WHY DOESN’T ANYONE USE MY LIBRARY? 

OR… 

HOW TO BUILD SOMETHING USEFUL AND HELP IT GAIN ADOPTION BY NURTURING IT FOR YEARS AND NOT BITING OFF MORE THAN YOU CAN CHEW. 

No content

@JIAARO 

WHY DOESN’T ANYONE USE MY LIBRARY? 

3 REASONS 

3 REASONS 1. They’ve never heard of it 

3 REASONS 1. They’ve never heard of it 2. It doesn’t fit their needs 

3 REASONS 1. They’ve never heard of it 2. It doesn’t fit their needs 3. Using something else instead 

STORY TIME 

THIS’LL BE EASY sound1 = open(“talking.wav”, “rb”).read() 

THIS’LL BE EASY sound1 = open(“talking.wav”, “rb”).read() sound2 = open(“music.mp3”, “rb”).read() 

THIS’LL BE EASY sound1 = open(“talking.wav”, “rb”).read() sound2 = open(“music.mp3”, “rb”).read() combined = sound1 + sound2 

THIS’LL BE EASY sound1 = open(“talking.wav”, “rb”).read() sound2 = open(“music.mp3”, “rb”).read() combined = sound1 + sound2 out_f = open(“this_will_totally_work.???”, “wb”) 

THIS’LL BE EASY sound1 = open(“talking.wav”, “rb”).read() sound2 = open(“music.mp3”, “rb”).read() combined = sound1 + sound2 out_f = open(“this_will_totally_work.MP3”, “wb”) out_f.write(combined) 

THIS’LL BE EASY sound1 = open(“talking.wav”, “rb”).read() sound2 = open(“music.WAV”, “rb”).read() combined = sound1 + sound2 out_f = open(“this_will_totally_work.WAV”, “wb”) out_f.write(combined) 

THIS’LL BE EASY # coding: -*- utf-8 -*- import wave sound1 = wave.open(“talking.wav”, “rb”) sound2 = wave.open(“music.wav”, “rb”) out_f = wave.open(“this_will_totally_work.WAV”, “wb”) # Who even knows what this does ¯\_(
)_/¯ out_f.setparams(sound1.getparams()) # manual accounting ಠ_ಠ sound1_frames = sound1.getnframes() sound2_frames = sound2.getnframes() out_f.setnframes(sound1_frames + sound2_frames) out_f.writeframes(sound1.readframes(sound1_frames)) out_f.writeframes(sound2.readframes(sound2_frames)) out_f.close() 

No content

RE-EVALUATE MY DECISIONS AND/OR LIFE 

THIS’LL BE EASY ▸ Channels (Mono, Stereo) ▸ Sample Rate (44.1 kHz, 48 kHz, 22,050 kHz, …0 ▸ Bit depth (8 / 16 / 24 / 32 bit) 

THIS’LL BE EASY import wave import audioop sound1 = wave.open(“talking.wav”, “rb”) sound2 = wave.open(“music.wav”, “rb”) out_f = wave.open(“this_will_totally_work.WAV”, “wb”) sound1_frames = sound1.getnframes() sound2_frames = sound2.getnframes() sound1_data = sound1.readframes(sound1_frames) sound2_data = sound2.readframes(sound2_frames) sound2_data_converted, _ = audioop.ratecv( sound2_data, sound2.getsampwidth(), sound2.getnchannels(), sound2.getframerate(), sound1.getframerate(), None ) out_f.setparams(sound1.getparams()) out_f.setnframes(sound1_frames + sound2_frames) out_f.writeframes(sound1_data) out_f.writeframes(sound2_data_converted) out_f.close() 

No content

No content

No content

3 REASONS 1. They’ve never heard of it 2. It doesn’t fit their needs 3. Using something else instead 

3. THEY’RE USING SOMETHING ELSE 

3. THEY’RE USING SOMETHING ELSE ▸ What are people using now? ▸ What are the criticisms? ▸ What do they like? 

A GOOD API is worth it 

from pydub import AudioSegment sound1 = AudioSegment.from_wav(“sound.wav”) sound2 = AudioSegment.from_mp3(“music.mp3”) combined = sound1.append(sound2, crossfade=1000) combined.export(“tada.mp3”, format=“mp3”) 

Opportunity is missed by most people because it is dressed in overalls and looks like work. ▸ Thomas Edison 

Opportunity is missed by most people because it is dressed in overalls and looks like work. ▸ Henry Dodd 

Opportunity is missed by most people because it is dressed in overalls and looks like work. ▸ Anonymous 

The successful man was out and on the job long before opportunity came a-knocking. And this same opportunity, by the way, is ofttimes disguised as hard work ▸ Anonymous 

Showing up is 80 percent of life. ▸ Woody Allen 

Showing up is 80 percent of life. ▸ Marshall Brickman 

Showing up is 80 percent of life. ▸ Woody Allen 

No content

BUILD Libraries, NOT FRAMEWORKS 

No content

No content

DON’T BITE OFF MORE THAN YOU CAN CHEW 

No content

No content

1. Github 2. Stack Overflow: pydub tag via RSS 3. PyPI: for pip install 

GITHUB ▸ octopress: Barebones landing page ▸ issues: Bugs/Features ▸ docs: markdown files in the repo ▸ AUTHORS file: has my email and twitter 

1. Github 2. Stack Overflow: pydub tag via RSS 3. PyPI: for pip install 

THE LONG HAUL 

DON’T FORGET ABOUT YOUR RESEARCH 

STORY TIME (CONTINUED) 

No content

LUCK 

RECAP 

RECAP ▸ Solve a real problem 

RECAP ▸ Solve a real problem ▸ Nice API 

RECAP ▸ Solve a real problem ▸ Nice API ▸ Don’t overcommit 

RECAP ▸ Solve a real problem ▸ Nice API ▸ Don’t overcommit ▸ Keep showing up 

THANKS! James Robert @jiaaro pydub.com [email protected] [email protected]