[boston-pig] Fwd: Python Meetup Notes 6/8

[Chris Curvey]

Here's the notes from last night's meeting.  Thanks, Dan!  (And thanks to
all who attended.)

-Chris

---------- Forwarded message ----------
From: Dan Milstein <danmil at comcast.net>
Date: Jun 8, 2006 10:16 PM
Subject: Python Meetup Notes 6/8
To: Chris Curvey <ccurvey at gmail.com>

Chris,

Here's what I've got -- feel free to email it around to all and sundry:

In attendance, clockwise from me:
  - Me (Dan Milstein)
  - Charlie - working with IVR
  - Ned (who showed up a few minutes in)
  - Gerard -- interested in learning
  - Nate   -- the plone master
  - Sally  -- from the Nature Conservancy?
  - Jack   -- also interested in learning
  - Chris  -- um, you

We talked about libs / tools etc that we like:

  - Charlie likes ElementTree for dealing with XML

  - I talked about the beauty that is BeautifulSoup

And, apparently, many folks have needs for screen-scraping python
apps, since there was much nodding of heads.  I showed off my two-
liner for finding all the important rows of the table that has the
text "Regular Season" in it somewhere:

     reg_season_tbl = soup.firstText('&nbsp;Regular
Season').findParent('table')
     trs = reg_season_tbl.fetch('tr', { 'class':['ysprow1', 'ysprow2']})


  - Sally talked about nose and its auto-running of tests tricks

Jeff Winkler's blog has more.  She also mentioned AgileTesting.com.

She also had great things to say about PyGoogle -- the Python
interface to the Google API.


  - Chris had great things to say about ZSI - the Zolera Soap Interface

It let him talk to PayPal's web servicey SOAP system.  And there was
much rejoicing.


  - There was some general talk about generating AJAX-y client-side
JS from Python

None of it seems far along, but we name-dropped
    - PyJamas (pygwt)
    - Bling
    - Kukit

Ned raised a concern (a question?) about the utility of using server-
side Python to write your client-side scripting (which is sort of
what those systems do), since the two tend to deal with different
data universes.  Along those lines (at least in my mind), no one was
clear on how you test the code for events like dragging and dropping
and the like.


  - Many people had good things to say about Selenium for web testing

Ned raised that they have some issues with it because it can't
generate all the events in JS that they need (e.g. double-clicking).
I talked some about a different approach I've been using for my code,
which is more focused on spec'ing pre- and post-conditions in the
database, the session, etc, and less on driving a browser.  There is
definitely some ways in which Selenium complements what I've got (and
may do it better, in fact).  I made a personal note to check that
Selenium out one more time.


  - I raved about GraphViz

And, again, there was much nodding -- it is a rock star tool for the
graph visualization.  Go AT&T labs, go.


  - Ned likes SmartyPants -- turns regular quotes into curly quotes

He talked about his start-up company, Tabblo.  We should all go to
their website:

http://www.tabblo.com/

It's a photo/story-sharing site, but it is not Flickr.

They are using Django, but they can't keep up with the trunk.  He
explained what is afoot with the big Django magic-removal branch, but
they haven't gotten sync'd up with it.  On a personal note, for me,
this cleared up the question of whether the "magic-removal branch"
I'd been hearing about meant removing the magic, *or* some deeply
magic way of doing removals.  I am somewhat relieved to learn that it
is the former.  Onward.

Tabblo is going to try to move to the new, non-magic stuff, but it's
lacking some features they want/need.  For example, you can't do
model inheritance.

They use python imaging library (PIL) to generate thumbnails and what
not.

Ned has been using pdflib, which costs money, but generates pdf with
precise control over formatting (from all sorts of languages). The
python wrapper has the C nature.  But it works.  reportlab is maybe
worth looking at.

tabblo uses memcached -- we discussed memcached vs. squid and some
questions about caching part of a page and caching db results and the
like.


  - Nate talked about sIFR as an interesting way to get rich
multimedia that gracefully degrades

People use it for, e.g. headlines


  - Sally talked about S5 (?), which lets you nicely autogenerate
slideshow/presentation type stuff

Maybe via ReST ->  S5 -> latex -> pdf.


  - Charlie talked about AwareTek -- Ron Stevens -- lots of
tutorial / evangelist stuff

Dive into Python is maybe a good book to learn from.

  - Nate talked about BarCamp
    - Weekend event, no planned agenda
    - An unconference conference
    - Nate org'd one about building online communities for musicians
    - How to solve a rubik's cube using grid computing


That's what I've got. Hope it's helpful...
-Dan




-- 
The short answer is "Yes."  The long answer is "No."
-------------- next part --------------
An HTML attachment was scrubbed...
URL: /pipermail/boston-pig/attachments/20060609/1bdc5427/attachment.htm