Common Lisp

SBCL come with very few built-in package.

To install additional ones, use asdf, e.g.

(require 'asdf)
(require 'asdf-install)
(asdf-install:install 'cl-ppcre)

will install Perl-style regular expression support. Then, when you restart SBCL, you'll need to load it with

(asdf:operate 'asdf:load-op :cl-ppcre)

To get a full-featured version of CL Hyperspec in Emacs

Look at dpans2texi.el.


To start Clojure REPL

We need to create a small startup script, say clj. The simplest one is something like:

#! /usr/bin/env bash
java -cp /usr/local/share/clojure/clojure.jar clojure.main

The problem is that we won't benefit from command-line history, backward search, etc. For that to be effective, we need to use JLine (rlwrap doesn't work on OS X, due to the lacking Readline library). Then, we just have to modify the above script with:

java -cp /.../jline-0.9.94.jar:/.../clojure.jar jline.ConsoleRunner clojure.main

where ... denotes actual path to related files. Alternatively, this could be setup using CLASSPATH, or by placing the jar file in the system-wide java extensions directory, e.g. /Library/Java/Extensions or /System/Library/Java/Extensions. The latter doesn't seem to be working as jline would be loaded by the extension classloader which is unaware of class path variable. Startup tips are described in http://clojure.org/getting_started, and on http://www.stackoverflow.com.

In addition to the core distribution

It is recommended to install clojure.contrib which is a set of additional features (like common mathematical operations) that are not present in the base package. To install it, just grab a binary (jar) distribution and make it available for Java. For that purpose, we can use a custom CLASSPATH, e.g. in my .profile:


or add an extra argument to -cp when invoking clojure.jar from Java.

To check that everything works well, we can use a simple script test.clj, like this:

(ns test
  (:use clojure.contrib.math
(deftest foo
  (is (= 4 (sqrt 16))))

that can be run under Clojure REPL. If it is ok, we should see something like:

Ran 1 tests containing 1 assertions.
    0 failures, 0 errors.
    {:type :summary, :test 1, :pass 1, :fail 0, :error 0}

Here is another example:

(ns test
  (:use clojure.contrib.pprint))
(pprint (for [x (range 10)] (range x)))

A better way to get a working clj starting script on OS X

is described at http://bit.ly/eda9IS. It is possible to add as many libraries as we want (e.g., incanter).

To facilitate editing and stay within our familiar Emacs

environment, we can use slime. A specific port for Aquamacs is available for download.

There's a difference between :use and :require

JVisualVM (jvisulavm comes with Java SDK) might help seeing

how much Java memory, cpu and threads Clojure actually uses. h/t http://www.fatvat.co.uk/2009/05/jvisualvm-and-clojure.html

A good set of exercices to translate from Lisp to Clojure:

L-99 – Ninety-Nine Lisp Problems. There is also functional-koans on Github.

Getting help

(doc histogram) brings a short description of the function with optional parameters (no pager though).

A bunch of documentation and resources is available from Mark


To activate folding mode for code blocks

M-1 C-x $; to revert, C-x $ (See set-selective-display)

To test whether we are running Aquamacs (or any other Emacs flavor),


(when (featurep 'aquamacs) ... )

On OS X, this just has to be negated to address built-in emacs.

Use M-x describe-bindings (or C-h b) to see

all shortcuts for current buffer.

Something that is lacking in most editors

is the possibility of creating temp buffer on the fly. E.g., C-x b (aka switch-to-buffer) allows to switch to a temporary buffer.

Managing multiple windows is a matter of 6 commands:

C-x 1 = delete-other-windows, C-x 2 = split-window-vertically, C-x 3 = split-window-horizontally, C-x + = balance-windows, C-x o = other-window, C-x C-b = list-buffers. Use M-x describe-bindings to view all the Buffer-menu key bindings.

To remove all distractive UI stuff, just put in your .emacs:

(if (fboundp 'scroll-bar-mode) (scroll-bar-mode -1))
(if (fboundp 'tool-bar-mode) (tool-bar-mode -1))
(if (fboundp 'menu-bar-mode) (menu-bar-mode -1))

C-c ( allows to navigate between references

e.g. equations, figures, etc. (With reftex-mode enabled, it is C-=.)

Spell-checking is done with ispell

Under OS X, an easy replacement of ispell is done with aspell. You need to populate /usr/local/lib/aspell-0.60/ with compiled dictionnaries. To check which dictionnaries are available on your system, type aspell dicts. A useful replacement of ispell-mode is the interactive flyspell-mode. To use it with LaTeX, just add in your .emacs file:

(add-hook ‘LaTeX-mode-hook ‘flyspell-mode)

To repeat a specific pattern or character

just use C-u nn pp, where nn is the number of times character pp as to be repeated. For example, C-u 78 - put a pseudo line of length 78; it is useful to choose this length as the default fill column width. To check which one is currently defined, type fill-column in the scratch buffer and evaluate this expression (C-x C-e). Common values are less than 80, e.g. 68, 70, or 78.

An 'annoying' (?) feature of ESS

is that R comment (#) are automagically tabbed to the right (40th column by default). To get ride of this, you can:

  • use ## instead, which yields comments aligned to the current level of indentation;
  • use ### instead, which results in comments aligned to the beginning of the line.

See http://ess.r-project.org/Manual/ess.html#Indenting. Another option is to put (setq ess-fancy-comments nil) in your .emacs file.

The auto-fill-mode also works with comments

Just M-q (fill-paragraph) on the region of interest. Another option is to use newcomment.el which provides comment-auto-fill-only-comments.

(require 'newcomment)
(setq comment-auto-fill-only-comments 1)
(setq-default auto-fill-function 'do-auto-fill)

h/t http://stackoverflow.com/questions/2214199.

Ido and Smex are two useful add-ons to Emacs

They help mastering buffers and provide enhanced completion.

ELPA stands for Emacs Lisp Package Archive

It provides package.el which acts as a package manager for Emacs based on a repository of pre-packed Emacs Lisp code. http://tromey.com/elpa/

To get ride of the automatic indentation of comments in ESS mode

upon pressing the Enter key or after C-M-\, use two or three '#'. E.g.,

# this is a comment that will be right-justified
## This one will not move

To check what a combination of keys stands for

just type C-h k followed by the combination.

Don't forget the very handy markdown-mode

Useful keybinding for ESS/Roxygen

C-c C-o (inside function body) generate help template

To ease the navigation across tab

use tabbar or remap F5 (not used on Air OS X), e.g.:

(global-set-key (kbd "<f5>")   'tabbar-forward-tab)
(global-set-key (kbd "C-<f5>") 'tabbar-backward-tab)

To view the current buffer in Safari

just use M-x browse-url-of-buffer

To open a file as root

just C-x C-f and type /sudo::/path/file

Interesting navigation keys

C-@ set mark
C-x C-x exchange point and mark
C-x n n narrow down to the specified region (between mark and point)
C-x n w revert to full buffer editing
M-< go to the top of the buffer
M-> go to the end of the buffer
C-l center buffer
C-v scroll one page forward
M-v scroll one page backward

Try word wrap with M-x longlines-mode.

Although Emacs highlights closing parenthesis (short blink)

you might prefer M-x show-paren-mode

To customize a given font face

just C-u C-x = to get its name.

In dired mode, instead of pressing Enter

use a: it won't open a new buffer for each directory you visit.

To get a nice overview of some pattern matching in a source file

M-x occur

M-I convert word to lowercase (see also M-x downcase-region).

In ESS mode,

C-c C-z will switch to the end of ess inferior buffer.

Use M-x remember to save quick notes, referencing the current buffer.

See http://stackoverflow.com/a/8920373/420055 to correct weird characters

in Emacs term or ansi-term when using zsh. (See also http://stackoverflow.com/a/7437783/420055 for git prompt and unicode characters).

To add a specific entry to the Info index page, check that your custom

Info dir are up to date, C-h v Info-directory-list, and use install-info to update the Info dir (linking or moving an Info file is not enough). E.g., we can put mu4e.info in /usr/local/share/info and use

$ sudo install-info /usr/local/share/info/mu4e.info /usr/local/share/info/dir


Somes Notes on Internal Implementation

lot of useful tricks to know about how Mathematica actually handle some of its calculus. Available at http://reference.wolfram.com/mathematica/tutorial/ (SomeNotesOnInternalImplementation.html)


To boot in 64-bits, just run in a Terminal

$ sudo systemsetup -setkernelbootarchitecture x86_64

To change Dock appareance (switch from 3D to 2D theme),

$ defaults write com.apple.dock no-glass -boolean YES
$ killall Dock

To revert, use -boolean NO.

There are three configuration files that can be used to customize the shell

(default to Bash): .bashrc, .bash_profile, and .profile. The first two are for non-interactive and interactive shell, respectively. The latter one is loaded when no .bash_profile exists, but its existence will supercedes .profile. The .bashprofile seems to be the recommended one (at least, according to Apple support). Some authors suggest to fill the .bash_profile with something like

if [ -f ~/.bashrc ]; then
    source ~/.bashrc

so that everything goes into .bashrc. There are also two system-wide configuration files that are located under /etc: /etc/profile and /etc/bashrc.

Apple's computers don't come with a true Readline library.

Instead, libedit is used a replacement, which might cause some troubles with applications relying on console. If necessary, it can be installed with a few steps:

$ curl -O ftp://ftp.cwru.edu/pub/bash/readline-6.1.tar.gz
$ tar xvf readline-6.1.tar.gz && cd readline-6.1
$ ./configure && make
$ sudo make install

For Python, it is not really a problem as we can simply do

$ sudo easy_install readline

and use the excellent ipython or bpython interactive shells.

Adding colors in the Terminal can be done

in (at least) two ways. First, we can just set an alias to ls, e.g. alias ls'ls -G'= (=ls --color=auto= don't under OS X, but we can use export LS_OPTIONS'–color=auto'= instead). Another option is to use CLI colors, that is:

export CLICOLOR=1
export LSCOLORS=dxfxcxdxbxegedabagacad

The order matters: DIR, SYMLINK, SOCKET, PIPE, EXE, BLOCKSP, CHARSP, EXESUID, EXEGUID, DIRSTICKY, DIRWOSTICKY. Color codes read as follows: a = black, b = red, c = green, d = brown, e = blue, f = magenta, g = cyan, h = light gray, x = default.

Parameter expansion is useful when we want to remove file extensions

in shell scripts, e.g. ${file%.*}. More generally, ${varname%pattern} allows to remove pattern at the end of a variable. h/t Jake Hofman

${!#} gives the last command line argument

and !! gives the arguments of the previously-executed command.

Instead of playing with diff

use comm (or grep -xF) to check for lines common to two files. h/t Jake Hofman

To make a given file executable, use chmod +x or chmod 755.

About file permission

we often find root (user) and wheel or admin (group). In fact, wheel is the system administrator group in BSD, while root is the system administrator user. It is common to add sudo permissions to users in the wheel group. The choice of wheel comes from the BSD UNIX group at UC Berkeley; "big wheel" means an influential or important person.

To go back home, ⇧⌘H in the Finder.

Some useful search features in Gmail

has:attachment pdf after:2011/01/01 before:2011/01/20
has:attachment (*.ppt || *.pptx || *.pps)
label:Label1 | label:Label2 (case insensitive)
label:Label1 filename:pdf
from:John OR from:Joe cc: bcc:
subject:"match this exactly"
from:John (keyword1 OR keyword2) (remove OR to match both keys)
keyword1 -keyword2 (has key1 but not key2)
in:anywhere (Spam and Trash excluded)

No solution for filtering large attachment, though. But see http://www.findbigmail.com or http://www.searchgmailbysize.com.

To know if you are running the 'new awk' (nawk)

just try to add a custom function in an awk script. E.g., from the gawk manual

function changeit(array, ind, nvalue)
    array[ind] = nvalue

    a[1] = 1; a[2] = 2; a[3] = 3
    changeit(a, 2, "two")
    printf "a[1] = %s, a[2] = %s, a[3] = %s\n",
	a[1], a[2], a[3]

and test it:

$ awk -f test_awk.awk

To view the latest files added to a directory

use ls like this:

$ ls -l -t /usr/bin | tail

The -c switch allows to view file sorted by modification date instead.

Instead of creating an animated GIF with ImageMagick


$ convert -delay 20 -loop 0 fig*.png spin_plot.gif

let's try to generate an SWF file thanks to the swftools:

$ png2swf -r 2 fig*.png -o afig.swf

We can enable summarize function for text

in Sytem Preferences > Keyboard > Services. Useful for counting words in an abstract.

mv file{,-old} will rename all file to sth like file-old

Set intersection between two files

$ sort file1 file2 | uniq -d h/t

h/t @mathieuen, 25 Mar

Use pkgutil –pkgs

to list all installed packages (with Leopard installer). Useful command include pkgutil --forget org.r-project.R.Leopard.fw.pkg when you want to install two concurrent versions of R.

To delete specific pages in Apercu.app

select them and Shift-Command Del.

The fmt program is a simple text formatter

which allows to insert hard line breaks in a whole file according to some fixed line width (but respecting word breaks).

To count how many rows have at least one empty field

$ grep ,, file.csv | wc -l

To show filesystems in human readable column format

and order by percentage used:

$ df -Ph | column -t | sort -n -k5

h/t climagic

To see recent files but just print the last 30 lines

$ ls -ltra | tail -n 30

h/t climagic

To delete any files over 30 days olds in /dir and below

$ find /dir -type f -a -mtime +30 -delete

h/t climagic

To indent every line by two spaces

$ ls -l | sed 's/^/  /'

h/t climagic

Simply change file extensions from ex1 to ex2

$ for f in *.ex1 ; do mv $f $( basename $f .ex1 ).ex2 ; done

Assuming you just ran an ssh command

then ping !$:s/user@// will ping the IP itself, removing the user@ part. h/t climagic

The new OS X Lion added a lot of UI cleaning

but the home Library folder is not visible anymore. To bring it back in the Finder, just run

$ chflags nohidden ~/Library

To find words that end in 'da'

$ grep "da$" /usr/share/dict/words

h/t climagic

To reformat plain text file

$ fmt -ut -w 80 filename.txt

h/t climagic

To show directories containing files older than 365 days

$ find . -mtime +365  | while read -r file ; do dirname "$file" ; done | uniq

h/t climagic

To be notified by email when a job is finished

$ C-z bg ; wait %1 ; echo "done" | mail -s "done" you@example.com

h/t climagic

How much a file will be compressed using gzip?

$ gzip -v9 -c file > /dev/null

To find all files larger than a Gb and report how large each one is

$ find / -type f -size +1G | xargs ls -lh > listofbigfiles.txt

h/t rmounce

Uninstalling Applications is quite easy

but for Libraries it is often harder to tell what's been installed by a pkg. While events are logged in in /var/log/install.log, we won't find useful information there.

Solution 1 (easy):

  • Run the pkg installer
  • Look into "Show Files" just after installer has been launched

Solution 2 (harder):

  • In /Library/Receipts there is everything we need, including an InstallHistory.plist where all packages are registered.
  • Otherwise, you can just look at
$ lsbom /Library/Receipts/package_name/Contents/Archive.bom

to get the list of files (with relative paths) and Contents/Info.plist to get the root directory (IFPkgFlagDefaultLocation).

To generate Makefile with Qt on OS X

$ qmake -project
$ qmake -spec macx-g++

To know if there are current users logged via ssh

$ lsof -i :ssh

In place of *nix `find` command

we can use

$ mdfind -name "proqol scoring"

Print list of folders in home directory, sorted by size:

$ du -d 1 -h ~ | sort -n -r

If you want to free up inactive RAM

just issue at the terminal $ purge. See also: http://apple.stackexchange.com/q/67031.

The tmutil command is useful to control Time Machine backups from the

Hold over the mouse over a word

and type Ctrl-Cmd-d to get its definition from the built-in dictionnary.

With Numbers

there's no TRANSPOSE() command, but we can use the following:

=INDEX(Table 1::$A$1:$A$10,COLUMN(),ROW())

The mdls utility

allows to extract any metadata from a given file, including EXIF and geolocalisation information from an image file.


Nice use of itertools.groupby

for Fasta files (https://www.biostars.org/p/225913/):

from itertools import groupby

def transcribe(sequence):
    return sequence.replace('T', 'U')

def translate_rna(s):
    codon2aa = {"AAA":"K", "AAC":"N", "AAG":"K", "AAU":"N",
		"ACA":"T", "ACC":"T", "ACG":"T", "ACU":"T",
		"AGA":"R", "AGC":"S", "AGG":"R", "AGU":"S",
		"AUA":"I", "AUC":"I", "AUG":"M", "AUU":"I",

		"CAA":"Q", "CAC":"H", "CAG":"Q", "CAU":"H",
		"CCA":"P", "CCC":"P", "CCG":"P", "CCU":"P",
		"CGA":"R", "CGC":"R", "CGG":"R", "CGU":"R",
		"CUA":"L", "CUC":"L", "CUG":"L", "CUU":"L",

		"GAA":"E", "GAC":"D", "GAG":"E", "GAU":"D",
		"GCA":"A", "GCC":"A", "GCG":"A", "GCU":"A",
		"GGA":"G", "GGC":"G", "GGG":"G", "GGU":"G",
		"GUA":"V", "GUC":"V", "GUG":"V", "GUU":"V",

		"UAA":"_", "UAC":"Y", "UAG":"_", "UAU":"T",
		"UCA":"S", "UCC":"S", "UCG":"S", "UCU":"S",
		"UGA":"_", "UGC":"C", "UGG":"W", "UGU":"C",
		"UUA":"L", "UUC":"F", "UUG":"L", "UUU":"F"}

    l = [codon2aa.get(s[n:n+3], 'X') for n in range(0, len(s), 3)]
    return "".join(l)

with open('dna.fasta') as h:
    faiter = (x[1] for x in groupby(h, lambda l: l[0] == ">"))
    for header in faiter:
	header = next(header)[1:].strip()
	seq = "".join(s.strip() for s in next(faiter))

It is possible to evaluate double inequalities in Python

The Python challenge is a series of puzzles

that can be solved with few lines: http://www.pythonchallenge.com

Easy formatting of date

year,month,day = time.strftime('%Y-%m-%d',time.localtime()).split('-')

h/t neilkod

Enumerate and generators

Nice trick when using enumerate:

next(i for i,v in enumerate(l) if is_odd(v))


regex.match vs. regex.search

match only finds a pattern at the beginning of the string; if you need to scan all the string, just use search or regex.match("(.*?*)pattern(.*?*)") eventually.

See also this SO post:

import random
import re
import string
import time

LIST_SIZE = 1000000

def generate_word():
    word = [random.choice(string.ascii_lowercase) for _ in range(LENGTH)]
    word = ''.join(word)
    return word

wordlist = [generate_word() for _ in range(LIST_SIZE)]

start = time.time()
[re.search('python', word) for word in wordlist]
print('search:', time.time() - start)

start = time.time()
[re.match('(.*?)python(.*?)', word) for word in wordlist]
print('match:', time.time() - start)


I always create small dataframe for illustrations

X <- replicate(2, rnorm(100))
y <- X[,1]+X[,2]+rnorm(100)
df <- data.frame(y=y, X=X)

Here is a one-line solution:

df <- transform(X <- as.data.frame(replicate(2, rnorm(100))), y = V1+V2+rnorm(100))

This is also a nice way to generate two uncorrelated predictors, while allowing a strong association between the outcome and each of them.

A quick and dirty way to simulate two-way ANOVA data

n <- 100
A <- gl(2, n/2, n, labels=paste("a", 1:2, sep=""))
B <- gl(2, n/4, n, labels=paste("b", 1:2, sep=""))
df <- data.frame(y=rnorm(n), A, B)

A good replacement to sink() for capturing output of R commands

is capture.output(). From the on-line help, it can even be combined to enscript like so:

ps <- pipe("enscript -o tempout.ps","w")
capture.output(example(glm), file=ps)

conflicts(detail=TRUE) gives details about masked functions

Export plot with shaded border (like OS X screencapture utility)

R> png("grv.png"); plot(replicate(2, rnorm(100))); dev.off()
$ convert grv.png \( +clone -background black -shadow \
    55x15+0+5 \) +swap -background none -layers merge  +repage grv2.png

Instead of replicate, we can use rply to do something

when no return values is expected, e.g. for an animation

r_ply(10, plot(runif(50)))

To get the column number of a column given its name

better than which (colnames(d)="a")= or grep("^a$", colnames(df)) use match("a", names(d)) With huge data.frame, it's even better to use the fastmatch package (use fmatch instead of match). h/t Matthew Dowle, http://bit.ly/yFeDTT

To update to a new version of R while keeping older installed programs

we can use pkgutil to look for available versions on a Mac system:

$ pkgutil --packages / | grep org.r-project

Mac versions usually include Leopard.fw.* in the installaed receipts. Then, we just have to tell the system to forget about a previous version, using e.g.,

$ sudo pkgutil --forget org.r-project.R.Leopard.fw.pkga


page(d, method"print")= to view a long data frame d through a pager.


How to prevent void in output of list

https://stackoverflow.com/q/41135630 On the use of the #:when clause in for/list form.


Using texdoc mathptmx show up a complete description of current font

selection scheme for use with common PostScript fonts. Example of a nice title typesetted in Times New Roman (h/t tex.stackexchange.com, 8768) :


Geometric\\ Invariant Theory

http://www.identifont.com is useful to try to identify

a given typeface.

http://detexify.kirelabs.org is a LaTeX symbol classifier

where any symbol can be drawn by hand.

An eps figure can be included in a pdftex or pdflatex document


A tilde can be printed as ∼\!\! in math mode

h/t Jake Hofman

There are various spaces available in math mode

\, = thin space; \; = medium space; \ = space; \quad = quad space; \! = negative space (useful with multiple integrands).

To get a prefixed superscript on something

use {}; i.e., {}_nC_k. h/t Jake Hofman

tex2im is a useful replacement of LaTeXit to convert LaTeX formulas

into high resolution pixmap graphics for inclusion in text processors or presentations. http://www.nought.de/tex2im.html

If you're looking for anything to read before going to sleep

try The LATEX2ε Sources (texdoc source2e).

To see what documents are available through texdoc

just run

$ ls `kpsewhich --var-value=TEXMFDIST`/doc/latex

h/t tex.stackexchange.com, 4358

There's a useful reference card for plain TeX available at

We can use system font with TeX (yes, TeX, not LaTeX) too

say tex_ttf.tex is:

\font\1="Inconsolata"\1 hello

Compile with:

$ xetex -no-pdf tex_ttf
$ xdvipdfmx -vv tex_ttf

It is possible to use a custom background on every page

of a (La)TeX document by using the atbegshi package.

A lot of useful explanations on LaTeX

may be found at the following address: http://nitens.org/taraborelli/latex (Dario Taraborelli).

A collection of fonts I like

Hoefler, Gentium, Fontin Sans, Myriad Pro, Minion Pro, Gill Sans, Inconsolota.

A simple rendering of the TeX logo in HTML is

T<span style"position: relative; top: .2em">E</span>X=. A better solution is to directly embed relevant code into a CSS stylesheet, like proposed by Edward O’Connor's on his blog: http://edward.oconnor.cx/2007/08/tex-poshlet.

There is no way to pass a range of citation as [1-5]

(instead of [1,2,3,4,5]), but here is a nice solution from http://tex.stackexchange.com (Credit: Martin Scharrer):

\foreach \n in {#1} {%

Then, we can use things like

\rcite{3,...,10}      % 3-10
\rcite{2,5,6,...,10}  % 2, 5-10
\rcite{3,5,...,10}    % 3, 5, 7, 9

To place all hyperlinks in dedicated footnote


h/t http://tex.stackexchange.com/questions/12774


To insert a non-breaking space

See this SO thread, or just use C-x 8 SPC.


Generated by Emacs 28.1 (Org mode 9.5.2) on 2022-08-23 mar. 14:14 (chl@aliquote)