Related Entries

India PyCon 2009
Quick wallpaper changer
Load testing with Grinder
Adding namespace to XML
Opera RSS to OPML

« Joy of Python: Setting Goals
» Annual reviews

Perl to Python: 20 stages

A post in the newsgroup that explains it all. And my work history when it comes to programming languages.

I wanted to post a link to 20 stages of conversion from Perl to Python that appeared in c.l.python newsgroup. It is an excellent piece and at least my experience is exactly as explained there.

I am one of those less fortunate mortals, who keeps getting attacked by other languages often. Here’s my work history so far.

  1. Once upon a time, I was happily coding in Oracle, SQL, PL/SQL, Perl, C and Oracle Forms.
  2. Then this thing called the web happened and to save paper, thought it will be nice to make HTML output. May be because by then I started working for a conservation oriented non-profit, this didn’t work out exactly as planned. Instead of people taking dull output from line printers, people starting printing out formatted HTML pages using laser printer.
  3. Then came PHP. Found it to be ideal for generating dynamic pages.
  4. Java came, kept coming back. Even now I don’t understand what the fuss is about it. It is a good language, but unless you work for software development shops, why go to such great length to write pages and pages of code. Object oriented or whatever! I liked it though.
  5. While looking for PHP or Java platforms to develop a slightly complex web application, discovered Zope. Liked it within the first 7 days.
  6. Though a bit scared of Python, went through the 20 steps mentioned above.
  7. It’s been 3 years. Java keeps coming back! ASP with VBScript makes guest appearances in my nightmares at work.
  8. Every two years stabler versions of Windows appear. Every time, within a month, I feel nostalgic about the time I used to work with HP-UX and other Unixes.
  9. Cold Fusion tried. HyperWave keeps trying.
  10. Oracle is now at a stage where Larry Ellison will put anything that passes by into the database engine.

What did I learn?

By the time you hear about difficulty to hire people proficient in Zope and Python; and by the time you hear about the virtues of super scalar architectures using J2EE and .Nut, you can actually go and make working applications that easily exceed user’s needs with 1/10th the code required in those "Enterprise Class Technologies". Jarno Virtanen has some points on why people prefer particular languages. I found it pretty interesting.

I do want to see Oracle’s tie up with Java in a bit more detail.