Differing Usage of Python and Ruby
I love using Ruby for quick one-off scripts that don’t even deserve to be placed
in a file. For instance, I did something rather stupid the other night (deleted
all the .specification files of my installed RubyGems), and, while my gems all
still worked fine, gem has no idea they existed, which is a Bad Thing.
Anyways, I wanted to reinstall all of my gems:
┌─[pearson@Bragi] - [~] - [Tue Aug 03, 01:42]
└─[$]> ls /usr/lib/ruby/gems/1.9.1/gems
addressable-2.1.2 do_sqlite3-0.10.2 linecache19-0.5.11 ruby_core_source-0.1.4
ansi-1.2.2 erasmus-0.0.3 liquid-2.1.2 ruby-debug19-0.11.6
api_cache-0.2.0 extlib-0.9.15 maruku-0.6.0 ruby-debug-base19-0.11.23
archive-tar-minitar-0.5.2 facets-2.8.4 mechanize-1.0.0 rubyforge-2.0.4
bluecloth-2.0.7 fast-stemmer-1.0.0 mg-0.0.8 rubygems-update-1.3.7
bson-1.0.4 gemcutter-0.6.1 mime-types-1.16 ruby-prof-0.7.3
builder-2.1.2 gherkin-2.1.5 mongo-0.18.3 ruby-prof-0.8.2
cheat-1.3.0 gherkin-2.2.0 mongo-1.0.6 showoff-0.2.4
classifier-1.3.3 gist-1.2.1 mongo_ext-0.18.3 sinatra-1.0
columnize-0.3.1 git-1.2.5 mongo_ext-0.19.3 stemmer-1.0.1
compass-0.10.3 gli-1.1.1 net-scp-1.0.2 syntax-1.0.0
contest-0.1.2 haml-3.0.15 net-ssh-2.0.23 term-ansicolor-1.0.5
crack-0.1.8 hoe-2.6.1 nokogiri-1.4.3.1 test-unit-2.1.1
cucumber-0.8.5 httparty-0.6.1 octopi-0.2.8 thor-0.14.0
curb-0.7.7.1 hub-0.0.0 open4-1.0.1 thoughtbot-shoulda-2.11.1
data_objects-0.10.2 jekyll-0.6.2 polyglot-0.3.1 ticgit-0.3.6
diff-lcs-1.1.2 jeweler-1.4.0 rack-1.2.1 ticgit-2010.02.08
directory_watcher-1.3.2 json-1.4.3 rake-0.8.7 treetop-1.4.8
dm-aggregates-0.10.2 json_pure-1.2.0 rdiscount-1.6.5 trollop-1.16.2
dm-aggregates-1.0.0 json_pure-1.4.3 rdoc-2.5.9 turn-0.7.0
dm-core-0.10.2 light_mongo-0.4.0 RedCloth-4.2.3 xml-simple-1.0.12
dm-core-1.0.0 limgur-2.0.0 repl-1.0.0 yard-0.5.8
dm-mongo-adapter-0.2.0.pre3 linecache-0.43 rest-client-1.6.0
While I probably could have done this with sed, Ruby’s a bit more familiar to
me, and allows me to play around in irb to make sure I’ve got things right.
sudo gem install $(ruby -e 'Dir.entries("/usr/lib/ruby/gems/1.9.1/gems").map {|d| d[/[A-z-]+/]}.map {|d| d.chop if !d.nil?}.each {|d| puts d}')
sudo gem install do_sqlite3 linecache19 open4 ruby-debug19 ruby-debug-base19
and that was it. Awesome.
The same thing in Python would’ve been something like
import os
import re
files = os.listdir("/usr/lib/ruby/gems/1.9.1/gems")
files = [re.match(r"([A-z-]+)", file).group()[:-1] for file in files]
for file in files:
print file
or, in one-line style:
python -c 'import os; import re; for file in os.listdir("/usr/lib/ruby/gems/1.9.1/gems"): print re.match(r"([A-z-]+)", file).group()[:-1]'
…except for the fact that I apparently can’t put for statements on one line
if they’re after anything else.
Anyways, I prefer Python (over Ruby) for any program that has me working with
another programmer, either directly or through a (small) library. You see,
Python has wonderful things like docstrings and help(), as well as a community
that values documentation so much they actually write it. Write documentation?
Willingly? Crazy, I know, but it really spoils you.
With Ruby, you often end up with some API documentation thrown up, and that’s about it. Even the Ruby core docs (which are pretty good) have nothing on the Python equivalent. And hell, the Python module for regular expressions not only has a rather long section detailing the meaning of various characters in regexes, but has a link to a full-fledged regex guide. This kind of documentation fanaticism trickles down to the one-man libraries in a way that makes me not cringe when having to learn how to call someone else’s code in a way that it does what I want.
So yeah, when it comes to writing things by myself, I love Ruby. But sometimes, the values I share with the Python community makes it worth putting up with crap like len.
