Programmers' preferences for package names

Fri, Jul 29, 2022

5 minute(s)

Are there trends in choosing package names in various programming ecosystems? Do package authors choose names for their packages that are alliterated with the name of the programming language? Let’s venture to find out.

First let’s install a couple of useful packages.

python -m pip install requests beautifulsoup4 pandas matplotlib

python -m pip install requests beautifulsoup4 pandas matplotlib

We can “bucket” the package names by their starting letter and count the number of packages in each bucket, i.e. a frequency plot. The following is the code I’m using:

Show Code

%matplotlib inline
import matplotlib.pyplot as plt
import numpy as np

plt.rc("font", size = 8)

from collections import defaultdict

reference = {}
total_reference = 0

def frequency_plot(items, lang, kind="packages"):
    bucket = defaultdict(lambda: 0)
    colors = defaultdict(lambda: "grey")
    for i in items:
        i = i.strip()
        bucket[i[0].lower()] += 1
    if kind=="packages":
        colors[lang[0].lower()] = "orange"
    keys = [k for k in sorted(list(bucket.keys())) if k.isalpha()]
    fig, axs = plt.subplots(1,1,figsize=(16,10),facecolor='#fffff8')
    ax = axs
    ax.bar(keys, [bucket[k] for k in keys], color=[colors[k] for k in keys], )
    total = sum(bucket[k] for k in keys)
    rects = ax.patches

    global reference
    global total_reference
    if kind != "packages":
        reference = bucket
        total_reference = total

    for rect, k in zip(rects, keys):
        height = rect.get_height()
        if kind == "packages":
            v = round(bucket[k] / total * 100 - reference[k] / total_reference * 100, 1)
            s = ["+", "-"][int(v < 0)]
        else:
            v = round(bucket[k] / total * 100, 1)
            s = ""
        if kind == "packages":
            label = f"{round(bucket[k] / total * 100, 1)}%\n({s}{abs(v)}%)"
        else:
            label = f"{s}{abs(v)}%"
        ax.text(
            rect.get_x() + rect.get_width() / 2, height*1.01, label, ha="center", va="bottom"
        )

    ax.set_title(f"Frequency count of {kind} in {lang} (Total: {total})")
    ax.set_ylabel("Count")
    ax.set_xlabel("First Letter")
    ax.set_facecolor("#fffff8")
    return ax

%matplotlib inline
import matplotlib.pyplot as plt
import numpy as np

plt.rc("font", size = 8)

from collections import defaultdict

reference = {}
total_reference = 0

def frequency_plot(items, lang, kind="packages"):
    bucket = defaultdict(lambda: 0)
    colors = defaultdict(lambda: "grey")
    for i in items:
        i = i.strip()
        bucket[i[0].lower()] += 1
    if kind=="packages":
        colors[lang[0].lower()] = "orange"
    keys = [k for k in sorted(list(bucket.keys())) if k.isalpha()]
    fig, axs = plt.subplots(1,1,figsize=(16,10),facecolor='#fffff8')
    ax = axs
    ax.bar(keys, [bucket[k] for k in keys], color=[colors[k] for k in keys], )
    total = sum(bucket[k] for k in keys)
    rects = ax.patches

    global reference
    global total_reference
    if kind != "packages":
        reference = bucket
        total_reference = total

    for rect, k in zip(rects, keys):
        height = rect.get_height()
        if kind == "packages":
            v = round(bucket[k] / total * 100 - reference[k] / total_reference * 100, 1)
            s = ["+", "-"][int(v < 0)]
        else:
            v = round(bucket[k] / total * 100, 1)
            s = ""
        if kind == "packages":
            label = f"{round(bucket[k] / total * 100, 1)}%\n({s}{abs(v)}%)"
        else:
            label = f"{s}{abs(v)}%"
        ax.text(
            rect.get_x() + rect.get_width() / 2, height*1.01, label, ha="center", va="bottom"
        )

    ax.set_title(f"Frequency count of {kind} in {lang} (Total: {total})")
    ax.set_ylabel("Count")
    ax.set_xlabel("First Letter")
    ax.set_facecolor("#fffff8")
    return ax

# English

For a reference case, let’s plot the distribution of words in the English language, per the list in /usr/share/dict/words on my MacOS 12.5.

Show Code

with open("/usr/share/dict/words") as f:
    words = set(f.read().split())

frequency_plot(words, "/usr/share/dict/words", kind="words")

with open("/usr/share/dict/words") as f:
    words = set(f.read().split())

frequency_plot(words, "/usr/share/dict/words", kind="words")

plt.savefig("./images/word-names.png", dpi=300, transparent=True);

plt.savefig("./images/word-names.png", dpi=300, transparent=True);

# Python

For Python, we can get the list of packages on PyPi using https://pypi.org/simple and use a HTML parser to get all the links on that page.

Show Code

import requests
from bs4 import BeautifulSoup

r = requests.get("https://pypi.org/simple")
soup = BeautifulSoup(r.text, 'html.parser')
packages = set()
for link in soup.find_all("a", href=True):
    packages.add(link.contents[0])

frequency_plot(packages, "Python")

import requests
from bs4 import BeautifulSoup

r = requests.get("https://pypi.org/simple")
soup = BeautifulSoup(r.text, 'html.parser')
packages = set()
for link in soup.find_all("a", href=True):
    packages.add(link.contents[0])

frequency_plot(packages, "Python")

plt.savefig("./images/python-package-names.png", dpi=300, transparent=True);

plt.savefig("./images/python-package-names.png", dpi=300, transparent=True);

The difference in percent of names of Python packages starting with “p” and words in the English language starting with “p” is +3.5%. The largest positive difference however is with the letter “d”, with “+4.2%”. I’m personally a little surprised it isn’t higher.

# Julia

When you install a package using Julia, it downloads a general registry into your “home” directory, and we can traverse that directory only one level deep.

Show Code

import os
packages = set()
for root, folders, files in os.walk(os.path.expanduser("~/.julia/registries/General/")):
    for folder in folders:
        if len(folder) > 1 and len("~/.julia/registries/General".split("/")) + 2 < len(root.split("/")):
            packages.add(folder)

frequency_plot(packages, "Julia")

import os
packages = set()
for root, folders, files in os.walk(os.path.expanduser("~/.julia/registries/General/")):
    for folder in folders:
        if len(folder) > 1 and len("~/.julia/registries/General".split("/")) + 2 < len(root.split("/")):
            packages.add(folder)

frequency_plot(packages, "Julia")

plt.savefig("./images/julia-package-names.png", dpi=300, transparent=True);

plt.savefig("./images/julia-package-names.png", dpi=300, transparent=True);

The difference in percent of names of Julia packages starting with “j” and words in the English language starting with “j” is +0.8%.

# Rust

https://crates.io conveniently has a data-access page that links to the latest dump which contains a csv file with the names of all the packages.

Show Code

import pandas as pd
packages = set(pd.read_csv("~/Downloads/2022-07-29-020018/data/crates.csv")["name"].dropna())

frequency_plot(packages, "Rust")

import pandas as pd
packages = set(pd.read_csv("~/Downloads/2022-07-29-020018/data/crates.csv")["name"].dropna())

frequency_plot(packages, "Rust")

plt.savefig("./images/rust-package-names.png", dpi=300, transparent=True);

plt.savefig("./images/rust-package-names.png", dpi=300, transparent=True);

In Rust, the difference is +3.9%, and it is the highest positive difference amongst all the letters.

# R

For R, similar to Python, we can parse the HTML from https://cran.r-project.org/web/packages/available_packages_by_name.html:

Show Code

import requests
from bs4 import BeautifulSoup

r = requests.get("https://cran.r-project.org/web/packages/available_packages_by_name.html")
soup = BeautifulSoup(r.text, 'html.parser')
packages = set()
for link in soup.find_all("a", href=True):
    packages.add(link.contents[0])

frequency_plot(packages, "R")

import requests
from bs4 import BeautifulSoup

r = requests.get("https://cran.r-project.org/web/packages/available_packages_by_name.html")
soup = BeautifulSoup(r.text, 'html.parser')
packages = set()
for link in soup.find_all("a", href=True):
    packages.add(link.contents[0])

frequency_plot(packages, "R")

plt.savefig("./images/r-package-names.png", transparent=True, dpi=300)

plt.savefig("./images/r-package-names.png", transparent=True, dpi=300)

In R, the difference is +6% for the letter “r”. This is the biggest difference we’ve seen so far.

# NPM

For NPM packages, https://replicate.npmjs.com/_all_docs contains a 228 MB json that contains all the packages.

Show Code

import json
import os
with open(os.path.expanduser("~/Downloads/_all_docs")) as f:
    data = json.loads(f.read())

packages = set()
for elem in data["rows"]:
    assert(elem["id"] == elem["key"])
    packages.add(elem["id"].split("/")[-1])

frequency_plot(packages, "NPM")

import json
import os
with open(os.path.expanduser("~/Downloads/_all_docs")) as f:
    data = json.loads(f.read())

packages = set()
for elem in data["rows"]:
    assert(elem["id"] == elem["key"])
    packages.add(elem["id"].split("/")[-1])

frequency_plot(packages, "NPM")

plt.savefig("./images/npm-package-names.png", dpi=300, transparent=True);

plt.savefig("./images/npm-package-names.png", dpi=300, transparent=True);

# Conclusion

In all these cases, there is a greater percentage of packages whose name starts with the same letter as the name of the programming language compared to the average distribution of words in the English language.

However, I don’t think it is enough to conclusively say that programmers prefer alliterating the package names with the programming language. It would be nice to test this assumption on C, C++, and Go (and possibly other languages), but these languages, as far as I can tell, don’t have central registries.