maintainers:
- Ivica Ico Bukvic ico@vt.edu
- Albert Graef aggraef@gmail.com
- Jonathan Wilkes jancsika@yahoo.com
- Downloads
- One Paragraph Overview
- Three Paragraph Overview
- Goals
- User Guide
- Relationship of Purr Data to Pure Data
- Build Guide
- Code of Conduct
- Project Governance
- Contributor Guide
- Human Interface Guidelines
- Core Pd Notes
- GUI Message Spec
Pure Data (aka Pd) is a visual programming language. That means you can use it to create software graphically by drawing diagrams instead of writing lines of code. These diagrams show how data flows through the software, displaying on the screen what text-based languages require you to piece together in your mind.
Pd has been designed with an emphasis on generating sound, video, 2D/3D graphics, and connecting through sensors, input devices, and MIDI as well as OSC devices.
Pd has a special emphasis on generating audio and/or video in real time, with low latency. Much of its design focuses on receiving, manipulating, and delivering high-quality audio signals. Specifically, the software addresses the problem of how to do this efficiently and reliably on general purpose operating systems like OSX, Windows, Debian, etc.-- i.e., systems designed mainly for multi-tasking.
Pd can easily work over local and remote networks. It can be used to integrate wearable technology, motor systems, lighting rigs, and other equipment. Pd is also suitable for learning basic multimedia processing and visual programming methods, as well as for realizing complex systems for large-scale projects.
Pd-L2ork has the following goals:
- Documentation. We like documentation. It's like code, except friendly.
- Be reliable. Binary releases must be usable for performances and installations. The git repo must always be in a workable state that can be compiled. Regressions must be fixed quickly.
- Be discoverable. Undocumented features are buggy. Missing help files are bugs. Patches for new functionality that lack documentation are spam.
- Be consistent. Consistent interfaces are themselves a kind of documentation. We like documentation, so it follows that we like consistent interfaces.
For a more in-depth look at Purr Data for new users and developers, see:
https://agraef.github.io/purr-data-intro/Purr-Data-Intro.html
For more resources see:
https://agraef.github.io/purr-data/
There are three maintained distributions of Pure Data:
- Purr Data. This is the 2.0 version of Pd-l2ork. It ships with lots of external libraries and uses a modern GUI written using HTML5.
- Pd-L2Ork 1.0, the version used by Ivica Bukvic for his laptop orchestra. Pd-l2ork 1.0 uses tcl/tk (and tkpath) for the GUI. You can find it here
- Pure Data "Vanilla". Miller Puckette's personal version which he hosts on his website and maintains. It doesn't come with external libraries pre-installed, but it does include an interface you can use to search and install external libraries maintained and packaged by other developers.
Packages for Windows and OSX:
https://github.com/jonwwilkes/purr-data/releases
Gnu/Linux installation instructions:
All of this needs to be done as root, using either sudo
or su
.
The following goes into /etc/pacman.conf:
[home_aggraef_Arch]
SigLevel = Never
Server = https://download.opensuse.org/repositories/home:aggraef/Arch/$arch
Then just run:
pacman -Sy
pacman -S purr-data
All of this needs to be done as root, using either sudo
or su
. Use "Debian_9.0" instead if you're running Debian Stretch rather than Buster; likewise if you're running Unstable or Testing.
You should import the repository key first, so that the packages can be updated automatically (this only needs to be done once):
wget -nv https://download.opensuse.org/repositories/home:aggraef/Debian_10/Release.key
apt-key add Release.key
Then add the repository to your apt sources as follows (taking Debian 10 as an example):
echo 'deb http://download.opensuse.org/repositories/home:/aggraef/Debian_10/ /' > /etc/apt/sources.list.d/home:aggraef.list
apt update
NOTE: This creates the source in a separate file /etc/apt/sources.list.d/home:aggraef.list. You can just remove this file when you don't need the repository any more.
Finally install the package:
apt install purr-data
All of this needs to be done as root, using either sudo
or su
. Use "Raspbian_9.0" instead if you're running Raspbian Stretch rather than Buster.
sudo su
You should import the repository key first, so that the packages can be updated automatically (this only needs to be done once):
wget -nv https://download.opensuse.org/repositories/home:aggraef/Raspbian_10/Release.key
apt-key add Release.key
Then add the repository to your apt sources as follows:
echo 'deb http://download.opensuse.org/repositories/home:/aggraef/Raspbian_10/ /' > /etc/apt/sources.list.d/home:aggraef.list
apt update
Finally install the package:
apt install purr-data
All of this needs to be done as root, using either sudo
or su
. Replace "18.04" with the version of Ubuntu you use.
sudo su
You should import the repository key first, so that the packages can be updated automatically (this only needs to be done once):
wget -nv https://download.opensuse.org/repositories/home:aggraef/xUbuntu_18.04/Release.key
apt-key add Release.key
Then add the repository to your apt sources as follows:
echo 'deb http://download.opensuse.org/repositories/home:/aggraef/xUbuntu_18.04/ /' > /etc/apt/sources.list.d/home:aggraef.list
apt update
Finally install the package:
apt install purr-data
NOTE: The openSUSE builds are somewhat experimental right now. In particular, the flite external is not supported, because the required dependencies are not available. Also note that some of the multimedia functionality (Gem video, in particular) will require proprietary media codecs and thus won't work in stock openSUSE. You will need to add third-party repositories like Packman to make these work.
We take Tumbleweed as an example here, if you're running Leap then replace "Tumbleweed" with "Leap_15.1" or whatever version you use.
sudo su
zypper addrepo https://download.opensuse.org/repositories/home:aggraef/openSUSE_Tumbleweed/home:aggraef.repo
zypper refresh
zypper install purr-data
NOTE: The instructions for Windows and OSX below talk about running the tar_em_up.sh
build
script, which is still the recommended way to build Purr Data right now.
However, Purr Data also has a new (and experimental) toplevel Makefile so that
just typing make
will build the package. You may find this easier. The
Makefile also offers the customary targets to clean (make clean
, or
make realclean
to put the sources in pristine state again) and to roll a
self-contained distribution tarball (make dist
). Please check the comments
at the beginning of the Makefile for more information.
Time to build: 10 minutes light install, 45 minutes to 1.5 hours full install Hard drive space required: roughly 2.5 GB
-
Install the dependencies
sudo apt-get install bison flex automake libasound2-dev \ libjack-jackd2-dev libtool libbluetooth-dev libgl1-mesa-dev \ libglu1-mesa-dev libglew-dev libmagick++-dev libftgl-dev \ libgmerlin-dev libgmerlin-avdec-dev libavifile-0.7-dev \ libmpeg3-dev libquicktime-dev libv4l-dev libraw1394-dev \ libdc1394-22-dev libfftw3-dev libvorbis-dev ladspa-sdk \ dssi-dev tap-plugins invada-studio-plugins-ladspa blepvco \ swh-plugins mcp-plugins cmt blop slv2-jack omins rev-plugins \ libslv2-dev dssi-utils vco-plugins wah-plugins fil-plugins \ mda-lv2 libmp3lame-dev libspeex-dev libgsl0-dev \ portaudio19-dev liblua5.3-dev python-dev libsmpeg0 libjpeg62-turbo \ flite1-dev libgsm1-dev libgtk2.0-dev git libstk0-dev \ libfluidsynth-dev fluid-soundfont-gm byacc
-
The gui toolkit may require installing the following extra dependencies sudo apt-get install gconf2 libnss3
-
Clone the Purr-Data repository (2 to 10 minutes)
git clone https://git.purrdata.net/jwilkes/purr-data.git
-
Compile the code (5 minutes to 1.5 hours) full
- to build only the core:
make light
(5 minutes) - to build core and all externals:
make all
(20 minutes to 1.5 hours) - to build everything except Gem:
make incremental
(10 to 20 minutes)
- to build only the core:
-
There should now be an installer file in the main directory of the repo. If you're using an apt-based Linux distribution it will be an apt package. Otherwise, it will be a tarball which you can unzip, enter, and run
make install
(as well asmake uninstall
to remove it).
To install using a pre-compiled binary, follow these instructions: http://l2ork.music.vt.edu/main/?page_id=56
To set up a development environment, first make sure you have the following package dependencies listed here: http://l2ork.music.vt.edu/main/?page_id=56
Then follow the steps outlined here: http://l2ork.music.vt.edu/main/?page_id=56#install-dev
Time to build: 50 minutes to 1.5 hours
Hard drive space required: roughly 2 GB
-
Install Homebrew (15 minutes) (asks for password twice-- once for command line tools, once for homebrew)
-
Install the dependencies (10 minutes):
brew install wget brew install autoconf brew install automake brew install libtool brew install fftw brew install python brew install lua brew install fluidsynth brew install lame brew install libvorbis brew install speex brew install gsl brew install libquicktime brew install pkg-config
-
Clone the Purr-Data repository (10 minutes)
git clone https://git.purrdata.net/jwilkes/purr-data.git
-
Change to the directory
cd purr-data/l2ork_addons
-
Run the installer (15 minutes)
./tar_em_up.sh -X
-
When the installer finishes, type
cd ..
-
There should now be a .dmg file in your current directory
Time to build: roughly 1.5 hours-- 30 minutes of this is for Gem alone
Hard drive space required to build: rougly 2.5 GB
Important note: check the name of your Windows user account. If it has a space in it-- like "My Home Computer" or "2nd Laptop", then stop. You may not use this guide. (Actually you can probably just install everything in ~/.. in that case, but I haven't tested doing it like that. Sorry. Get a better OS...)
-
Download and install msys2 (5 minutes)
There are two installers-- one for 32-bit Windows systems (i386) and one for 64-bit Windows (x_64). Be sure you know which version of Windows you are running and download the appropriate installer.
Note: don't run it after it installs. You'll open it manually in the next step. -
Download and install inno setup (5 minutes)
-
Run MinGW-w64 Win32 Shell (less than a minute)
msys2 adds three Start Menu items for different "flavors" of shell:- MinGW-w64 Win32 Shell <- click this one!
- MinGW-w64 Win64 Shell
- MSYS Shell
-
Install the dependencies (5-10 minutes)
Once the shell opens, we need to install the dependencies for building Purr Data. First we need to update all the packages:pacman -Syu
After closing and reopening the shell as prompted, you may need to do it again:
pacman -Syu
Now everything should be up-to-date. Issue the following command:
pacman -S autoconf automake git libtool \ make mingw-w64-i686-dlfcn mingw-w64-i686-fftw \ mingw-w64-i686-fluidsynth \ mingw-w64-i686-ftgl mingw-w64-i686-fribidi \ mingw-w64-i686-ladspa-sdk mingw-w64-i686-lame \ mingw-w64-i686-libsndfile mingw-w64-i686-libvorbis \ mingw-w64-i686-lua mingw-w64-i686-toolchain \ mingw-w64-i686-libjpeg-turbo \ rsync unzip wget
-
Download the source code (3-6 minutes)
Issue the following command to create a new directory "purr-data" and clone the repository to it:git clone https://git.purrdata.net/jwilkes/purr-data.git
-
Enter the purr-data/l2ork_addons directory (less than a minute)
cd purr-data/l2ork_addons
-
Finally, build Purr-Data (45-80 minutes)
./tar_em_up.sh -Z
-
Look in the top level directory of the Git repository and click the setup file to start installing Purr Data to your machine.
- No sarcasm, please
- Don't appear to lack empathy
- You can't live here. If you're spending hours a day writing Purr Data code or-- worse-- spending hours a day writing emails about code that has yet to be written, you're doing it wrong
- If working on something for the first time, ask to be mentored
- If no one asked you to mentor them, don't teach
- It is better to let small things go then to risk taking time away from solving bigger problems
It is a bad idea to break this Code of Conduct even if no one complains about your behavior.
- the three maintainers listed at the top of this document are the ones in charge of this project
- unanimous decisions are preferred
- 2 out of 3 can break a disagreement
- there will only ever be three maintainers of this project at any given time. If you'd like to temporarily step in as one of the three, send an inquiry to the list and we cany discuss it
Contributing is easy:
- Join the development list: http://disis.music.vt.edu/cgi-bin/mailman/listinfo/l2ork-dev
- Fork Purr Data using the gitlab UI and then try to build it from source for your own platform using the Build Guide above. If you run into problems ask on the development list for help.
- Once you have successfully built Purr Data, install it and make sure it runs correctly.
- Start making changes to the code with brief, clear commit messages. If you want some practice you can try fixing one of the bugs on the issue tracker labeled "good-first-bug"
- One you are done fixing the bug or adding your feature, make a merge request in the Gitlab UI so we can merge the fix for the next release.
A few guidelines:
- There should be a short and clear commit message for each merge request.
- Short and clear title and description are required for each merge request.
- There should be a short branch name related to the issue, like "update-readme".
- No prototypes, please. Purr Data's biggest strength is that users can turn an idea into working code very quickly. But a prototyping language that is itself a prototype isn't very useful. That means Purr Data's core code and libraries must be stable, consistent, well-documented, and easy to use.
- Develop incrementally. Small, solid improvements to the software are preferable to large, disruptive ones.
- Try not to duplicate existing functionality. For backwards compatibility Purr Data ships many legacy libraries which unfortunately duplicate the same functionality. This makes it harder to learn how to use Pd, and makes it burdensome to read patches and keep track of all the disparate implementations.
- Keep dependencies to a minimum. Cross-platform dependency handling is unfortunately still an open research problem. In the even that you need an external library dependency, please mirror it at git.purrdata.net so that the build system doesn't depend on the availability of external infrastructure.
Here are some of the current tasks:
- writing small audio/visual Pd games or demos to include in the next release
- skills needed: ability to write Pd programs
- status: I wrote a little sprite-based game that will ship with the next version of Pd-L2Ork. In it, the character walks around in an actual Pd diagram shoots at the objects to progress, and to make realtime changes to the music. What I'd like is to include a new, smallish game with each release that has a link in the Pd console. It can be a little demo or game, just something fun that shows off what can be done using Pure Data.
- designing/implementing regression test template
- skills needed: knowledge about... regression tests. :) But also some expertise in using Pd so that the tests themselves can be written in Pure Data. At the same time, they should be able to be run as part of the automated packaging process (i.e., in -nogui mode).
- status: some externals have their own testing environments, but they are limited as they require manual intervention to run and read the results inside a graphical window. We currently have a crude test system that at least ensures that each external library instantiates without crashing. Here's an email thread with Katja Vetter's design, which looks to be automatable: http://markmail.org/message/t7yitfc55anus76i#query:+page:1+mid:chb56ve7kea2qumn+state:results And Mathieu Bouchard's "pure unity" (not sure if this is the most recent link...): http://sourceforge.net/p/pure-data/svn/HEAD/tree/tags/externals/pureunity/pureunity-0.0/
- adding support for double precision to the external libraries that ship with purr-data
- skills needed: knowledge about data types in C language(specially float and double)
- status: the core classes of purr data and the freeverb~ external library have been changed to support both float and double but still the remaining external libraries only have support for single precision. The task ahead is to add double precision support to these external libraries. As per the current resources we have the merge requests that have been used to add double precision support to the core libraries: https://git.purrdata.net/jwilkes/purr-data/merge_requests?scope=all&utf8=%E2%9C%93&state=merged&author_username=pranay_36 And Katja Vetter's double precision patches to the pd-double project which were actually used for adding double precision support to the core libraries of purr-data. https://github.com/pd-projects/pd-double/commit/982ad1aa1a82b9bcd29c5b6a6e6b597675d5f300
Pd is a multi-window application that consists of three parts:
- A main window, called the "Pd Window" or "Console Window". This window displays informational and error messages for Pd programs.
- One or more "canvas" windows-- aka "patch" windows, used to display the diagrams that make up a Pd program.
- One or more dialog windows used to configure the various parts of Pd.
All should look simple and uncluttered. Although "canvas" windows cannot (yet) be traversed and edited using only the keyboard, all three parts of Pd should be designed so that they can be manipulated using only the keyboard.
It should also be possible to produce sound and interact when a new user runs program for the very first time. In every release, there should be a link at the bottom of the Console Window to a short game written in Pd that demonstrates one or more of the capabilities of the Pd environment. The game should be designed to be fun outside of its efficacy as a demonstration of Pd.
Pd ships with "DejaVu Sans Mono", which is used for the text in canvas windows. Fonts are sized to fit the hard-coded constraints in Pd Vanilla. This way box sizes will match as closely as possible across distributions and OSes.
These hard-coded sizes are maximum character widths and heights. No font fits these maximums exactly, so it's currently impossible to tell when looking at a Pd canvas whether the objects will collide on a system using a different font (or even a different font-rendering engine).
Dialogs and console button labels may use variable-width fonts. There is not yet a suggested default to use for these.
The console printout area currently uses "DejaVu Sans Mono". Errors are printed as a link so that the user can click them to highlight the corresponded canvas or object that triggered the error.
Nothing set in stone yet.
The following is adapted from Pd Vanilla's original source notes. (Found in pd/src/CHANGELOG.txt for some reason...)
Sections 2-3 below are quite old. Someone needs to check whether they even hold true for Pd Vanilla anymore.
First, the containment tree of things that can be sent messages ("pure data"). (note that t_object and t_text, and t_graph and t_canvas, should be unified...)
BEFORE 0.35:
m_pd.h t_pd anything with a class
t_gobj "graphic object"
t_text text object
g_canvas.h
t_glist list of graphic objects
g_canvas.c t_canvas Pd "document"
AFTER 0.35:
m_pd.h t_pd anything with a class
t_gobj "graphic object"
t_text patchable object, AKA t_object
g_canvas.h t_glist list of graphic objects, AKA t_canvas
Other structures:
g_canvas.h t_selection -- linked list of gobjs
t_editor -- editor state, allocated for visible glists
m_imp.h t_methodentry -- method handler
t_widgetbehavior -- class-dependent editing behavior for gobjs
t_parentwidgetbehavior -- objects' behavior on parent window
t_class -- method definitions, instance size, flags, etc.
1.0 C coding style. The source should pass most "warnings" of C compilers (-Wall on Linux, for instance-- see the makefile.) Some informalities are intentional, for instance the loose use of function prototypes (see below) and uncast conversions from longer to shorter numerical formats. The code doesn't respect "const" yet.
1.1. Prefixes in structure elements. The names of structure elements always
have a K&R-style prefix, as in ((t_atom)x)->a_type
, where the a_
prefix
indicates "atom." This is intended to enhance readability (although the
convention arose from a limitation of early C compilers.) Common prefixes are:
w_
(word)a_
(atom)s_
(symbol)ob_
(object)te_
(text object)g_
(graphical object)gl_
(glist, a list of graphical objects).
Also, global symbols sometimes get prefixes, as in s_float
(the symbol whose
string is "float"). Typedefs are prefixed by t_
. Most private structures,
i.e., structures whose definitions appear in a ".c" file, are prefixed by x_
.
1.2. Function arguments. Many functions take as their first
argument a pointer named x
, which is a pointer to a structure suggested
by the function prefix; e.g., canvas_dirty(x, n)
where x
points to a canvas
(t_canvas *x)
.
1.3. Function Prototypes. Functions which are used in at least two different files (besides where they originate) are prototyped in the appropriate include file. Functions which are provided in one file and used in one other are prototyped right where they are used. This is just to keep the size of the ".h" files down for readability's sake.
1.4. Whacko private terminology. Some terms are lifted from other historically relevant programs, notably "ugen" (which is just a tilde object; see d_ugen.c.)
1.5. Spacing. Tabs are 8 spaces; indentation is 4 spaces. Indenting curly brackets are by themselves on their own lines, as in:
if (x)
{
x = 0;
}
Lines should fit within 80 spaces.
2.0. Max patch-level compatibility. "Import" and "Export" functions are provided which aspire to strict compatibility with 0.26 patches (ISPW version), but which don't get anywhere close to that yet. Where possible, features appearing on the Mac will someday also be provided; for instance, the connect message on the Mac offers segmented patch cords; these will devolve into straight lines in Pd. Many, many UI objects in Opcode Max will not appear in Pd, at least at first.
3.0. Compatibility with Max 0.26 "externs"-- source-level compatibility. Pd objects follow the style of 0.26 objects as closely as possible, making exceptions in cases where the 0.26 model is clearly deficient. These are:
3.1. Anything involving the MacIntosh "Handle" data type is changed to use char * or void * instead.
3.2. Pd passes true single-precision floating-point arguments to methods; Max uses double. Typedefs are provided:
t_floatarg, t_intarg for arguments passed by the message system
t_float, t_int for the "word" union (in atoms, for example.)
3.3. Badly-named entities got name changes:
w_long --> w_int (in the "union word" structure)
3.4. Many library functions are renamed and have different arguments; I hope to provide an include file to alias them when compiling Max externs.
4.0. Function name prefixes. Many function names have prefixes which indicate what "package" they belong to. The exceptions are:
typedmess, vmess, getfn, gensym (m_class.c)
getbytes, freebytes, resizebytes (m_memory.c)
post, error, bug (s_print.c)
which are all frequently called and which don't fit into simple categories. Important packages are:
(pd-gui:) pdgui -- everything
(pd:) pd -- functions common to all "pd" objects
obj -- fuctions common to all "patchable" objects ala Max
sys -- "system" level functions
binbuf -- functions manipulating binbufs
class -- functions manipulating classes
(other) -- functions common to the named Pd class
5.0. Source file prefixes.
PD:
s system interface
m message system
g graphics stuff
d DSP objects
x control objects
z other
PD-GUI:
gui GUI front end
- Brackets on the same line as declaration or expression:
if (a) {
- Single line comments only:
//
- Use double-quotes for strings
- Use underscores to separate words of function names and variables
Purpose: a set of functions to communicate with the gui without putting language-specific strings (like tcl) into the C code. The new interface is a step toward separating some (but not all) of the GUI logic out from the C code. Of course the GUI can still be designed to parse and evaluate incoming messages as commands. But the idiosyncracies of the GUI toolkit can be limited to either the GUI code itself or to a small set of modular wrappers around sys_vgui.
The public interface consists of the following:
gui_vmess(const char *msg, const char *format, ...);
where const char *format
consists of zero or more of the following:
- f - floating point value (
t_float
) - i - integer (
int
) - s - c string (`char* )
- x - hexadecimal integer value, with a precision of at least six digits.
(hex value is preceded by an 'x', like
x123456
)
For some of Pd's internals like array visualization, the message length may vary. For these special cases, the following functions allow the developer to iteratively build up a message to send to the GUI.
gui_start_vmess(const char *msg, const char *format, ...);
gui_start_array(); // start an array
gui_f(t_float float); // floating point array element (t_float)
gui_i(int int); // integer array element (int)
gui_s(const char *str); // c string array element
gui_end_array(); // end an array
gui_end_vmess(); // terminate the message
The above will send a well-formed message to the GUI, where the number of array elements are limited by the amount of memory available to the GUI. Because of the complexity of this approach, it may only be used when it is necessary to send a variable length message to the GUI. (Some of the current code may violate this rule, but that can be viewed as a bug which needs to get fixed.)
The array element functions gui_f, gui_i, and gui_s may only be used inside an array. Arrays may be nested, but this adds complexity and should be avoided if possible.
The public functions above should fit any sensible message format. Unfortunately, Pd's message format (FUDI) is too simplistic to handle arbitrary c-strings and arrays, so it cannot be used here. (But if it happens to improve in the future it should be trivial to make a wrapper for the public interface above.)
The current wrapper was made with the assumption that there is a Javascript Engine at the other end of the message. Messages consist of a selector, followed by whitespace, followed by a comman-delimited list of valid Javascript primitives (numbers, strings, and arrays). For the arrays, Javascript's array notation is used. This is a highly idiosyncratic, quick-and-dirty approach. But the point is that the idiosyncracy exists in a single file of the source code, and can be easily made more modular (or replaced entirely by something else) without affecting any of the rest of the C code.