audience, Yuta Tsuji and I did. And it is a paper which was published in 2014, but we did the work in 2013.
So that's that paper that I showed you before. And this is how we submitted it. And we submitted it to the Journal of the American Chemical Society, one of the prime chemical journals. And here is the submission letter. So part of what you have to submit, it's a computer-based process. It's an experience the first time you do it. It becomes routine by the time you do it 400 times.
[LAUGHTER]
It takes about a half hour in an optimal situation if you have all the material prepared. If you're not prepared, it may take two or three hours. One of the things you have to prepare is to suggest some reviewers or referees. You can also suggest some people who you don't want to be reviewers. If you suggest 10 such people, a minus mark will appear next to your name in the mind of the editor, if not worse. So one should use that sparingly. But you need to provide the name, affiliation, email address of the reviewers.
And how you choose those is very interesting. Should you choose among them your friends and former students, if the editor should not know about, that they are friends and former students? Should you choose the experts? Obviously, you should choose the person-- what you're trying to do is to read the editor's mind-- the subeditor's mind. The subeditor's the one that knows. And there is no point not putting on there someone whom the subeditor, from knowing the field and reading your paper, would not put on. So you might as well put them on.
There is a submission letter, and you see that at left. It should say something positive about the work but should not overdo it. Overdoing it is going to cause another small minus mark in the mind of the editor. The editor knows what's going on. But it doesn't hurt to say why this work might be important.
So we did that. That was submitted in November 2013, November 20.
[MICROPHONE FEEDBACK]
Sorry. I must be giving some resonance. And the comments were received on December 17-- relatively quick review time of about three and a half weeks-- and the paper was rejected. Now I should tell you that the subeditor, Wes Borden, is a good friend of mine.
[LAUGHTER]
I was his TA when he was an undergraduate when I was in graduate school. So I've known him for 40 years. And we've written-- I and the editor-- have written four or five papers together. So we know each other. That's what happens when you're in a trade for a long time. Nevertheless, could he have done something else? I suspect because he sent us out to four reviewers, that probably he tried. What I mean by that is he probably got a bad review as the first one, and then he sent it to some others, hoping to get some good reviews.
Now I don't show you the actual reviews. I'll spare you those. Though I'll show you some others. But I'll just show you the recommendation. One says publish in JACS after minor revision. One says published in JACS without charge-- that sounds good-- without change. But then the third one says do not publish, and a fourth one says publish elsewhere.
[LAUGHTER]
But I want you to see what an editor who is a friend does with two positive and two negative reviews. And that is he cannot accept it. Now could he send it out to the fifth reviewer? I don't think there's any point. He did the right thing here on this.
Now what I want to show you is some of my best referees' comments, because I know in general, our interest is in other people suffering, in general. Or in prurient interest of other kinds. So here are some of the best reviewers' comments-- best in quotation marks, I'm sorry.
"The speculations in this paper are the sort of thing that one expects to hear at research seminars or in social chemical gathering. Certainly many of them have been made at my own seminar by bright young students. No one else, however, has had the conceit or effrontery to think them worth publishing." Wow.
OK, paper two. That was submitted to a chemical journal but was reviewed by a physicist.
"This paper would not be acceptable for publication in Physical Review. The authors should calculate the binding energy of the structure and compare it with graphite, not just propose it as a possible structure." This was a paper in which we proposed, the first in the literature, of metallic-- of many since then-- hypothetical metallic allotropes of carbon. And that's why he's saying this.
"This extended [INAUDIBLE] method contains errors. It is absolutely useless except for publishing papers in chemistry journals. You chemists should raise your standards."
[LAUGHTER]
I thought you'd like to see that.
The third reviewer, "I am not now and never have been an admirer of Hoffmann's efforts in the inorganic organometallic field. To a bridge player the, sideline kibbutzers--" OK, for those of you who are non-native speakers of English, to kibbutz, a verb or kibbutzing, is incursion of Yiddish into English, and is interesting, but you'll have to look it up in a dictionary to see what it means. But the general idea is someone who stands by-- someone's playing a chess game and someone's standing there, making comments to the players. Like you should have moved your rook there. That's a kibbutzer.
OK, so probably the third comments were written by the same person who wrote the first comment.
[LAUGHTER]
He also wrote something for my 10-year review, which Harold knows about. Anyway, I thought you'd see this as fun.
[LAUGHTER]
OK so my friend rejects the paper from JACS. What do I do? Actually, I'll give you some advice on what to do. But what do we do is we resubmitted it to another good journal, Angewandte Chemie. It got a review in January very quickly from three reviewers. A favorable decision, needing revision, and then here is my letter to the subeditor at Angewandte Chemie making my revisions.
And what I do in this is I answer the letter briefly, but then I, down here, paste together the three reviewers' comments in an organized form. They come in rather disorganized. Everyone has their own style, and they come in all kinds of fonts and-- not just funds but-- I assemble them and then I answer them in a recognizable way. And then I also say here, they asked me to reduce the paper so we did. I send to you a track changes marked up version of this paper so you can see clearly what was cut and what was not.
So now that comes to-- now I come to advice. That's the case story. The paper got published. It's a good paper. Responding to the reviews through the editor or subeditor. If his or her decision is a firm one that they reject the paper, as you saw, that was their decision in that first comment I got from JACS. Do not fight. Don't tell the editor-- under any circumstances, do not tell the editor how many of your friends like the paper. And especially don't tell them that a Nobel laureate liked the paper, anything like that. Just curse them under your breath, or out loud in a private room in your house.
[LAUGHTER]
And then go out and have an ice cream cone with a friend, or have a glass of whisky at home. And then the next day just thank the editor for the reviews and resubmit the paper elsewhere. If an editor leaves you any opening-- and you'll learn how to recognize this-- then you seize on it, you revise the paper.
But now things change. All of a sudden, the audience for that revision is the editor and the reviewers, if he should, god forbid, send it back to the reviewers for a second opinion. That may happen. The audience is there and you must do everything possible to reach that audience sensibly and to make it easy for them. So that's why I pasted the referees' comments in a neat document, why I answered them point by point. And what you must not do-- and the same applies to your marriage as to your relationship with the editor-- when someone says something you don't like, don't escalate.
[LAUGHTER]
That is just-- instead, you answer-- if you perceive anything to be wrong or aggressive, you answer it politely and do everything you can not escalate. The idea is obvious. The idea is to impress the editor with your professionalism and your willingness to take the comments in a friendly, professional manner. Anything you do that detracts from that is going to get the paper rejected. It's as simple as that. And it may have therapeutic value to you to get angry. But it doesn't serve any purpose at all in this process.
And this is why also, you want to make clear to the editor-- don't tell the editor I've cut the paper by 12% and cut it by 2%. The editor is not stupid. But you make it easy for them by actually pointing out the changes that you have made. And if you have made some changes on your own, which were not in the criticism, put those in too. It gives the editor the impression that you are a thinking human being who is able to revise a paper in different ways.
OK, so what's clear from what I say is that in writing papers, you keep your audiences firmly in mind. This we learn at Cornell by being teaching assistants in introductory chemistry or physics. That's where we learn. Our students are very intolerant if you don't keep them in mind. And they teach us. Editors, reviewers, readers are all audiences that we try to reach.
Now technical editors or something else. Technical editors have one interest in life, same as everyone else, it is how to make life easy for themselves. Which may be also what makes it easy for the journal. So they dominate the material that's put in the instructions to the author about how the paper is to be prepared.
So for instance, one of the things that you will see described by many journals is to put that text front, put the figures at the end, and put the figure captions on a separate page. That's not a human being talking. That's someone talking who wants an easy job of how to have the paper set in their format.
Now let me give you an example--
[MICROPHONE FEEDBACK]
Can someone turn down this just a little bit here? Does somebody know where to don't get the feedback?
Let me give you an example again. A case study. So this is a paper I got, a pre-print from a friend. And I have removed the first page to protect the guilty. So this is someone I know who wants me to read a paper of theirs. He's sending me essentially a pre-print. Here's page two and three, next two pages, next two pages, references, figure captions on the page at right, and then the figures.
So this person has composed the paper in the way that the technical editor has asked them to propose the paper. Now you are a human being, but who is the first human being who reads this paper? The editor, maybe goes over it a little bit, the subeditor. But the first human being who reads this paper after you submit it is the reviewer. So supposing the reviewer reads the paper, here is page two and three. I've eliminated page one.
One incidental piece of advice about writing papers. Do not do not wait to introduce a figure till page five in a paper. The natives like pictures, and the quicker you have a figure in a paper, especially if there are chemical structures involved, the more will people read the paper. So here, someone has--
[MICROPHONE FEEDBACK]
--waited to page four. Can you turn that down a little bit? To introduce a figure, and it was-- I put a little square on it. That's the first time a figure appears, a reference to figure one. So to find figure one, you have to go through the paper to the end. There you find figure one. But it doesn't have the caption, which shows you what the axes mean in figure one. To do that, you have to go back to another page. Now this is only four figures. Imagine there are 20 figures in this paper? This is not a way to make friends.
[LAUGHTER]
Now in today's day, a figure should appear, god knows, just where it's introduced, where it's called out. And it should have a caption so you can understand it. And with today's layout features, this is just in Word, not in anything complicated. This is not that complicated to do. So now here is the point. The technical editor of the journal tell you to do it in this way. But the editor is human, unlike the technical editor.
And in fact, the option when you submit usually carries with it the ability to, in addition to the manuscript file, which can be submitted often in Word or LaTeX format sometimes, you can submit it also PDF. And anyway, you have an option to upload a version in a PDF, and the editor, who is human, knows that the reviewers would prefer to read it in the normal way with the figures appearing where they are, and they'll send it out for review that way because they are more like you than the technical editor.
OK. Editors, in general, you should respect their tough job. You see here, never, never invoke authority. In general Americans don't do that, but people from other cultures sometimes think that by invoking the favorable opinion of some experts you will influence an editor. This is bizarre, and it will not work.
The reviewers. The reviewers are gatekeepers. When I am a reviewer, something switches on in my mind. The context is criticism. And this is-- incidentally, I'll come back to it-- one of the differences between our [INAUDIBLE] and reviewed papers. The context of a reviewer is to criticize. That leads to excess, of course, as you saw in some of those comments. But the idea is to find things that are wrong or that will help the reader.
In general, make the reading of the paper easy for the reviewers. Use large fonts. Do not cram in five pictures into one, even though you see in phys rev letters that those pictures are so crammed in. Of course, everyone can enlarge the pictures. But just like everyone can go to the back of the manuscript to find the figure captions, you are making it harder for that person to read the paper.
That's the basic idea. Wherever you can, make it easy. So figures especially, make the figures large. They can be combined at the end into smaller ones if need be, but not in that initial stage of your paper.
There is a special problem for non-native speakers, writers of English, with the universal language of science, which is broken English. But there are interesting questions of who is responsible for improving that English. For instance, do the editors of journals have a responsibility for improving the English in a paper? But one thing you can be sure is the paper will be sent out to the reviewers in whatever form you have it. It may be improved later on. So it's important to work on getting the English improved and getting friends to think about this.
Let's talk about journals, where to publish. So there is, established through time and practice by the people active in the field, there is established a pecking order of journals. Sometimes things are indistinct. It's certainly field-specific, subfield-specific within any science. So the pecking order for high energy nuclear physics will differ from that of condensed matter physics. And we can all remember some.
Here I picked inorganic chemistry. I'll say something about Science and Nature. But Journal of American Chemical Society, Angewandte Chemie, publish articles, the top journals in chemistry, publish articles across all fields of chemistry. But in particular they publish in inorganic chemistry, a good number of journals. After that, the two American Chemical Society journals of inorganic chemistry and organometallics are probably in the next rank. There is a European upstart that is getting better by the minute. That's European Journal of Inorganic Chemistry. And down at the bottom is Inorganica Chimica Acta
But the bottom is bottomless, as we'll see in a while, because there are new journals being founded all the time for no good reason. There is something to measure journals, which is an impact factor, which the journals are very concerned with and which they put on their mastheads or webpage.
So you take all the papers, scientific papers, but not editorial matter, that are published in a journal in years, let's say 2006 and 2007, and say you get 365 articles that are published in that rather small journal. The bigger journals will have in the thousands. And then you look in 2008, and all the papers that are published in the world anywhere, in their own journal or in other journals, and you add up all the citations and then you divide the second number by the first and you get the impact factor.
So it's the number of times papers in that journal published in 2006, 2007 have been cited in 2008. There are other definitions. There is a problem with these. One of the problems has to do with that most fundamental of human qualities, and that is laziness. So being lazy, people cite review journals much more than they cite original papers.
We're now close to something else, which Robert Merton has described as a obliteration citation, where the more something becomes familiar the less it becomes cited, perhaps. But basically anything that carries review articles is going to get cited more, because of this basic human quality, than anything that does not publish review articles. Angewandte Chemie, for instance, benefits from that.
There are also problems of fashion and such. There is a special problem with Science, Nature, Cell. Nature is now, of course, a machine. There is of a few nature journals, but let's talk about the original. These have acquired in their own fields-- Cell will not publish an inorganic chemistry article. These acquired a certain reputation out of proportion to their value. And there are problems with these journals, which those of us who have submitted to them encountered there. But anyone submits to them.
One is that there is an initial editor screening. The editors are probably a very good scientific background, but they haven't worked necessarily in all fields, and into their criterion for initial screening, unfortunately, comes something called newsworthiness. And that is intangible that's perceived by them. And it is a factor. And they may, essentially, they could call it triage, but they could reject some articles without them being reviewed. And this is very annoying if you're one of those who submitted one of those articles.
The other thing they do is heavy editing to emphasize newsworthiness. And the third thing they do which is bad is they put a limit on the number of end notes, typically like 30 for a Nature article. This is terrible because when you put that together with human propensity, my general advice to you would be, if you're faced with publishing in those, if you're at 35, if you have a choice between eliminating an end note to your article or to another person's article, eliminate it to your article. But that's not an advice that many people follow.
Some countries, universities, also place excessive importance on publication in these. In China people are paid extra money in bonuses for publication in this. And there is an interesting-- I still have to get hold of their ranking of the journals to get it, but the rewards are in real money, which is as tangible as you can get in this.
All I can tell you about this, it's possible to get a Nobel Prize without publishing a single article in Nature or Science. The person who introduced me is responsible entirely for the only article in Nature that I have published. And I never forgive myself for being pushed into that, because I had a good track record until then.
Something is changing. Something's changing under our feet. Something's changing very rapidly in your scientific lifetime.
First there is the question of open source publishing, which, again, we could talk about-- we could talk another time. There are being founded hundreds of open source journals, which are of poor quality, curiously not motivated-- though they pretend to be-- by the open source ethic of information being accessible to all, but motivated by the usual thing which motivates publishers, which is making money. And that's because the financing model shifts as you go from normal journals to open source journals.
There is a question of pre-print servers, and-- how is it pronounced? Is Paul here? Because I have-- is it archive or ArXiv? I don't-- I have--
Archive is the official. OK, we could schedule a nice little discussion between Paul Ginsberg and myself about this. Many chemists have problems with ArXiv. I don't have the problems because of that. But I have other problems, which I don't want to discuss here, with ArXiv. In general it's very popular in physics, mathematics, astronomy, but in chemistry it is not popular. In fact, a good number of chemical journals will not accept for publication a paper if it has been submitted to ArXiv. This is a decision by the editors of those journals. But behind it is something interesting.
The other thing that is at work now that is beginning are blogs. So the young people in the audience, for better or for worse, are in there with blogs. The interesting question is do blogs make a contribution to science? Being of a certain age, I would be skeptical. But I have now seen two blogs in chemistry, specifically in theoretical chemistry, by Henry Rzepa and by Bachrach, Steve Bachrach, I think, which are both good blogs. They actually, outside of the normal process, raise interesting questions.
And we've had something happen in chemistry, which I was a little involved in, of the following kind. Someone wrote a paper. Sason Shaik in Israel. A blog comment was made on the paper by Henry Rzepa. That blog comment was so perceptive that Rzepa was a co-author of the next paper that Sason wrote about that subject. And there is a third paper in which I joined them, but it came out of a blog. And there may be more things like this.
There is another whole question of access to data. This is something tied up with legal questions, NSF mandates, that we put in data management plans, and that everyone in the world should have access to even the raw data of doing the scientific research. That's still-- there'll be legal tests of this. It all came upon us as a result of the litigious of American society and should not have happened.
There is another interesting thing about archival storage. What should we do with those terabytes of material that we generate, that our computers generate?
I want to talk a little bit about how to key read and keep up with the literature. So here is the good old days. So you can see that this young man-- I'm proud that I am not 20 pounds heavier than that, and I can't do anything about the hair. But this is a photograph of me downstairs in the Clark Physical Sciences Library at a time when I could say, without any exaggeration, that we had the best library in the world. I can't say that today.
It has nothing to do with the librarians. It has everything to do with the money. I'm pointing to somewhere. Where's the money? It's over there, they hope. It has something to do with the money that goes into the libraries. And this was the old days.
So what would happen in the old days? The journals were all printed, of course. And they came to the library and they were put on the shelf, which was very important to me and to others who kept up with it. This is where they were put for the week that they came in. After that week, they were taken off this shelf-- shelf had two sides. It had room, when it got crowded, for about 100 journals on there.
After this one week-- I'm telling you this in some detail because most of you don't remember how it was. But after that week it was put onto another set of shelves for the whole year. And at the end of the year, the whole year's number of issues was sent to a binder, which bound them in a book form and then they went on the shelves in the library.
So there were three stages in this process. But this was the important stage for me. When I was at the very beginning of my career I would go here every day. After a while I got busy and I started traveling, but I still tried to get here on a Saturday or Sunday. Now why on a Saturday or Sunday? In part, it's my work ethic, but in part, it's because on Monday they would be taken away from this shelf and put on the other shelf. And if I came on Saturday or Sunday I could be sure that I have seen everything that's been published that came into the library that week. That means I could keep up with all the literature.
Of course, I didn't look at all the journals, because they at that time I wasn't as interested in physics as I am today. I didn't look at many of the physics journals and I didn't look at the relatively few astronomy journals we had in there. And we were spared the biology literature, which was in Mann Library elsewhere.
But I was obsessive about keeping up with the literature. One of the reasons I was obsessive was this was the new. And science is the cult of the new. This was the new stuff that was coming in.
The second thing was it kept me ahead of my graduate students and postdocs. Let me tell you something here. This is psychological. If a research director wants to make an impression of a work ethic on his or her students, one thing, of course, is to be in their office all the time. That helps. But there is, in terms of a single action, what has probably the greatest impact is when some graduate student or postdoc has been working on some specific subject and the research director sees in the literature a new article, not necessarily scooping the people, but relevant to that field of research. And if he sends a little email-- it wasn't an email then-- to the student, saying, hey, I just saw this in the literature. It's something the student should have seen but they didn't. If a research director does this two or three times, they've got that student.
[LAUGHTER]
Because it's obvious why. It's obvious to the student that the research director cares about this. It's just ordinary human nature. So this is another reason for keeping up with the library.
OK, now the most important thing was I even built in a redundancy. That is, somehow I thought I might have missed some of these. So I looked for something which would have the interesting papers or all of the papers in another form than the hard copy that came in the library. Some of you of a certain age will remember something called Current Contents. That was a print publication, which published the tables of contents-- not the abstracts-- of every journal in the world that came in to a certain office. It's been taken over, it became Citation Index, and is now the Web of Science. There is a line for those things.
I read Current Contents. Current Contents provided what I call, in general, is a useful principle for keeping up with the literature, which is what I call optimal redundancy. So if you do things normally, by the nature of things, you're going to miss something. It just happens. So you build in redundancy. Something where you have another way of looking at the thing.
Now some people are so uptight about-- relatively few-- about missing things that they do too much of this. So optimal redundancy is two or two and a half or something like that. That's what I found with time.
So now you see my problem. That print issues have disappeared. The shelf has disappeared. Maybe it's still there. It's gone. I wish I had that piece of furniture. It's gone, and the journals are on the web. That's fine, I can access them. And but the problem is that I don't know what I've read. I'm getting older I forget what I've read. So how do I build a system which, given that I have to read a lot, that I don't repeat myself? After a while the titles of the papers begin to swim before your eyes. They're all the same.
So what I've done, I managed to find it in Google Reader, but what I use are something called RSS aggregators, which you will know better what that means than I. And I used to use Google Reader, but in the way that Google has of abolishing anything that it can't make any money from, it abolished Google Reader. And so now there are other places. And there is something called Feedly, there are other aggregators. I use one called the Old Reader, which, as the name implies, tries to essentially emulate the environment of Google Reader.
So I get journals and I get a feed of the titles and abstracts of every journal that I want, and I put them in. You see a list of the journals over here, Chemical Physics Letters, Chemical Reviews, Chemical Science. And I have it so set up here, I'm looking at the title and the abstract and the graphical abstract, too. But I could set it up that I could just see the title. And as I scroll past it-- which may or may not mean that I've read it-- but let's be nice to me. Let's say as I read the title and look at the abstract-- graphical abstracts are a wonderful invention. There is an interesting story about where they came in in Science. But as I read this it goes from boldface to plain font. So it tells me and the number goes down.
So what this means is in Dalton Transactions there are 18 articles I haven't seen. And as I open Dalton Transactions and as I go down this, then that number decreases. So this tells me what I've read. It manages the same thing, but it still doesn't have the feel of a paper. So actually, a few more words about this. Let me tell you the scope of where I am at. And this is not meant to impress you.
I'll tell you in just a moment a piece of advice at 600 papers. In a number of ways I've been at it. Recently Neil and I published four papers in Journal of Chemical Physics two years ago. And the editor reminded me that I had first published my first two papers in that journal 50 years before. So Ben can beat me on this, but it's a nice feeling when that happens.
Anyway, what I have on here is about 100 journals. They generate around 1,500 new articles per week, which I look at on a weekend. I look at the titles of those. I look at the abstracts of about 50. I look inside the paper of about 10. And I print out or save about five. So that's the process of weaning.
Now I know that-- and I tell you that that process takes me the same time that going to the library did, in the old library, and that is two to three hours a week, this process of keeping up with the literature. What's very nice with a computer, which was not true of the library, is that process can be fitted into the nooks and crannies that a modern day existence allows.
What I mean by that is if you're stuck in a doctor's office for an hour, you can do this. And so that this is very different, whereas you couldn't do it before. So actually, it goes-- now I know it's intimidating, 100 journals and 1,500 articles. But it took me 52 years to get there. So give yourself that time.
So here is the practical advice when you're starting out this. You already have, if you've written one paper, you have from the references in that paper, even though for you, it looks like the world began yesterday and there's probably not a reference more than 10 years old in that, doesn't matter. Probably you have in there five references to JACS or Angewandte Chemie, if you're a chemist. If you're in a biological world, you would have other references. If you are in physics, you would have Phys Rev Letters, Phys Rev B, and other things. So put those on that list.
After that, follow a simple practical test. If you find yourself looking up a journal to get a reference for whatever reason in the world-- your professor has made you do it, you're doing a literature search for a paper-- if you find yourself looking up a journal 10 times in the normal way, which means typing in the name of the journal in Google, getting to the website of the journal, then putting in the reference that somebody gave you and getting the paper, if you do that 10 times and you notice the same journal, add that journal to this list. That's all. Just a practical test of utility to you.
This is a very different philosophy than the philosophy of a targeted search, which is the problem for both journals and for ArXiv. If you put in that you want to look at papers on high temperature superconductors-- this doesn't matter whether it's ArXiv or journal literature-- you will get 2,000 references. So you refine the search, and it turns out that you're working on cuprate superconductors which involve bismuth. So you ask it to supply you with references of cuprate superconduction, and, my god, it goes down to five, and you like that.
OK. So you can cover the same ground by a multitude of-- you can gain the redundancy by using a multitude of search topics. But given human nature, we tend to focus on very precise areas not because that's what we should be doing or need to be doing, it's simply to get the number of hits on a search down to a manageable level. So getting that down to a manageable level-- I'm talking about how to define search terms-- is what forces you into a situation where you become too specialized and you will only see articles in that field. And that is important at a certain stage, but it's not a good thing overall.
There are many more things to talk about. How to search the literature. SciFinder. How to devise optimum searching strategies. And our librarians, Leah McEwan and Diane Dietrich, are here to help us with that in a number of ways. SciFinder, Web Science, Web of Science, and databases each have their utility.
There is the interesting question of citation practice, which, what is the optimum? You cannot cite everything. That's too much. But there are real problems in cit-- I'm writing an article about this, which I'll have ready in a while.
There are other things. How do you behave? I told you how to behave with the editors. Don't get angry. How do you behave when you find that someone else has done the work that you've been slaving to do? That's an interesting psychological problem.
There is plagiarism. Like the sexual misdeeds of our religious ministers, of much attraction to the general public, but not overall a serious problem for good science, I think. But it is something which the journals are very concerned with.
Then I haven't told you anything about how to behave as a reviewer of a paper, which is still another subject which is worth discussion. But I think we have to-- and even though Neil said I would talk about how to write a better scientific paper, I haven't done that.
There are many theories of this, but here, as in many other things, a virtue theory is not a bad idea. That is, you write scientific papers by modeling yourself on papers which are good papers. And try to figure out what it is that is good about them. It's very hard in the beginning. There are many things pulling at you when you're writing a scientific paper. I could talk about that another time. But I think it's enough, so we stop here. Thank you.
[APPLAUSE]
Nobel laureate Roald Hoffmann gives a personal reflection, with practical advice, on the publication process in science, including submission letters, anonymous review, dealing with editors, and the inevitable revision of a manuscript. Recorded Feb. 5, 2015 as part of the Career Advancement Program for Engineers and Scientists.
