Hacker News new | past | comments | ask | show | jobs | submit login
Guidelines for keeping a laboratory notebook (rice.edu)
87 points by Tomte 4 months ago | hide | past | web | favorite | 23 comments

My paper notebooks were always horrific. Writing has always been painful and awkward for me. But I could type like nobody's business thanks to programming. I survived college in the early 80s by being one of the first students to get a word processor.

By the time I was keeping a notebook, my work was generating mountains of computer readable data, source code, and so forth. We managed by agreeing on a format for data files, where the filename referenced a notebook page, and it worked OK.

Today, it's unavoidable that people are going to keep their notes electronically, and there are no perfect solutions for doing this. Wet chemists still like paper notebooks, since it's hard to get a computer close to the bench, and to type while wearing rubber gloves. Academic workers are expected to supply their own computers, and are nervous about getting them damaged or contaminated. Plus, drawing pictures and writing equations on a computer are both awkward.

Computation related fields lend themselves well to purely electronic notebooks, no surprise. Today, a lot of my work fits perfectly in a Jupyter notebook.

Commercial notebook software exists, but it tends to be sold largely for enterprise use, i.e., the solution it solves is how to control lab workers and secure their results, not how to enable independent, creative work.

Professional lab researchers use physical notebooks because they are lo-tech and immutable. A pre-numbered page in a bound book can't be torn out surreptitiously. Pencil and white-out are disallowed. They're simple and tamper-evident, which is very important if you're designing pharmaceuticals, for instance.

Obviously there's crypto, signing, checksums, and so on, that you can use on computers. But when the answer is as simple as writing in pen in a pre-numbered, bound notebook, and having your supervisor sign your notes, might as well just do that.

Definitely good points. And every field has its own level of security risk. In my case, the greatest risk is me not being able to figure out what I did in the past. I'm not in a tightly regulated / audited industry.

It is not only regulation. There might be a legal fight over a discovery and for this there's nothing like a tamper-evident paper document to prove that a lab did X on date Y.

> Computation related fields lend themselves well to purely electronic notebooks, no surprise. Today, a lot of my work fits perfectly in a Jupyter notebook.

Some notes and ideas regarding Jupyter notebooks as lab notebooks from "Keeping a Lab Notebook [pdf]": https://news.ycombinator.com/item?id=15710815

A chemist friend of mine has handwriting that is illegible even to him, and his solution is a high quality dictaphone. He used to have it transcribed, but today he just hooks it up to a computer and lets it do the work.

This strikes me as a very weird document, it contains some good advice about keeping a detailed and accurate record. It links to a cool story about why this is important (http://www.ruf.rice.edu/~bioslabs/tools/notebook/ringerstory...).

Then separately it has the whole thing that reads like it's about a protocol to prove you aren't committing fraud. Except it's hard to imagine circumstances outside of undergraduate classes where anyone is going to read your notebook trying to convict you of committing fraud. It's also hard to see how this advice helps because it just makes fraud slightly harder, now instead of editing in place you have to edit by copying things over to a new notebook.

I was at LBNL while the element 118 crisis happened. While ultimately it was figured out using VMS timestamps, well kept lab notebooks helped clear a lot of people from potential guilt.

Considering how bad the replication crisis is, if our society ever gets its shit together, many of these people should actually be convicted of fraud or at least removed from positions where they could commit further frauds.

Is there any good material to read about the forensics that went into the events surrounding the element 118 fraud? It sounds interesting.

Possible, but I doubt it would tell the actual story. I knew two of the coauthors in the Ninov paper (one of my closest friends) and two of the guys who did the computer forensics as it turned out, so I kind of had an inside view.

Anyway, use a paper notebook. It's worth the trouble for lots of reasons. Also use analog chart recorders if you can.

It's important to realize that lab notebooks (in professonal academic research labs) are quasi-legal documents with specific requirements, in that they must document every modification made to an experimental setup, and every measurement made. It's similar to a ship's log (I presume, not being in the shipping business). This is not about keeping a personal notebook, although there's nothing wrong with applying the rigorous standards of a lab notebook to personal notes. "Lab notebooks" in student lab courses are meant to familiarize students with the practices of genuine lab notebooks, but are obviously much less stringent.

plaigarism was egregiously rampant, in my undergrad experience. A handwritten report, experimental notebook, and pre-precis of intended manuscript was mandatory. The blatant xerox copy pasta that occurred, right down to the whiteout over the original authors PII and the toner banding was astonishing. There were several incidents were the entire assignment had to be redone by the entire class, with variations per individual, along with the stated expectation, that this was to be a solo effort. The instructors and the profs had to wrestle with this at personal expense under threat of having the entire year of subject disqualified, for everyone. Anti fraud mechanisms are important for undergrad as an enforcement of ethical habits, ive went through it first hand.

This is (or atleast historically was) important (or perceived to be) in industry labs to prove discovery for intellectual property purposes.

With the US patent system moving from “first to invent” (notebook timestamps become determining) to “first to file”, this is much less of an issue now (with reapect to IP provenance).

In the U.S. patent system, priority now goes to the first inventor to file. If Alice files before Bob, there's always the theoretical possibility that Alice derived her invention from Bob's invention and thereby wasn't really an inventor. In that situation, Bob might try to prove this, so as to be able to (re)claim priority and thereby be the one who is issued a patent instead of Alice. If Alice has a timestamped notebook, that can be powerful evidence of her independent development of the invention, which will help her fend off Bob's claim to priority.

(Much the same is true if Carol discloses confidential information to Dave under an NDA, and Dave realizes that his team had previously known the information and so the information shouldn't be subject to the NDA: If Dave's team has contemporaneous written records of their prior knowledge or independent development of the information, it can really help their case if Carol sues them for breach of the NDA.)

I've found this [1] other tutorial as a great basis to keep a good lab notebook.I love how he tried a cromatography of different pens to find something that can survive any accident.

[1] https://colinpurrington.com/tips/lab-notebooks

No mention on pen or pencil.

People tend to hate pencils. I love them.

They're more permanent and chemically stable than ink - unless, or course - the eraser monster hides around your book.

> Please use a ball point pen for all entries, so that the marks will not smear nor will they be erasable.

They just give this. But, there is a lot of interesting discussion about different inks vs pencil for archiving things that I would've loved to read.

Something i was told was to use the lefthand page for rough work, it [leftpage] was to be considered an appendix or attachment of sort and was not considered for scrutiny. it was highly advisable to use pencil on the left page in wet solvent conditions [water, alchohol, benzene etc.]

The right hand side of the note book was a final and standing statement of your notes and was to be indelible ink with bracket and single cross through in the case of transcription errors. the right hand page was the "final answer"

Top of the document says use a ballpoint pen, which doesn't smear and can't be erased like a pencil.

Try some gel ink pens. I'll concede they are awkward to write with without staining the paper with your hand, but they resist almost any common solvent, and can't be erased by cherrypickers.

I've been using the big grid+ offering from Vela Sciences for my daily work. Only use pencil though, Graph Gear 1000. Recently switched from 0.5mm to 0.9mm. Have only been keeping "good" notes for a few years now and so far pencil hasn't been a problem. Was just rereading Jan 2017 this morning and even in 0.5mm, looks as good as new. I'm not in a real lab setting though so I just go with what feels good. If I was forced to use ink, I probably wouldn't take as many notes.

The classic book on this topic:

Notebooks of the Mind: Explorations of Thinking https://www.amazon.com/dp/0195108965

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact