CodeToVideo is a unique LONG-TERM source code archival strategy.
It easily and beautifully transmogrifies your source code to the widest-supported audio/video media files capable of being shared, stored and viewed in ways source code has never been shared before!
This project built itself:
- Self-Built CodeToVideo project [1 min]
- The entire composer/composer project [30 min]
Virtually everything!!! See Supported Languages for specifics.
Certainly all of the big languages: JavaScript, C, C++, C#, PHP, Python, Ruby, JSON, XML, and many, many more!
It allows source code to be quickly, concisely and verifiably transmitted across space-time via mediums that heretofore were unavailable for the transmission of bulk text files, particularly source-code, in a readily readable manner.
Namely: YouTube videos, Twitter tweets, Facebook Posts, social media IMs, even Google Photos.
But also: BluRay players (almost all support h.264 playback), many modern DVD players, Google ChromeCast, virtually all NASes.
Unlike virtually all other professional classes, professional code craftsmen rarely, if ever, rewatch their own works.
While professional sports players, lawyers, and many more have realized for decades the power of JUST WATCHING your past actions, many coders have never even thought to do this, much less done it.
By quicly flitting in front of your eyes at 2.5 seconds per page of code, your subconscious will pick up on coding practices. If you watch the CodeToVideos of software afficiondos substantially more advanced than yourself, you will undoubtedly grow in proficiency yourself, as this occurs across all creative disciplines.
More importantly for long-term survivability of source code:
- Greater fidelity (due to both h.264 and BluRay error-correcting technologies) in case of bitrot over text.
- Even fuzzy text in images can be readable by future entities whereas bitrotted text files could not.
- Your code can be transferred to 35mm film reels and be viewable without any technological society at all.
Because I am a firm believer in reincarnation, this project has a timeline horizon of 1,250 years.
That means that I am targeting for my 21st Century source code written in the 2020s to be readable by both human and non-human intelligences as far out as the Year 3150 minimally.
While we have no idea what compression algorithms of today will still be accessible, or even if classical
computers based on silicon-transitors will still even exist, much less the concept of TEXTUAL DOCUMENTS
(due to global societal trends of less and less literacy in favor of audio/video):
The idea is that even in the FAR FUTURE, even with alien intelligences without so much the concept of a "text file", audio/video representations of source code should be both more future proof and easily accessed than expecting an entity wholly unfamiliar with the 21st Century of Earth to figure out zip files, or even ASCII, much less UTF-8 and xv.
Also, it is far more likely that h.264 and future world-web video encodings will persist far far longer than the written word.
When stored on BluRay or 35mm film reels in fire-proof, water-proof pouches in side safes, it should be virtually indestructable.
Burned BluRays stored outside of the Sun in individual cases, with 33% parity PAR2 files, should last in excess of 50-200 years (see [docs/bluray-longevity.png]. 35mm film stored properly should last eons and would not require any existing technological edifice to view (just Sunlight/torchlight).
We currently recommend storing each CodeToVideo as five separate copies with 33% parity each on each 25 GB one-layer BluRay disc, with each disc duplicated 2-3x and stored in separate geographical areas in secure locations (ideally waterpoof/fireproof pouches, individually to avoid bitrot). Because BluRay bitrot typically occurs 1% every 5 years on inferior medium and 0.5% every 10 on superior media, this should mean an average recoverability rate of 99.9% after 500 years.
Hopefully, your descendants will periodically transfer these archives to new media every 50-100 years.
Hopefully, there are digital archivists who like to do it this for fun, like me ;-) at least every 20.
code-to-video /path/to/MyProject
This will create a ~/Videos/CodeToVideo/MyProject.mp4 with the following attributes:
- 25 fps h264 at -crf 16 [near-ideal quality].
- 1 source code snashot every 2.5 seconds.
- No audio.
- ~3 MB of video for every 500 kb of source code.
The default seconds per page is 2.5 seconds. You cna change this by adding a second decimal parameter:
# 5 seconds per page
code-to-video /path/to/MyProject 5
You can show multiple pages per second by using a fraction of 1 fps:
# 2 pages per second
code-to-video /path/to/MyProject 0.5
# 3 pages per second
code-to-video /path/to/MyProject 0.3333
If you are only interested in future OCR recovery of the source code, you can completely disable any pause and show one page per video frame by passing in 0:
code-to-video /path/to/MyProject 0
This greatly reduces the size of the videos, but removes most practicality.
This is the ideal scenario for storing large amounts of source code or other text data on 35mm film reels and is what the GitHub Arctic Code Vault project does for its most important source code documents.
You can easily add an audio file to the video by passing it in as an optional third parameter.
code-to-video /path/to/MyProject 2.5 "/path/to/music.file"
The audio format will be reencoded to 128 kbps 48000 Hz AAC for maximum future-proofing.
CodeToVideo requires the following:
- Modern scrot
- alecthomas/chroma
- ffmpeg
- GNU Bash
- GNU nl
- GNU split
- GNU tail
- A color TTY terminal that supports ANSI colors
sudo cp bin/code-to-video
sudo mkdir /usr/share/codetovideo
sudo cp watermark.png /usr/share/codetovideo
On Arch Linux:
sudo pacman -S scrot
On Debian Linux:
sudo apt install scrot
Everywhere else:
git clone --depth=1 https://github.com/resurrecting-open-source-projects/scrot.git
cd scrot
./autogen.sh
./configure --prefix=/usr
make
strip src/scrot
sudo make install
See [docs/building.md] for further instructions.
This project is licensed under the Creative Commons Attribution International License v4.0.
Theodore R. Smith [email protected]
GPG Fingerprint: 4BF8 2613 1C34 87AC D28F 2AD8 EB24 A91D D612 5690
CEO: PHP Experts, Inc.
Reachable on Twitter.