[Chicago] guido @ google

Brian Ray bray at sent.com
Thu Dec 22 17:05:51 CET 2005

On Dec 22, 2005, at 8:08 AM, Ed Summers wrote:

> Well, it's not a book -- but surely this bodes well for the python  
> community:
> http://www.oreillynet.com/pub/wlg/8821

I would be interested to hear how Google uses Python. And what use  
they will have for the writer of Python. There have been many job  
offerings from Google asking for people with Python experience.

Here in the office, we are having a heated language flame war debate  
concerning web programming--keep in mind most of is are C hackers and  
really none of us are web experts. My argument was that the web is  
just like any other program and does not require some statically  
linked ball of twine--like PHP. PHP is fast because it is one big  
static library. You can get as much performance with a language which  
scales well and allows clustering, like Python. Also their seems to  
be more thought focused on program design and less on. "how do I  
write this next line of code?". I think I mentioned the grasshopper  
versus team of ants analogy. I then mentioned that Google may be  
using Python in this way. The result was the others thought I was  
wrong in that Google must not be using Python because thought Python  
was not fast enough.

On speed, the is the Python philosophy still write everything in  
Python then rewrite anything not quick enough in "C". I wonder if  
Guido will be writing "C" for Python at Google.

One thing that intrigues me about Google is their concentration on  
making data available quickly, cleanly and concisely.  Much of this  
idea is part of they Python experience. Then when I take a look at  
other great web technology build with Python, especially those I can  
take apart because I have the source code, like for example just  
about any Django application, I see that this technology too also  
uses some of the well founded theories derived from Python.

With plain C/C++, I immediately know who wrote what piece of code  
before looking at blame or cvs logs. Their own techniques are  
signatures and bad practices are on everything they write. With  
Python, I can not always do the same thing--which is a good thing if  
your working in a large group. Maybe, the restrictiveness causes  
people to write code which matches each others. I only then can tell  
who wrote which piece of Python code when I take a step back and  
start looking at class diagrams and looking at architecture decisions.

Still, Guido at Google cool. Thanks Ed!

-- Brian Ray

More information about the Chicago mailing list