#+title: Notes on Emacs #+author: Various #+filetags: NOTE #+options: toc:t This is a small collection of Emacs related posts and documentation by other authors that I find useful. * Emacs Features Random links and notes about various of Emacs' features. ** Eshell - [[https://www.masteringemacs.org/article/complete-guide-mastering-eshell][Mastering Eshell]] by Mickey Petersen - [[http://www.howardism.org/Technical/Emacs/eshell-fun.html][Eschewing Zshell for Emacs Shell]] by Howard Abrams - [[https://github.com/howardabrams/dot-files/blob/master/emacs-eshell.org][Abrams' EShell configuration]] ** Fonts and Themes - To show all available fonts, use the function ~x-family-fonts~. - Load custom themes with ~customize-themes~ instead of calling ~load-theme~ directly. However, this does not set ~custom-enabled-themes~ correctly (Emacs 26.1), so directly customizing ~custom-enabled-themes~ seems to be the best thing (although one has to know the names of the themes then). ** Magit Walkthrough :PROPERTIES: :CREATED: [2018-08-11 Sat 21:05] :END: https://emacsair.me/2017/09/01/magit-walk-through/ * Random Notes ** Insert URL from Safari :IRREAL: :PROPERTIES: :CREATED: [2018-08-04 Sat 17:44] :END: Responding to [[http://irreal.org/blog/?p=6924][yesterday's post]], Sacha asks if I could post the code for =jcs-insert-url= for others to use. I thought I'd already done that but apparently not. That's probably because except for the part identical to =jcs-get-link=, which I /did/ [[http://irreal.org/blog/?p=2895][write about]], it's pretty trivial. In any event, here it is: #+BEGIN_SRC emacs-lisp (defun jcs-insert-url () "Insert URL of current browser page into Emacs buffer." (interactive) (insert (jcs-retrieve-url))) #+END_SRC The =jcs-retrieve-url= function does all the work, of course, and is just the code that I abstracted out of =jcs-get-link= to actually retrieve the URL from Safari: #+BEGIN_SRC emacs-lisp (defun jcs-retrieve-url () "Retrieve the URL of the current Safari page as a string." (org-trim (shell-command-to-string "osascript -e 'tell application \"Safari\" to return URL of document 1'"))) #+END_SRC One obvious problem with all this is that it works only for macOS. Not to despair, though, because in the comments to the original post, [[http://irreal.org/blog/?p=6924#comment-3732979999][Brad Collins suggests a solution]] that uses [[https://github.com/xuchunyang/grab-x-link][grab-x-link]] to do the same thing for FireFox and Chrome on other systems. Be sure to read Brad's comment because there is---or at least was---an issue with the MELPA version. Finally, Sacha took the part about looking for ways to make your workflow easier seriously and came up with a bit of Elisp to [[http://sachachua.com/blog/2018/01/org-mode-inserting-a-function-definition/][insert a function definition at the point]], regardless of where it's defined. That's very handy and I immediately stole her code and used it to insert the two functions above. My old method was to switch to =init.el=, find the function, copy it to the kill ring, switch back to the original buffer, add the source block fences, and insert the code between them. Sacha's code did all of that for me and I didn't even have to leave my current buffer. That's splendid. If you find yourself having to add function definitions to your text, be sure to read Sacha's post. It will save you a lot of time. [[http://irreal.org/blog/?p=6926][Link]] ** Calc for Programmers :IRREAL: :PROPERTIES: :CREATED: [2018-08-05 Sun 10:04] :END: After writing about Florian Adamsky's post on [[http://irreal.org/blog/?p=7040][acronyms in AUCTeX]], I snooped around on his site and came across a [[https://florian.adamsky.it/2016/03/31/emacs-calc-for-programmers-and-cs.html][nice post]] on [[https://www.gnu.org/software/emacs/manual/html_node/calc/index.html][Emacs Calc]] from a programmer's and computer scientist's point of view. As regular readers know, I've been working to increase my calc-fu lately so I read the post with interest. Adamsky demonstrates some of the Calc functions that are useful to programmers and computer scientists. This includes such things as entering and displaying numbers in various radixes and performing the standard logical operations on (the usually binary representation of) numbers. He even shows how to add a new “units” representation to Calc---in this case bits/bytes/bits per second. Calc is a large subsystem and famously hard to master but worth the effort. It's been described as a “poor man's Mathematica.” It's not nearly as powerful as Mathematica, of course, but it's surprising how many things it can do. If you're a programmer/computer scientist and an Emacs user you should spend a little time investigating Calc. It really can make your life easier. An easy way to get started is to read Adamsky's post. It covers only a small slice of Calc but will give you an idea of its power. [[http://irreal.org/blog/?p=7044][Link]] ** Parsing with ~org-element~ :IRREAL: :PROPERTIES: :CREATED: [2018-08-10 Fri 17:55] :END: The other day, I saw [[https://www.reddit.com/r/emacs/comments/89bxe0/org_mode_longterm_time_tracking_for_freelancers/][this query]] on the reddit Emacs subreddit. I already have solutions for this type of problem but I'm always interested in the how people use Org mode to record and report data so I followed the link that primitiveinds provided for [[https://alexpeits.github.io/programming/2017/02/12/org-timesheets.html][his solution]] to generating time reports. Even if, like me, you already have your time tracking and reporting needs under control, primitiveinds' solution is worth looking at for its own sake. It works by looking for CLOCK entries in an Org buffer and accumulating the relevant information in the CLOCK line as well data about the associated task. That might seem like it would require routine but tedious text manipulation but primitiveinds leverages the org-element functionality to easily handle the task. He starts by calling =org-element-parse-buffer= to generate a tree representation of the Org buffer. Then he uses =org-element-map= to examine each CLOCK element (and only CLOCK elements) to extract the necessary information. It's a great technique that can easily be adapted for other parsing of Org data. The code that primitiveinds presents is easy to follow and he provides a nice explanation of what it's doing. If you need to programmatically examine Org data for further processing, you should take a look at primitiveinds' post. It's definitely worth a read. [[http://irreal.org/blog/?p=7073][Link]] The relevant code: #+begin_src emacs-lisp (nconc '(("date" "project" "hours" "task")) '(hline) (let ((ast (org-element-parse-buffer 'element))) (org-element-map ast 'clock (lambda (x) (let* ((val (org-element-property :value x)) (task (org-element-property :parent (org-element-property :parent x)))) `(,(let ((year (org-element-property :year-start val)) (month (calendar-month-name (org-element-property :month-start val))) (day (org-element-property :day-start val))) ;; (insert (org-element-property :raw-value val)) (format "%s %s, %s" month day year)) ,(org-element-property :PROJECT task) ,(org-element-property :duration x) ,(org-element-property :title task) ))))) '(hline) '(("" "total:" ":=vsum(@2..@-1);T" ""))) #+end_src ** Emacs Lisp Byte-Code :IRREAL: :PROPERTIES: :CREATED: [2018-08-11 Sat 21:40] :END: Very few Emacs users---no matter how advanced---ever need to worry about the specifics of the Elisp bytecode, or even, for that matter, that it exists. Still, as guys like Chris Wellons [[http://nullprogram.com/blog/2014/01/04/][have shown]], it can sometimes be useful to have a basic understanding of the bytecodes. R Bernstein has put together a comprehensive, book-length [[http://rocky.github.io/elisp-bytecode.pdf][documentation on Elisp bytecodes]]. After a short introduction, the documentation considers the bytecode environment including the compiler, interpreter, and bytecode optimization. Then there's a long section on the individual bytecode instructions. Finally, there are sections on the changes in bytecodes between Emacs versions, a table of opcodes, and a reference section. There's also a GitHub repository of the [[https://github.com/rocky/elisp-bytecode][document source]]. As I said, you probably will never need this but if you do, you'll be /very/ glad to have Bernstein's documentation. It's another example of the vibrant Emacs community. [[http://irreal.org/blog/?p=7166][Link]] ** Formatting Tables :IRREAL: :PROPERTIES: :CREATED: [2018-10-28 Sun 09:13] :END: If you're like me, you automatically think of the Org mode table editor (or Orgtbl minor mode) when you think of tables in Emacs. It's hard to beat that functionality and Orgtbl mode makes it available everywhere in Emacs, even if you're not in an Org buffer. Sometimes, though, you'd like to have special formatting for some or all of the table. That's where =delim-col= comes in. =Delim-col= is /built-in/ Emacs functionality that allows you to do things like adjust what string separates the columns, add a beginning or ending string to each item, add an ending string for each row, and adjust the padding in the table. It can be really handy for copying and pasting and then reformatting tables from an external source. I didn't know about =delim-col= until I read about it [[https://emacsnotes.wordpress.com/2018/09/24/delim-col-a-handy-tool-for-creating-pretty-tables-and-converting-those-to-different-table-formats/][over at Emacs Notes]], where you'll find a good explanation of the facility and what it can do. The Emacs Notes post also offers at bit of Elisp to make choosing the strings and delimiters a bit easier. By default you have to set them using a series of =setq= statements if you want something different from the built-in choices. The Emacs Notes codes arranges for you to be prompted for the values. You probably won't need the =delim-col= functionality very often but when you do it's much easier than using something like a keyboard macro. Take a look at the post and see if you don't agree. [[http://irreal.org/blog/?p=7540][Link]] ** Org Mode Cookbook :IRREAL: :PROPERTIES: :CREATED: [2018-08-04 Sat 12:51] :END: Way back in 2014, I [[http://irreal.org/blog/?p=2575][posted]] about Eric Neilsen's excellent [[http://ehneilsen.net/notebook/orgExamples/org-examples.html][Emacs org-mode examples and cookbook]]. I recently came across a reference to it and was reminded what a great resource it is. It's easy to browse through and just read one or two entries when you have time. In skimming through it, I learned---or perhaps relearned---how to [[http://ehneilsen.net/notebook/orgExamples/org-examples.html#sec-10][insert in-line calculations in a document]]. As I wrote in the original post, Neilsen is a researcher and his cookbook is oriented at using Org mode to produce documents of various types. Still, that covers a lot of territory and there are many good examples of powerful Org mode use cases in it. The Document has moved or, really, taken up a second residence. It was originally hosted at [[http://fnal.gov/][Fermilab]], where Neilsen works, and it's still there but it's also available at his own site. The two documents are identical so it doesn't matter if you use the new link or the original one pointing to FNAL. If you're an Org user, especially if you use Org to produce documents, you should take a look at Neilsen's cookbook and bookmark it for future use. [[http://irreal.org/blog/?p=6894][Link]] ** How to paste then copy :PROPERTIES: :CREATED: [2018-08-11 Sat 21:47] :END: Question: how to set a mark such that all subsequent copy operations move their text to that exact mark. Answer: use ~cua-selection-mode~! See https://www.reddit.com/r/emacs/comments/8ekz0u/how_to_pastethencopy/. /Update/: turns out it doesn’t work so well, disabled it again. ** Tramp and Telnet over non-standard ports :PROPERTIES: :CREATED: [2018-12-29 Sat 15:58] :END: Syntax: ~/telnet:HOST#PORT:~, works also with other protocols. ** Fractals in Emacs :PROPERTIES: :CREATED: [2018-08-04 Sat 13:01] :END: From https://nullprogram.com/blog/2012/09/14/ #+begin_src emacs-lisp (defun sierpinski (s) (pop-to-buffer (get-buffer-create "*sierpinski*")) (fundamental-mode) (erase-buffer) (labels ((fill-p (x y) (cond ((or (zerop x) (zerop y)) "0") ((and (= 1 (mod x 3)) (= 1 (mod y 3))) "1") (t (fill-p (/ x 3) (/ y 3)))))) (insert (format "P1\n%d %d\n" s s)) (dotimes (y s) (dotimes (x s) (insert (fill-p x y) " ")))) (image-mode)) (defun mandelbrot () (pop-to-buffer (get-buffer-create "*mandelbrot*")) (let ((w 400) (h 300) (d 32)) (fundamental-mode) (erase-buffer) (set-buffer-multibyte nil) (insert (format "P6\n%d %d\n255\n" w h)) (dotimes (y h) (dotimes (x w) (let* ((cx (* 1.5 (/ (- x (/ w 1.45)) w 0.45))) (cy (* 1.5 (/ (- y (/ h 2.0)) h 0.5))) (zr 0) (zi 0) (v (dotimes (i d d) (if (> (+ (* zr zr) (* zi zi)) 4) (return i) (psetq zr (+ (* zr zr) (- (* zi zi)) cx) zi (+ (* (* zr zi) 2) cy)))))) (insert-char (floor (* 256 (/ v 1.0 d))) 3)))) (image-mode))) #+end_src ** Presentations with ~org-reveal~ Website: https://github.com/yjwen/org-reveal Sample: #+BEGIN_SRC org #+title: Foo! #+author: bar #+options: author:t toc:t num:nil date:nil timestamp:nil #+reveal_theme: sky * Slide 1 * Slide 2 ** Subslide 1 ** Subslide 2 - Item 1 - Item 2 | a | b | d | |---+---+---| | 1 | 2 | 3 | * Slide 3 #+END_SRC ** Writing a PhD thesis with Org Mode From: https://write.as/dani/writing-a-phd-thesis-with-org-mode *TLDR*: I started using Emacs about 3 years ago. I couldn't be more grateful to have seen the light, and to have been rescued from the darkness of Windoze, Goggle and/or friends. After enlightenment, I've taken upon myself the task of customising an environment to write my PhD thesis with Org Mode.* *** Why Post created in response to the [[https://www.reddit.com/r/emacs/comments/9ynsvc/write_a_thesis_using_emacs_and_orgmode/][current thread]] in /r/emacs/ on thesis writing with Org Mode.\\ I see most people's reason to avoid Org mode for scientific writing is the fact that supervisors or co-authors use Mic. Word. I'll try to argue that that's not enough reason to accept subpar tools. *** What I'll talk about I'll mention a bit of my motivations, and then I'll discuss how to make use of (mostly) built in Org functionality such as tagging, export, [[https://orgmode.org/manual/In_002dbuffer-settings.html][setupfiles]] and includes, reference management, keyboard shortcuts and advanced searching; all with the purpose of building a useful thesis writing environment. Readers should have a minimum knowledge of Org mode, the Org export system and LaTeX. *** My requirements Here in the Netherlands, most PhD thesis consist of an introduction, 3 to 4 research chapters (as submitted for publication), a summary, bibliography and appendices. What this means for me is that my writing environment has to /necessarily/ satisfy the following *minimum requirements*: - Inclusion of (parts of) external files - Keeping track of references - Include and reference figures - Version control documents - Support for sharing with my supervisor in whatever format he wants Failure to comply with any of these means the editor is unfit for purpose^{#fn.1”>1}. Unfortunately, this set of requirements are not seamlessly satisfied by likes of Mic. Word or G. Docs. I reckon they can probably be configured to satisfy them, but why bother. Additionally, a PhD thesis writing environment should also provide the following features: - Extended searching facilities for both text and references - Simple syntax for tables and equations - Support within a proper text editor - Shortcuts to reach my files and build the thesis To the best of my knowledge, /only/ Emacs with Org Mode + ox-latex provide all of these out of the box. *** Moulding Org Mode for thesis writing Most of my inspiration comes from reading Kitchin's blogs and code, and reading the Org Mode documentation, mailing list and Emacs Stack Exchange. Here' I'll go one by one through all of the requirements listed above, and how to deal with them. **** Prelude: File structure I have a main /thesis.org/ document, with latex heading declarations and a commented setup file. I also have /research.org/ files, in different directories, with their own latex heading declarations and commented setup files. The first lines of /thesis.org/ look like the following: #+BEGIN_src org # -*- mode: org; org-latex-title-command: ""; org-latex-toc-command: "" -*- #+TITLE: Thesis Title #+LATEX_CLASS: mimosis # Setupfile with #+LATEX_HEADER, #+OPTIONS and explanations #+SETUPFILE: thesis.setup #+LATEX_HEADER: \KOMAoptions{fontsize=12pt,headings=small} #+LATEX_HEADER: \bibliography{~/Papers/bibtex/Publications} #+EXCLUDE_TAGS: journal noexport * Frontmatter :ignore: #+LATEX: \frontmatter #+INCLUDE: ./Title.org #+LATEX: \tableofcontents * Mainmatter :ignore: #+LATEX: \mainmatter * Introduction * Research 1 #+INCLUDE: "../research1/research.org::*Abstract" :only-contents t Some stuff. #+INCLUDE: "../research1/research.org" :lines "5-" * Research 2 ... #+END_src And the first lines and structure overview of the multiple /research.org/ files: #+BEGIN_src org #+TITLE: Research #+LATEX_CLASS: elsarticle #+LATEX_CLASS_OPTIONS: [authoryear,preprint,11pt] #+SETUPFILE: paper.setup #+LATEX_HEADER:\bibliography{./ref/Publications-research} #+EXCLUDE_TAGS: thesis noexport * Frontmatter :ignore:journal: #+LATEX: \begin{frontmatter} ** Author List :ignore: Abstract :ignore: Keywords :ignore: #+LATEX: \end{frontmatter} * Introduction ... #+END_src **** Inserting (parts of) external files I write my research chapters with LaTeX classes targeting the journal's format. That means that a research chapter may be written with the =elsarticle= class, whereas the thesis as a whole is written with the =mimosis= [[https://github.com/Submanifold/latex-mimosis][class]], a derivative of KOMA =scrbook=. Here's the class configuration for both: #+begin_src emacs-lisp (add-to-list 'org-latex-classes '("elsarticle" "\\documentclass{elsarticle} [NO-DEFAULT-PACKAGES] [PACKAGES] [EXTRA]" ("\\section{%s}" . "\\section*{%s}") ("\\subsection{%s}" . "\\subsection*{%s}") ("\\subsubsection{%s}" . "\\subsubsection*{%s}") ("\\paragraph{%s}" . "\\paragraph*{%s}") ("\\subparagraph{%s}" . "\\subparagraph*{%s}"))) (add-to-list 'org-latex-classes '("mimosis" "\\documentclass{mimosis} [NO-DEFAULT-PACKAGES] [PACKAGES] [EXTRA] \\newcommand{\\mboxparagraph}[1]{\\paragraph{#1}\\mbox{}\\\\} \\newcommand{\\mboxsubparagraph}[1]{\\subparagraph{#1}\\mbox{}\\\\}" ("\\chapter{%s}" . "\\chapter*{%s}") ("\\section{%s}" . "\\section*{%s}") ("\\subsection{%s}" . "\\subsection*{%s}") ("\\subsubsection{%s}" . "\\subsubsection*{%s}") ("\\mboxparagraph{%s}" . "\\mboxparagraph*{%s}") ("\\mboxsubparagraph{%s}" . "\\mboxsubparagraph*{%s}"))) #+END_src Research chapters print the bibliography on their own, and they may contain acknowledgements that shouldn't be present in the middle of the thesis, so they should be excluded. In other to insert research chapters into my thesis, I use Org's =#+INCLUDE= derivative: #+begin_src org #+INCLUDE: file.org #+end_src In order to not include the some parts of the file, i.e., to exclude the title, setupfile and headers, I narrow down the lines: #+begin_src org # Include line 5 until the end of the file #+INCLUDE: file.org :lines 5- #+end_src In order to exclude parts of the file, I tag research chapter headings that are only meant for publication with a =:journal:= tag (such as the bibliography or acknowledgements). This way they are automatically excluded from the thesis (see the =#+EXCLUDE_TAGS:= derivative in the /thesis.org/ file). Also, I could have thesis specific content in the /research.org/ document tagged with =:thesis:=, and it would be excluded in the /research.org/ export, but I currently don't have any. Now, the most important piece of advice I can give anyone is to *learn how to use tags*, =EXCLUDE_TAGS= and the org-plus-contributions =ignore= tag. With the =ignore= tag we separate the structuring of the text as a physical document from the structuring of the text as a semantic unity. This allows an extremely fine control over pieces of text to include into another document. For example, in a research chapter written with the =elsarticle= class, the abstract has to be included in the Frontmatter. By tagging a headline as follows: #+begin_src org ** Abstract :ignore: #+end_src I can write the research abstract in it's own heading, pretend that the heading itself does not exist (so it does not trigger =/begin{document}=), only its contents, and then include the contents in the thesis in an arbitrary location: #+begin_src org # in thesis.org #+INCLUDE: "research.org::*Abstract" :only-contents t #+end_src The =:ignore:= tag is one of the best Org mode features, in my opinion. It's key to my workflow, and a shame to see it's not a part of Org core, but rather a contribution to be found in /ox-extra.el/. To activate it, add the following to your /init/: #+BEGIN_src emacs-lisp (require 'ox-extra) (ox-extras-activate '(ignore-headlines)) #+END_src The realisation that it's possible to have such fine control over where to include or exclude pieces of text opens the door to all sort of interesting experiments: putting figures and captions directly into beamer or org-reveal presentations, creating conference posters, writing blog posts, etc. **** Keep track of references For backwards compatibility I still use Mendeley to track literature. I export bibtex files for each research project individually, and also a master bibtex for use in the thesis. These documents are saved to =~/Papers/bibtex/=, but for the research chapters, I keep local copies under =./ref/Publications-research.bib=.\\ To insert citations, I use [[https://github.com/jkitchin/org-ref][org-ref.]] It's documentation says it all. After setting up local bibliography files with the derivative =#+BIBLIOGRAPHY=, press =C-c ]= to see a list of publications and insert them in place. I also prefer to have =parencite= citations instead of =cite=, because they work nicely with BibLaTeX. My setup for org-ref: #+begin_src emacs-lisp (with-eval-after-load 'org-ref ;; see org-ref for use of these variables (setq org-ref-default-bibliography '("~/Papers/bibtex/Publications.bib") org-ref-pdf-directory "~/Papers/MendeleyDesktop/" org-ref-get-pdf-filename-function 'org-ref-get-mendeley-filename bibtex-completion-pdf-field "file" org-latex-prefer-user-labels t org-ref-default-citation-link "parencite" ;; bibtex-dialect 'biblatex ) (defun org-ref-open-pdf-at-point-in-emacs () "Open the pdf for bibtex key under point if it exists." (interactive) (let* ((results (org-ref-get-bibtex-key-and-file)) (key (car results)) (pdf-file (funcall org-ref-get-pdf-filename-function key))) (if (file-exists-p pdf-file) (find-file-other-window pdf-file) (message "no pdf found for %s" key)))) ;; https://github.com/jkitchin/org-ref/issues/597 (defun org-ref-grep-pdf (&optional _candidate) "Search pdf files of marked CANDIDATEs." (interactive) (let ((keys (helm-marked-candidates)) (get-pdf-function org-ref-get-pdf-filename-function)) (helm-do-pdfgrep-1 (-remove (lambda (pdf) (string= pdf "")) (mapcar (lambda (key) (funcall get-pdf-function key)) keys))))) (helm-add-action-to-source "Grep PDF" 'org-ref-grep-pdf helm-source-bibtex 1) (setq helm-bibtex-map (let ((map (make-sparse-keymap))) (set-keymap-parent map helm-map) (define-key map (kbd "C-s") (lambda () (interactive) (helm-run-after-exit 'org-ref-grep-pdf))) map)) (push `(keymap . ,helm-bibtex-map) helm-source-bibtex) (setq org-ref-helm-user-candidates '(("Open in Emacs" . org-ref-open-pdf-at-point-in-emacs)))) #+end_src **** Include and reference figures For each research project I keep a =./media= directory, where all my figures live. You can include figures in Org mode by using the following syntax: #+begin_src org #+NAME: figurename #+CAPTION: This is a figure caption [[path_to_figure][link_description]] #+end_src Currently there is a bug in the ELPA version of Org mode, such that relative paths to figures in =#+INCLUDE= 'd files aren't adapted with respect to the including file, so the latex export cannot find them. I've [[https://code.orgmode.org/bzg/org-mode/commit/d81a1d088c74e605c99e90a2835c55df5144f43e][submitted a fix]] which should land in the next release of Org. **** Version control documents [[https://magit.vc/][Magit]]. I thought about having the research chapters as git submodules in a thesis git project directory, but I currently don't. This would allow me to always have the thesis code in a saved state, even if I further work on my research chapters to answer to reviewers questions. **** Support for sharing with my supervisor Unfortunately, my supervisor likes to write comments in Mic. Word. I give in that sharing your writing with colleagues is a fundamental part of research.\\ Fortunately, [[https://github.com/jkitchin/scimax/blob/master/ox-word.el][ox-word]] export via Pandoc & LaTeX is capable of creating nice looking, structured Word files which I send to my supervisor. I then manually work through each comment step by step, though I'm looking for a way to improve this part of my workflow. I think the Emacs community is missing a minor mode to track Word document changes from within Org Mode. There are some ideas laying around on how to implement it [[https://lists.gnu.org/archive/html/emacs-orgmode/2015-06/msg00246.html][hidden deep in the mailing list]], or in [[https://emacs.stackexchange.com/questions/34923/merging-changes-to-from-docx-files-into-org-files][this Emacs Exchange thread]]. I may update this post with more information later. **** Extended search facilities By extended search facilities I mean the ability to quickly search for information in references, and to keep notes linked to the literature. For searching I make use of [[https://github.com/jkitchin/org-ref/issues/597][org-ref + pdfgrep]], as shown in my org-ref setup. For notes linked to documents I've recently started to use [[https://github.com/weirdNox/org-noter][Org-noter.]] **** Simple syntax for tables and equations Org tables are a pleasure to work with. The following: #+BEGIN_EXAMPLE | a | b | c | |---+---+---| | 1 | 2 | 3 | #+END_EXAMPLE Turns into: | a | b | c | |-----+-----+-----| | 1 | 2 | 3 | Equations can be written in LaTeX: #+BEGIN_EXAMPLE \frac{d \vec{M} (t)}{dt} = \vec{M} (t) \times \gamma \vec{B} (t) #+END_EXAMPLE will become /omitted/ **** Support within a proper text editor No need to talk about the synergy of using Emacs to edit text. I personally started using Spacemacs without Evil mode, because I find it aesthetically pleasing and because it offers great support for the languages I use the most, and excellent integration with Helm and Org.\\ The following configurations make the Org editing experience a bit nicer, in my opinion: #+BEGIN_src emacs-lisp ;; Writegood https://github.com/bnbeckwith/writegood-mode (add-hook 'org-mode-hook 'writegood-mode) ;; https://github.com/cadadr/elisp/blob/master/org-variable-pitch.el (use-package org-variable-pitch :load-path "~/Elisp") (add-hook 'org-mode-hook 'org-variable-pitch-minor-mode) (setq visual-fill-column-width 120 visual-fill-column-center-text t) (add-hook 'org-mode-hook 'visual-line-mode) ;; https://github.com/joostkremers/visual-fill-column (add-hook 'org-mode-hook 'visual-fill-column-mode) (add-hook 'org-mode-hook 'org-display-inline-images) ;; I have a modified version of the following: ;; https://github.com/lepisma/rogue/blob/master/config.el (load-file "~/Projects/rogue/config.el") (add-hook 'org-mode-hook '(lambda () (setq-local line-spacing 5))) ;; Aesthetical enhancements. (setq org-fontify-quote-and-verse-blocks t org-hide-macro-markers t org-fontify-whole-heading-line t org-fontify-done-headline t org-hide-emphasis-markers t) #+END_src **** Shortcuts to reach my files and build the thesis I have a hydra (defined in Spacemacs as a transient-state) to move between my Thesis files: #+begin_src emacs-lisp ;; Spacemacs hydra. (spacemacs|define-transient-state thesis-menu :title "Ph.D. Thesis Menu" :doc " ^Main Files^ ^Chapters^ ^Actions^ ^^^^^^^^------------------------------------------- _m_: Thesis _1_: Research 1 _o_: Open Thesis.pdf externally _t_: Title page _2_: Research 2 _c_: Async compile file _i_: Introduction _3_: Research 3 _a_: things _s_: thesis.setup _4_: Research 4 ^ ^ " :bindings ("a" things :exit t) ("m" (find-file "~/thesis/thesis.org") :exit t) ("t" (find-file "~/thesis/titlepage.org") :exit t) ("s" (find-file "~/thesis/thesis.setup") :exit t) ("i" (find-file "~/thesis/intro/intro.org") :exit t) ("1" (find-file "~/thesis/ch1/research.org") :exit t) ("2" (find-file "~/thesis/ch2/research.org") :exit t) ("3" (find-file "~/thesis/ch3/research.org") :exit t) ("4" (find-file "~/thesis/ch4/research.org") :exit t) ("o" (shell-command "open ~/thesis/thesis.pdf" :exit t)) ("c" (org-latex-export-to-pdf :async t) :exit t)) (global-set-key (kbd "H-t") 'spacemacs/thesis-menu-transient-state/body) #+end_src ** Gnus and notmuch Currently working: ~notmuch~ for ~nnmaildir~ backend. However, I also have a local ~nnimap~-backend, which uses Maildir format locally. It would be nice to have this working with ~notmuch~ as well. Relevant functions: - ~nnir-compose-result~ :: Goes over the results of notmuch (stored in the buffer ~*nnir*~ (with an additional leading space) and decides which lines to keep; - ~nnir-notmuch-remove-prefix~ :: a server-local variable to decide what to remove from the lines in ~*nnir*~. ** Summary of Search and Replace Commands in Emacs [2016-08-13 Sat 17:50] http://www.omps.in/blog/2016/01/20/day-10-search-and-replace-using-regex/ ** Bookmarks with Org-mode :PROPERTIES: :CREATED: [2018-07-08 Sun 09:29] :END: [[http://karl-voit.at/2014/08/10/bookmarks-with-orgmode/]] ** Inserting a function definition :CHUA: :PROPERTIES: :CREATED: [2018-08-04 Sat 17:40] :END: From [[http://sachachua.com/blog/2018/01/org-mode-inserting-a-function-definition/][Sacha Chua]]. While nudging jcs to add a definition of =jcs-insert-url= to the blog post about [[http://irreal.org/blog/?p=6924][Making Things Easier]], I realized it might be handy to have a quick function for inserting a function definition without thinking about where it's defined. This tries to use the definition from the source, and it can fall back to using the stored function definition if necessary. There's probably a better way to do this, but this was small and fun to write. =) Naturally, I used it to insert itself: #+begin_src emacs-lisp (defun my/org-insert-defun (function) "Inserts an Org source block with the definition for FUNCTION." (interactive (find-function-read)) (let* ((buffer-point (condition-case nil (find-definition-noselect function nil) (error nil))) (new-buf (car buffer-point)) (new-point (cdr buffer-point)) definition) (if buffer-point (with-current-buffer new-buf ;; Try to get original definition (save-excursion (goto-char new-point) (setq definition (buffer-substring-no-properties (point) (save-excursion (end-of-defun) (point)))))) ;; Fallback: Print function definition (setq definition (concat (prin1-to-string (symbol-function function)) "\n"))) (insert "#+begin_src emacs-lisp\n" definition "#+end_src\n"))) #+end_src ** Tramping into GCloud instances from within emacs :IRREAL: From https://gist.github.com/jackrusher/36c80a2fd6a8fe8ddf46bc7e408ae1f9 via [[https://irreal.org/blog/?p=8126][Irreal]]. #+begin_src emacs-lisp ;; make sure you've set your default project with: ;; gcloud config set project (require 'tramp) (add-to-list 'tramp-methods '("gcssh" (tramp-login-program "gcloud compute ssh") (tramp-login-args (("%h"))) (tramp-async-args (("-q"))) (tramp-remote-shell "/bin/sh") (tramp-remote-shell-args ("-c")) (tramp-gw-args (("-o" "GlobalKnownHostsFile=/dev/null") ("-o" "UserKnownHostsFile=/dev/null") ("-o" "StrictHostKeyChecking=no"))) (tramp-default-port 22))) ;; ... after which it's as easy as: ;; ;; C-x C-f /gcssh:compute-instance:/path/to/filename.clj #+end_src ** Finding the largest Emacs buffer I found this somewhere, but forget where … Approach: sort all buffers by size. #+begin_src emacs-lisp (mapcar (lambda (x) (list x (buffer-size x))) (cl-sort (buffer-list) (lambda (x y) (< (buffer-size y) (buffer-size x))))) #+end_src ** Emacs reencode buffers with other encoding - Use the function =revert-buffer-with-coding-system= to re-open a buffer in another encoding. - Customize the variable =file-coding-system-alist= to control whether certain files should always be opened with a specific encoding. * Tutorials ** Online - [[https://github.com/alphapapa/emacs-package-dev-handbook][Emacs Package Developer's Cookbook]] ** Workshop for Emacs Macros (By [[https://www.howardism.org/][Howard Abrams]], 2015 Nov 24) Basic idea of this file is to be both a *guide* and a *practice pad* for leveling up on Emacs’ keyboard macro system. To use, just [[https://github.com/howardabrams/pdx-emacs-hackers/raw/master/workshops/keyboard-macros.org][download this file]] to your system, and edit it in Emacs. Then just follow along... If you are not familiar with org-mode files, here are the basics for this workshop: - Hit a tab on the header expands or collapses it - Narrow to show just one section with: C-x n s - Return by widening normally with: C-x n w - Hyperlinks can be clicked on to view the Emacs Manual Pretty much treat the rest of this document as a text file. *** Basic Usage Let’s begin with something simple to try out this project. In the /block area/ below, hit: C-c ' (yes, the apostrophe). When done with the tasks, hit: C-x C-s to close and return here. #+BEGIN_SRC org - This foobar should be Interesting - Nulla foobar Posuere - Nullam foobar Tempus - Etiam foobar laoreet quam sed Arcu - Donec foobar hendrerit tempor Tellus - Mauris foobar mollis tincidunt Felis - etiam laoreet quam sed arcu - nullam rutrum - sed diam - lorem ipsum dolor sit amet, consectetuer adipiscing elit - mauris ac felis vel velit tristique imperdiet - fusce suscipit, wisi nec facilisis facilisis, est dui fermentum leo, quis tempor ligula erat quis odio - donec hendrerit tempor tellus #+END_SRC Notice that the frame is split with this document still visible? Good, let’s play around. Let’s capitalize the first letter of each line. Position point at the beginning of the first line, and then: - F3 - M-c - C-n - C-a - F4 Press F4 repeatedly to capitalize the first word on each line. Don’t capitalize /every/ line... leave a few alone for a moment. Now you have the basics. Let’s make a new macro to capitalize the /last/ word on each line. Pop back up to the top of the list, and: - F3 - C-e - M-b - M-c - C-n - F4 Once again, leave a few lines alone. Just one more macro, as I want to show you something interesting. Jump back to the beginning, and: - F3 - C-a - M-f - Space - foobar (or type any single word you find amusing) - C-n - F4 Type: C-x C-k C-k Notice it behaves the same as F4, but you can keep typing C-k to execute your macro. Start typing C-n and/or C-p and notice the minibuffer. It alternates between the following: - M-c C-n C-a - C-e M-b M-c C-n - C-a M-f SPC f 2*o b a r C-n Typing C-k at any point will execute that macro. You’ve now see the Emacs Macro Ring, and can manipulate it. You are now at Level 2. Type C-c ' (apostrophe) to close that side window. *** Naming Macros While that =foobar= macro may be quite handy, looping around the keyboard macro ring to select it may be inefficient. Name it! C-x C-k n Give it a name of foobar, and now, you can: M-x foobar to have the macro run. Perhaps it may be more useful to bind that macro to some un-used key, like C-c k or F2 ... type: C-x C-k b And then type your binding, e.g. C-c k You know that C-c followed by a single letter is supposed to be free of usage and available for your nefarious reasons, right? Read the [[info:emacs#Save%20Keyboard%20Macro][the Emacs Manual]] for details. If you actually tried this example, then you have earned enough experience points to be level 3. *** What do you mean On Every Line? Keyboard macros are slick, and while we /could/ prefix the F4 with the numbering prefix to run it multiple times at once, e.g. C-9 F4, a typical use case is to run it on every line in a region. Position the cursor in the block, and type: C-c ‘ #+BEGIN_SRC org I will not take out the trash I should not stop swearing I may not stop to smell the roses I do not want to go shopping I could not care less #+END_SRC With the point on the first line, type: - F3 - C-a - M-f - M-f - Space - not (type the word, =not=) - F4 Now select the remaining 4 lines any way you like, and type: - C-x C-k r Read [[info:emacs#Basic%20Keyboard%20Macro][the Emacs Manual]] for complete details, and welcome to level 4. *** Numbering Numbering things seems to be important to people. Let’s see if we can easily number the worst song ever inflicted on humanity. Once again, position the point inside the block, and type: C-c ' #+BEGIN_SRC org 1 Drummers Drumming 2 Pipers Piping 3 Lords a Leaping 4 Ladies Dancing 5 Maids a Milking 6 Swans a Swimming 7 Geese a Laying 8 Golden Rings 9 Calling Birds 10 French Hens 11 Turtle Doves 12 Partridge in a Pear Tree #+END_SRC Follow along at home. First, position the point at the end (on the line with the partridge), and: - C-x C-k C-c - 1 - Return (to set a counter to 1) - F3 - C-a - F3 - Spacebar - C-n - F4 Yes, hitting the F3 key twice (once to start the macro, and again to insert the value of the counter) is a bit odd. You can also use the binding: C-x C-k C-i Each time you hit F4, you increment the counter that is inserted. By the way, if you didn’t do the whole =C-x C-k C-c= to set the counter to 1, the first value would be 0. Read [[info:emacs#Keyboard%20Macro%20Counter][the manual]] for details. *** Fixing Macros Now that you are a Level 5 Elf Keyboard Macro-er, and you can now sling macros everywhere, you may notice that sometimes, in the middle of a long macro, you flub it. If you hit F4 too soon (it happens), hit: C-u C-u F3 to pick up where you left off and enter the rest of the macro. Hit F4 when you are really finished. If you need to fix a keyboard macro with more finesse, type: C-x C-k C-e And behold the glory. Tis a simple language that should be pretty obvious to a skilled Macro-er as yourself. Type C-h m once in that buffer to display details of how to edit the macro. When you are finished editing, type: C-c C-c I’ll let you play around with your own example for this one. Read [[info:emacs#Edit%20Keyboard%20Macro][the manual]] for complete details. *** Variations on a Theme To gain the Level 7 Keyboard Macro-er title, one should know how to customize a macro /while/ running it. The following section of HTML code needs some textual changes. We want to add one of the following phrases to the /end/ of every paragraph that has a =class= of =change=: * Because I said so. Got it? * Because I'm the boss. Got it? * You heard me. Got it? * Just do it. Got it? To begin, first move to the following block (hint: C-c M-f) and hit TAB to collapse the block (you gotta see all the instructions, right?) Next, issue a C-c ' on this block to show it in a new window. If you have trouble with your HTML mode, change the =html= to =text=. #+BEGIN_SRC html

Proin neque massa, cursus ut, gravida ut, lobortis eget, lacus. Praesent augue. Sed diam. Nunc eleifend leo vitae magna. Nunc rutrum turpis sed pede.

Nullam rutrum. Nunc rutrum turpis sed pede.

Phasellus at dui in ligula mollis ultricies. Curabitur lacinia pulvinar nibh. Donec pretium posuere tellus. Praesent fermentum tempor tellus. Proin quam nisl, tincidunt et, mattis eget, convallis nec, purus.

Fusce sagittis, libero non molestie mollis, magna orci ultrices dolor, at vulputate neque nulla lacinia eros. Sed diam. Nam vestibulum accumsan nisl.

Aliquam feugiat tellus ut neque. Nam vestibulum accumsan nisl. Praesent fermentum tempor tellus.

Vivamus id enim. Suspendisse potenti. Curabitur lacinia pulvinar nibh. Mauris ac felis vel velit tristique imperdiet.

Donec vitae dolor. Mauris ac felis vel velit tristique imperdiet. Nunc aliquet, augue nec adipiscing interdum, lacus tellus malesuada massa, quis varius mi purus non odio. Proin quam nisl, tincidunt et, mattis eget, convallis nec, purus. Nam euismod tellus id erat.

Nullam rutrum.

#+END_SRC Type the following: - F3 - C-s - Type: class="change" - C-e - C-s - Type:

- C-p - C-e - C-x q - Type: Got it? - C-n - F4 In this particular case, we actually didn't do anything special, so move to the beginning of the buffer, and hit F4, and the cursor will go to the end of every paragraph that needs changing, and stop with a prompt: Proceed with macro? (Y, N, RET, C-l, C-r) Type C-r and begin typing one of our phrases, and when you are done, type: C-M-c The prompt will be re-displayed, so finish the macro with 'Y'. Check out [[info:emacs#Keyboard%20Macro%20Query][the manual]] for details on this =C-x q= business. ** Emacs Calc Tutorials By Andrew Hyatt, found here: https://github.com/ahyatt/emacs-calc-tutorials. License is GPLv3. Order as given by https://blog.markhepburn.com/2013/12/07/andrew-hyatts-emacs-calc-tutorials *** README #+BEGIN_EXAMPLE This repository contains tutorials about emacs calc originally writen on the Emacs community on Google+. The best way to read is probably just to open the org files directly, which Github will display correctly. If anyone would like to correct anything, add any tutorials, or request anything, the normal Github bug / request / or pull request process will work. #+END_EXAMPLE *** HEX OK, seems like there's interest in some quick calc tips. Here's today's: How to convert decimal to hexidecimal. Let's say you want to convert number 12345 to hex. #+BEGIN_EXAMPLE M-x calc d 6 (sets the number radix to 16, meaning all output will be in hex) 10#12345 (inputs the number 12345 in base 10) The output reads: 1: 16#3039 #+END_EXAMPLE The answer is therefore =0x3039=. And then you can do a =d 0= to set the number radix back to normal, base 10. Here's how to do the other way. Let's convert =0xABCDEF= to base 10. #+BEGIN_EXAMPLE M-x calc 16#ABCDEF The output reads: 1: 11259375 #+END_EXAMPLE *** Date Ever want to know how many seconds old David Hasselhoff is? calc can do many things, but it doesn't know much about Hasselhoff, so first I do a query on Google for [david hasselhoff]. I get a knowledge card on the right saying he was born July 17, 1952. It doesn't give a time, so we'll just assume it was at midnight. #+BEGIN_EXAMPLE M-x calc t N (put the current time on the stack) ' (press ' to enter algebraic mode, then you input the date). - (subtract the two to get the number of days David has been alive) 24 (we're going to multiply by 24, the number of hours in a day) 60 (the number of minutes in an hour) 60 (the number of seconds in a minute) * * * Final result: 1: 1910255938.01 #+END_EXAMPLE There you have it, he's... wait, how many seconds? That's really hard to read. Back into calc! #+BEGIN_EXAMPLE d g (toggle digit grouping) The final final result: 1: 1,910,255,938.01 #+END_EXAMPLE Ah, that's a 1.9 billion seconds. Sweet! *** Time Hey, what's the time? It's time to get ill! No, actually I meant the time in seconds since the epoch. Yesterday I went over doing math with time, which is fun but not something I use everyday. Much more useful is converting to and from Unix timestamps. Let's start by getting the time now in seconds since the epoch: #+BEGIN_EXAMPLE M-x calc t N (get the time now) t U (convert the time to seconds since the epoch) Result: 1: 1359424746 #+END_EXAMPLE Oh, and you want to insert that into your last used buffer? #+BEGIN_EXAMPLE y (that doesn't mean "yes", that means yank into the last buffer) #+END_EXAMPLE Done! Just to be complete, let's convert another date we have to input: #+BEGIN_EXAMPLE '<12:00pm Jul 4, 1776> (single quote to enter algebraic mode, then the date) t U (converts the time to seconds since the epoch) #+END_EXAMPLE But wait, what will happen? This is considerably before the epoch. #+BEGIN_EXAMPLE Result: 1: -6106003200 #+END_EXAMPLE Oh calc, you never let me down. Let's do the other way. Remember the Billenium? #+BEGIN_EXAMPLE 1e9 t U (converts the time in seconds since the epoch to text) Result: 1: <9:46:40pm Sat Sep 8, 2001> #+END_EXAMPLE Wow, I never realized how close the Billenium was to September 11th. Kind of spooky... *** Random I use calc whenever I need a random number. The interface is easy and the random numbers are (supposedly) high quality. So, let's start with something simple: A random number between 0 and 100: #+BEGIN_EXAMPLE M-x calc 100 (the upper bound, all values will be between 0 and this) k r (creates a random number between 0 and the number on the stack) Result: 1: 66 (of course, yours will be different) #+END_EXAMPLE I want another one! #+BEGIN_EXAMPLE k a (creates another number with the same upper bound as the last) #+END_EXAMPLE Now that I’ve had a taste of that sweet sweet randomness, I want a vector of 50! #+BEGIN_EXAMPLE 100 (the upper bound, again) 50 (the number to generate) k h (generate a vector of 50 random numbers between 0 and 100) 1: [60, 72, 61, 74, 77, 97, 10, 90, 8, 29, 82, 81, 51, 58, 7, 88, 99, 1, 37, 89, 93, 84, 52, 94, 2, 35, 5, 48, 87, 47, 14, 6, 79, 18, 67, 76, 70, 9, 43, 65, 69, 23, 55, 11, 53, 78, 50, 30, 13, 42] #+END_EXAMPLE OK, that's nice. But how about a number between 0 and 1? #+BEGIN_EXAMPLE 1.0 k r Result: 1: 0.636988102539 #+END_EXAMPLE OK, how about number between -50 and 50? For that we need to use what calc calls an interval form: #+BEGIN_EXAMPLE [ (Starts interval form) 50 (You can't just type -50 in calc) n (negate, givint -50) .. (the middle part of the interval form) 50] (closing the interval form) #+END_EXAMPLE What you see now in calc is: #+BEGIN_EXAMPLE [-50 .. 50] #+END_EXAMPLE And you could have just typed it in with: #+BEGIN_EXAMPLE '[-50 .. 50] #+END_EXAMPLE which would be a lot easier, really. #+BEGIN_EXAMPLE k r #+END_EXAMPLE This produces a random number from the bounds of the interval, in this case both -50 and 50 are possible, if you wanted them to be exlusive bounds, you'd use the form =(-50 .. 50)=. Finally, you can re-arrange a list: #+BEGIN_EXAMPLE '[1 2 3 4] (our starting vector) -1 (signals to use the vector above, could also be the size of the vector) k h Result: 1: [3, 1, 4, 2] #+END_EXAMPLE But =k a= will not give you more variants, unfortunately. *** Unit Conversion You load 16 tons, and what do you get? I mean, in kilograms. #+BEGIN_EXAMPLE M-x calc ' 16 tons (' to enter algebraic mode, so you can type out the units) u c kg (u c for "unit convert", and kg being the target unit). Result: 1: 14514.95584 kg #+END_EXAMPLE Calc treats units as special. If you added something, such as: #+BEGIN_EXAMPLE 3 + Result: 1: 14514.95584 kg + 3 #+END_EXAMPLE But you can remove the units from the above using: #+BEGIN_EXAMPLE u r (remove units) Result: 1: 14517.95584 #+END_EXAMPLE OK, that's all well and good. But I've always wondered how much is Grandpa Simpson's gas mileage when he said "My car gets 40 rods to the hogshead and that's the way I likes it." For that, we need to define the units. Calc knows about a lot of units, but maybe not the rod and hogshead. In fact, in the calc info pages, defining what a "rod" is the example for how to define your own units. Let's get started! #+BEGIN_EXAMPLE '16 ft (The equivalent to one rod) u d rod Rod (defines a new unit rod, with optional description "Rod") #+END_EXAMPLE Now a hogshead is a unit of measurement that varies by what liquid it contains. I don't know what the unit is for gasoline, but let's use sherry as a substitute, in which a hogshead is 245 liters. #+BEGIN_EXAMPLE '245 liters u d hogshead (don't bother with a description this time) '40 rod '1 hogshead / #+END_EXAMPLE Wait, what units should we be using? #+BEGIN_EXAMPLE u v (show the units table, a handy table of all units) u c mi/gal (the units come from the unit table) Result: 1: 1.87280731429e-3 mi / gal #+END_EXAMPLE But wait, we can do better. Why upgrade this measure to something that isn't even standard? Miles per gallon is just a bit better than rods per hogshead (in fact, that was what the original joke was about). #+BEGIN_EXAMPLE u c si (convert everything to scientific units) Result: 1: 796.212244896 / m^2 #+END_EXAMPLE Not that I understand this number, but at least in miles per gallon, I can see that that's not such great fuel economy, but what you do expect from Grandpa? OK, one more cool thing, then I'm out of here. Calc can split up numbers into multiple units. Here's 42 inches in feet and inches: #+BEGIN_EXAMPLE '42 in u c ft+in (Convert to a mixture of feet and inches) Result: 1: 3 ft + 6. in #+END_EXAMPLE Calc, you're sooo coool! *** Pi and Precision This one's about p and P and mostly about pi. First, let's pi it up: #+BEGIN_EXAMPLE M-x calc P (this gives you pi) Result: 1: 3.14159265359 #+END_EXAMPLE Well, I guess that's a reasonable pi. But, c'mon, this is calc. Can't we get a bit more digits? How about 100? #+BEGIN_EXAMPLE p 100 (sets precisions to 100) P (need to ask calc again for pi, it doesn't recalculate) Result: 1: 3.141592653589793238462643383279502884197169399375105820974944592307816406286208998628034825342117068 #+END_EXAMPLE Well, but actually evaluating it robs it of its never-ending charm. Let's just use it as a variable. How about calculating the area of a circle with a 5 km radius? #+BEGIN_EXAMPLE '5000 m 2 ^ 'pi (enter pi as a variable) * We get: 1: 25000000 m^2 pi #+END_EXAMPLE Yeah, sure that’s what I said I wanted, but I’ve changed my mind - now I want a number. #+BEGIN_EXAMPLE = 1: 78539816.3397448309615660845819875721049292349843776455243736148076954101571552249657008706335529267 m^2 #+END_EXAMPLE Whoops, looked like I forgot to set the precision back to normal. And I can't read this. Let's make it a bit nicer. #+BEGIN_EXAMPLE Control-_ (normal emacs undo) p 7 d g (turn digit grouping on) = Result: 1: 7.853983e7 m^2 #+END_EXAMPLE Oh, that's because I didn't have enough precision to render it without resorting to scientific notation. Let's just bump the precision up again. #+BEGIN_EXAMPLE Control-_ (undo, since we have to redo the pi conversion with more precision) p 10 = Result: 1: 78,539,816.35 m^2 #+END_EXAMPLE Ah, that's better. *** Strings Did you know you could work with strings in calc? For an example, let's find out what "Hello world" is in binary: #+BEGIN_EXAMPLE M-x calc d 2 (change the to binary mode) "Hello world (Enter the string "Hello world" which turns into a vector of numbers) Result: 1: [2#1001000, 2#1100101, 2#1101100, 2#1101100, 2#1101111, 2#100000, 2#1110111, 2#1101111, 2#1110010, 2#1101100, 2#1100100] #+END_EXAMPLE And similarly, we can convert back. If someone gave you the binary number: =01001000011011110110110001100001= and asked what the string was, I'd have no idea... but calc knows: #+BEGIN_EXAMPLE d " (changes to string mode) C-x b scratch (whaaa, leave calc?) 01001000011011110110110001100001 (enter the number we're parsing) C-a (go to the start of the line) C-x ( (start a macro) 2# (prefix the number with a binary indicator) C-u 8 C-f (Jump forward 8 characters) (insert a space to separate the numbers) C-x ) (end the macro) C-x e (repeat the macro) e e (repeat twice twice more) C- (set mark) C-a (goto beginning of line) C-x g (copy region into calc) Result: 1: "Hola" #+END_EXAMPLE And there you have it! Maybe there is an easier way to convert from the giant binary number to a vector of bytes, but I don't know it yet. *** Fractional Arithmetic This one is pretty short, but it's about one of my favorite features of calc: the ability to handle fractions as fractions instead of rendering them as real numbers. Quick, what's =5/8 + 9/21=? Um, ok... better start multiplying things... wait, let's just tell calc to do it. #+BEGIN_EXAMPLE M-x calc 5:8 (this is how you enter a fraction) 9:21 + Result: 1: 59/56 #+END_EXAMPLE So easy! If we want to convert it to a float you can do this: #+BEGIN_EXAMPLE c f (convert to float) Result: 1: 1.05357142857*10.^0 #+END_EXAMPLE And if you want it back as a fraction, then just do: #+BEGIN_EXAMPLE c F (convert to fraction) Result: 1: 59/56 #+END_EXAMPLE That's so awesome! You could also enter fractions this way: #+BEGIN_EXAMPLE m f (set fraction mode, integer division will result in fractions) 5 8 / Result: 1: 5/8 #+END_EXAMPLE Now you can live in the nice world of fractions as much as you like. It's a nice world, full of pleasant to look at integers taking up little horizontal space *** Algebra I think it's time to write about one of the amazing things that calc can do: algebra! Before we get into how to solve equations, I just want to write about on some cool things you can do with the calc display. Let's say you have a formula you want to work with =a + sqrt(b) = 5=. Let's enter that into calc: #+BEGIN_EXAMPLE M-x calc 'a + sqrt(b) = 5 (' starts algebraic mode). Result: 1: a + sqrt(b) = 5 #+END_EXAMPLE Well, that's not so surprising, that's what we put in. Kind of disappointing, though. Is that it calc? We love your brains, but what about your looks? That's important too! #+BEGIN_EXAMPLE d B (turn on calc-big-language mode) Result: ___ 1: a + V b = 5 #+END_EXAMPLE Hey, that's an ASCII square-root symbol. What other cool things can you do here? #+BEGIN_EXAMPLE a^2 Result: 2 1: a 3:4 (enter the fraction 3/4) Result: 3 1: - 4 #+END_EXAMPLE Here's how to get back: #+BEGIN_EXAMPLE d N (calc-normal-language) #+END_EXAMPLE And an alternative, in which all operators are explicitly represented as functions: #+BEGIN_EXAMPLE 1: a + sqrt(b) = 5 (re-enter the formula) d U (calc-unformatted-language) Result: 1: eq(add(a, sqrt(b)), 5) #+END_EXAMPLE But, wait, did you think that's all? What if you wanted to enter that equation in Mathematica? #+BEGIN_EXAMPLE d M (calc-mathematica-language) Result: 1: a + Sqrt[b] == 5 #+END_EXAMPLE Ooh! Calc! Do c++ next! #+BEGIN_EXAMPLE d C (calc-c-language) 1: a + sqrt(b) == 5 #+END_EXAMPLE Latex! #+BEGIN_EXAMPLE d L (calc-latex-language) Result: 1: a + \sqrt{b} = 5 #+END_EXAMPLE I could keep going, but trust me, there's more. And you can even define your own languages by constructing syntax tables, but I won't get into that now. *** More on Algebra Jim is 42 years old. He has one brother, and their total age is 100. What is the brother's age? OK, this isn't a very hard problem, but let's just introduce calc algebra by solving it. #+BEGIN_EXAMPLE M-x calc '42 + x = 100 (' to enter algebraic input) a S x (solve for x) Result: 1: x = 58 #+END_EXAMPLE Let's make this harder. Jim and Dan's ages sum to 100. Jim is 5 years older than Dan. How old are they? #+BEGIN_EXAMPLE '[j + d = 100, d + 5 = j] a S j,d Result: 1: [j = 52.5, d = 47.5] #+END_EXAMPLE Nice! And of course it can give you more than just numerical solutions: #+BEGIN_EXAMPLE 'sin(x) + tan(y) = pi / 2 a S y (solve for y) Result: 1: y = arctan(pi / 2 - sin(x)) #+END_EXAMPLE Sometimes there are more than one solution. For example: #+BEGIN_EXAMPLE 'x^2 = 25 a S x Result: 1: x = 5 #+END_EXAMPLE Wait, what happened to -5! That's a valid solution, why didn't calc tell us about it? What's happening here is that calc is telling us about the first valid thing it can find, which is basically how it operates. But you can always get everything: #+BEGIN_EXAMPLE 'x^2 = 25 a P x (find the polynomial solutions) Result: 1: [5, -5] #+END_EXAMPLE Sometimes there aren't a finite number of results because you aren't dealing with polynomials. You can just get a generalized solution: #+BEGIN_EXAMPLE 'sin(x)^2 = 25 H a S x (solve for x, giving the generalized solution) Result: 1: x = arcsin(5 s1) (-1)^n1 + 180 n1 #+END_EXAMPLE This uses the calc notation =n1=, which you just means any integer. You can also see another notation =s1= which means any sign. In this case =5 s1= means that that number can be 5 or -5. Looking at how awesome calc is, it's just a shame I never knew about it in high school... *** Financial I recently chatted with emacspeak creator T.V. Raman, and told him I was writing a series of short tutorials about calc. He is really a calc fanatic, and told me a story in which he astounded a loan officer by calculating scheduled loan payments with just a few keystrokes in calc. Raman is living proof that calc is a useful tool for so many situations, and it always pays to have emacs running. He also mentioned that he found the explanation in the calc tutorial about the financial functions to be the clearest he's ever read. So, yes, calc can do finance. Let's say that you were sitting in front of a loan officer, and she told you that for your loan of $500,000, you need to pay in 30 installments with a 5% interest rate. How much do you need to pay each month? Wait a second! Stop right there, loan officer! I have calc! #+BEGIN_EXAMPLE M-x calc 500000 (the amount of the loan) 30 (the number of payments) '5% (equivalent to typing 0.05) b M (calc-fin-pmt, computing the amount of periodic payments to amortize a loan) Result: 1: 25,000 #+END_EXAMPLE OK, but that's a bit obvious, since $25,000 is just 5% of $500,000. If the number of payments was much smaller, we'd get a larger value. Let's take another question: if you wanted to only pay $10,000 in each installment? How many installments would it take to pay off the loan? #+BEGIN_EXAMPLE '5% 10000 (the payment we want to make) 500000 (the loan amount) b # (calc-fin-nper, calculate the number of installments needed) Result 1: nper(0.05, 10,000, 500,000) #+END_EXAMPLE What? Oh, I see, I also go the message: "Payment too small to cover interest rate: 10000". Oh, right, 5% of $500,000 is already $25,000, so we'd never pay it off at that rate. What if we payed $50,000 instead? #+BEGIN_EXAMPLE '5% 50000 (the payment we want to make) 500000 (the loan amount) b # Result: 1: 14.2066908 #+END_EXAMPLE So, it would take just over 14 payments to pay off the loan. OK, one more cool one: Let's say you meet an investment banker who gives you the following deal. I've got a investment for you, she says. Just give me $100,000 and I'll give you $10,000 at the end of each year for the next 12 years. Assuming the interest rate will stay at 3% for the next 12 years. Is it a good deal? Hey, what are you asking me for? I have no idea! Calc knows, though, because it can tell you the break-even point for the cost of an investment that gives periodic payments. #+BEGIN_EXAMPLE '3% (the interest rate) 12 (the number of payments) 10000 (the payment you get each time) b P (calc-fin-pv, calculate the "present value" of the investment, the break-even point for the investment) Result: 1: 99,540.0399357 #+END_EXAMPLE In other words, the break-even point for the initial cost is $99,540. If the investment costs more than this, it's no good at that assumed interest rate. Better reject the deal. Trust calc more than any investment banker. This is just a small sampling of some of the financial calculations that calc can perform. The next time you are making an investment, fire up calc. You'll not only have confidence in the deal, you may just amaze someone with the power of emacs, just like T.V. Raman did. *** Calculus Quick, integrate =2x + sin(y)=! Well, frankly, it's been so long since I've done calculus by hand I can't remember anymore. Well, knowing calculus is good, but knowing calc is even more useful! #+BEGIN_EXAMPLE M-x calc '2x + sin(y) (The single quote enters algebraic mode) a i y (Calculate the integral with respect to y) Result 1: 2 x y - 180 cos(y) / pi #+END_EXAMPLE You can also integrate over specific regions by using C-u a i, whereupon it will prompt you for the start and end point of the integration. As the manual mentions, the results are often not as simplified as they could be. Calc is impressive, but it isn't as sophisticated as Mathematica. An example of some issues are if we just take the derivative of the integral we just calculated. We should get back to our original formula. #+BEGIN_EXAMPLE a d y (Calculate the derivative with respect to y) Result: 1: 2 x + 3.14159265358 sin(y) / pi #+END_EXAMPLE Clearly this should be 2x + sin(y), but calc may have made an error. OK, let's make calc do something cool so we can forget this unfortunate incident. Hey, how about making a Taylor series of a function? #+BEGIN_EXAMPLE '2x + sin(y) (re-enter the formula) a t y 6 (Calculate the Taylor series of a term, over y, for 6 terms) Result: 1: 2 x + y - y^3 / 6 + y^5 / 120 - y^7 / 5040 + y^9 / 362880 #+END_EXAMPLE This isn't a bad approximation, see [[https://www.google.com/search?q=y+-+y%5E3+%2F+6+%2B+y%5E5+%2F+120+-+y%5E7+%2F+5040+%2B+y%5E9+%2F+362880][Google’s answer]] for comparison. So, yes, calc can do college-level math, even if the answers aren't perfectly simplified. It's not Mathematica, but it is free and integrated into emacs, so it's definitely nice to have. *** Bit Manipulation Quick! What bits are set on the number 925817? What, are you going to convert it to binary and note positions of 1s? Ha! I laugh at such primitive techniques. #+BEGIN_EXAMPLE M-x calc 925817 b u (unpack the bits into a vector) Result 1: [0, [3 .. 6], 13, [17 .. 19]] #+END_EXAMPLE How many bits is that? #+BEGIN_EXAMPLE v # (count the number of items in a vector) Result: 1: 9 #+END_EXAMPLE This is convenient! So yes, calc has some nice functions for binary numbers. The interesting thing about calc's binary number functions aren't just that you can do bitwise operations such as AND and OR, but it has the notion of a word size that it works with. Well, it'd have to do things like NOT. Let's check it out. First, we'll see what the number 925817 looks like in binary. #+BEGIN_EXAMPLE 925817 d 2 Result: 1: 2#11100010000001111001 #+END_EXAMPLE This is nice, but it'd be better to see the whole word. #+BEGIN_EXAMPLE d z (Display leading zeroes) Result: 1: 2#00000000000011100010000001111001 #+END_EXAMPLE Ah, that's more like it. The word size by default is 32 bits, as you can see. Or, wait, can you see? Hard to count. Let's verify it. #+BEGIN_EXAMPLE d 0 (go back to base-10 mode) 0 (we start with 0) b n (calculate the not) Result: 1: 4294967295 #+END_EXAMPLE Now we already know how to count the 1's... #+BEGIN_EXAMPLE b u v # Reuslt: 1: 0000000032 #+END_EXAMPLE Whoops, looks like we still have leading 0s. But we've confirmed it, so let's just let it go for now. Now, let's see what the number 925817 is if you reverse all the bits. I can't take credit for this particular bit of cleverness, this technique comes straight from calc's info pages. #+BEGIN_EXAMPLE d z (get rid of leading 0s) 925817 b u (unpack into a vector) 31 - (tranform each bit position by subtracting it from 31, the tab just switches the items around on the stack) b p (repack the vector) Result: 1: 2651090944 #+END_EXAMPLE Woody Allen once praised New York by saying how he loves that you can go to Chinatown and eat a crab in the middle of the night, but in reality what kind of crazy person would need to do that? I feel the same way about all these features. Will I really ever need to reverse the bits of a number? Not sure, but I do love the way that calc has me covered for whatever I really want to do. By the way, want to go to 64-bit mode? Just change the word size. #+BEGIN_EXAMPLE b w 64 (change the word size to 64) #+END_EXAMPLE Now let's reverse the bits of 925817 again to see what we get. It'll be amusingly huge! #+BEGIN_EXAMPLE 925817 b u (unpack into a vector) 63 - (tranform each number by subtracting it from 31, the tab just switches the items around on the stack) b p (repack the vector) Result: 1: 11386348903201767424 #+END_EXAMPLE Ah, that's what it was. I was just about to give that same answer myself. One more cool thing. If you give a negative word size, calc will interpret binary number as 2's complement numbers. For example: #+BEGIN_EXAMPLE b w 32 (set the word size to 32) 2 (just to choose a simple number) b n (bitwise not) Result: 1: 4294967293 #+END_EXAMPLE And now with 2's complement! #+BEGIN_EXAMPLE b w -32 (set the word size to -32, in other words, a 2's complement version of 32 bit) 2 b n Result: 1: -3 #+END_EXAMPLE Hope this helps you twiddle those bits in all the ways that make you happy.