diff --git a/.github/workflows/yaml-lint.yaml b/.github/workflows/yaml-lint.yaml
index 50a37311..a1b86a46 100644
--- a/.github/workflows/yaml-lint.yaml
+++ b/.github/workflows/yaml-lint.yaml
@@ -7,7 +7,7 @@ jobs:
     runs-on: ubuntu-latest
     strategy:
       matrix:
-        python-version: [3.8]
+        python-version: [3.11]
 
     steps:
     - uses: actions/checkout@v3
@@ -17,8 +17,9 @@ jobs:
         python-version: ${{ matrix.python-version }}
     - name: Install dependencies
       run: |
+        apt-get install -y build-essential libicu-dev
         python -m pip install --upgrade pip
-        pip install yamllint
+        pip install yamllint pycld2 pyicu-binary
         if [ -f requirements.txt ]; then pip install -r requirements.txt; fi
     - name: Lint _config.yml with yamllint
       run: |
diff --git a/Makefile b/Makefile
index 9c629cf0..8f1e6e55 100644
--- a/Makefile
+++ b/Makefile
@@ -5,13 +5,21 @@ all : commands
 commands :
 	@grep -h -E '^##' ${MAKEFILE_LIST} | sed -e 's/## //g' | column -t -s ':'
 
+_data/glossary.yml : ./glossary.yml
+	@mkdir -p _data
+	@cp $< $@
+
+sort-glossary : _data/glossary.yml
+	@yamllint glossary.yml
+	@python utils/sort-glossary.py
+
 ## site : rebuild GitHub Pages site locally.
-site : _data/glossary.yml
+site : sort-glossary
 	rm -rf .jekyll-cache .jekyll-metadata _site
 	bundle exec jekyll build
 
 ## gh-site : builds the website for GitHub pages (part of the GH Actions workflow)
-gh-site : _data/glossary.yml
+gh-site : sort-glossary
 	@rm -rf _gh-site
 	@mkdir -p _gh-site
 	@cp -r `ls -A | grep -v '.git' | grep -v '_gh-site' | grep -v '_site'` _gh-site
@@ -19,7 +27,7 @@ gh-site : _data/glossary.yml
 	@cp $< _gh-site/$<
 
 ## serve : serve GitHub Pages site locally.
-serve : _data/glossary.yml
+serve : sort-glossary
 	rm -rf _site
 	bundle exec jekyll serve -I
 
@@ -27,7 +35,7 @@ serve : _data/glossary.yml
 clean :
 	@rm -rf _site
 	@find . -name '*~' -exec rm {} \;
-	@rm -f _data/glossary.yml
+	@rm -rf _data/*
 
 ## check : check glossary consistency.
 check :
@@ -37,8 +45,3 @@ check :
 ## checkall : check glossary consistency including missing terms in all languages.
 checkall :
 	@python utils/check-glossary.py -A _config.yml glossary.yml
-
-# Create copy of glossary file for GitHub Pages site.
-_data/glossary.yml : ./glossary.yml
-	@mkdir -p _data
-	@cp $< $@
diff --git a/utils/sort-glossary.py b/utils/sort-glossary.py
index 299490ce..f38a6518 100644
--- a/utils/sort-glossary.py
+++ b/utils/sort-glossary.py
@@ -6,7 +6,8 @@
 
 import icu
 
-languages = [
+# set up supported languages
+LANGUAGES = [
     ('aa', 'Afar'),
     ('ab', 'Abkhazian'),
     ('af', 'Afrikaans'),
@@ -212,13 +213,15 @@
     ('zu', 'Zulu')
 ]
 
-def _sort_terms(count_dict):
+
+def _sort_terms(count_dict, data_path):
     # sort and reassign terms
     for lang in count_dict:
+        # check 2-letter language codes vs 3-letter language codes
         # std_lang = standardize_tag(lang)
         # print(f"{lang} -> {std_lang} -> {Language.get(std_lang).to_alpha3()}")
 
-        # create a locale from the language code and sort the terms with a collator
+        # create a locale from the language code and a collator to perform sorting
         icu_locale = icu.Locale(lang)
         collator = icu.Collator.createInstance(icu_locale)
 
@@ -227,16 +230,18 @@ def _sort_terms(count_dict):
             lang_path = data_path.joinpath(lang)
             lang_path.mkdir(parents=True, exist_ok=True)
 
-            # sort
+            # sort and store sorted terms separate from the original list
             sorted_terms = sorted(count_dict[lang]["terms"], key=collator.getSortKey)
             count_dict[lang]["sorted_terms"] = sorted_terms
     return count_dict
 
-def _setup_dict(glossary):
+
+def _setup_dict(glossary, data_path):
+    # data structure to hold counts and terms
     count_dict = {}
-    lang_codes = []
 
-    for cc in languages:
+    lang_codes = []
+    for cc in LANGUAGES:
         count_dict[cc[0]] = {}
         count_dict[cc[0]]["count"] = 0
         count_dict[cc[0]]["name"] = cc[1]
@@ -248,6 +253,7 @@ def _setup_dict(glossary):
     # total number of glossary terms
     # print(len(glos))
 
+    # count terms and store them in the data structure
     for slug in glossary:
         for lang in slug.keys():
             if lang in lang_codes:
@@ -259,18 +265,24 @@ def _setup_dict(glossary):
                         "def": slug[lang]["def"]
                     }
                 )
-    return _sort_terms(count_dict)
+
+    # return the data structure including sorted terms
+    return _sort_terms(count_dict, data_path)
+
 
 def _build_lang_glossary(count_dict):
     glossary_by_lang = {}
     for lang in count_dict:
         sorted_glossary = []
+
+        # process the data structure to create a new sorted glossary per language
         for sorted_term in count_dict[lang]["sorted_terms"]:
             if sorted_term in count_dict[lang]["term_entry_map"]:
                 term_map = count_dict[lang]["term_entry_map"][sorted_term]
                 slug = term_map["slug"]
                 _def = term_map["def"]
 
+                # use an OrderedDict to retain insertion order
                 sorted_glossary.append(OrderedDict({
                         "slug": slug,
                         lang: {
@@ -278,31 +290,46 @@ def _build_lang_glossary(count_dict):
                             "def": _def
                         }
                     }))
+
+        # only include languages with terms
         if sorted_glossary:
             glossary_by_lang[lang] = sorted_glossary
     return glossary_by_lang
 
+
 def setup_yaml():
     """ https://stackoverflow.com/a/8661021 """
     def represent_dict_order(self, data):
         return self.represent_mapping('tag:yaml.org,2002:map', data.items())
     yaml.add_representer(OrderedDict, represent_dict_order)
 
-# load main glossary file
-glos = yaml.safe_load(Path('glossary.yml').read_text())
-data_path = Path("_data/")
 
-# sort terms
-count_dict = _setup_dict(glos)
+def main():
+    try:
+        # get path
+        current_path = Path(__file__).resolve()
+
+        # load main glossary file
+        data_path = current_path.parent.parent.joinpath("_data/")
+        glossary_path = data_path.joinpath("glossary.yml")
+        glos = yaml.safe_load(glossary_path.read_text())
+
+        # sort terms
+        sort_dict = _setup_dict(glos, data_path)
+
+        # rebuild glossary per language
+        sorted_glossary_by_lang = _build_lang_glossary(sort_dict)
+
+        # setup yaml for outputting
+        setup_yaml()
+        for lang in sorted_glossary_by_lang:
+            yaml.dump(sorted_glossary_by_lang[lang], Path(f'_data/{lang}/glossary.yml').open('w'))
 
-# rebuild glossary per language
-sorted_glossary_by_lang = _build_lang_glossary(count_dict)
+        # output counts
+        # pprint.pprint(count_dict)
+    except Exception as e:
+        print(e)
 
-# setup yaml for outputting
-setup_yaml()
-for lang in sorted_glossary_by_lang:
-    pprint.pprint(sorted_glossary_by_lang[lang])
-    yaml.dump(sorted_glossary_by_lang[lang], Path(f'_data/{lang}/glossary.yml').open('w'))
 
-# output counts
-# pprint.pprint(count_dict)
+if __name__ == '__main__':
+    main()