Medium to Large Scale Python Deployments

[Domenic R. Merenda]

"Michael Chermside" <mcherm at destiny.com> wrote in message
news:mailman.1023460091.14067.python-list at python.org...
> Folks, I think this is a pretty useful thing to collect.

> In fact, it might be really nice to collect a list of large (successful)
> Python projects and post it somewhere so the Python community could use
> it for advocacy.

   Michael, I agree with you.  I was formerly the Vice President of Business
Development for BeOpen.com, while we had PythonLabs there, and I can tell
you that we were compiling such a list for that very reason.  Using the
WayBack Machine, I was able to grab the list.  Here ya go:


Digital Creations
http://web.archive.org/web/20001119052700/http://www.digicool.com/
Digital Creations built its flagship product, Zope, in Python. Zope is an
advanced, highly adaptable web application server and content management
system in production use across the globe. Zope enables web developers and
content managers to rapidly deploy interactive, dynamic, scalable,
community-oriented web sites. Zope is available at
http://web.archive.org/web/20001119052700/http://www.zope.org/.


 eGroups (acquired by Yahoo!)
http://web.archive.org/web/20001119052700/http://www.egroups.com/
eGroups empowers people to share interests and ideas by delivering the
world's best web-based email group service, a technology built in Python.
Email groups facilitate group communication through the ease and convenience
of a user's email account. Python has allowed eGroups to rapidly evolve its
service offering as well as enable it to scale to unprecedented levels.

eGroups decision to use Python helped them achieve rapid growth, leading to
their acquisition by another market leader, Yahoo!.


 Google.com
http://web.archive.org/web/20001119052700/http://www.google.com/
Google's mission is to organize the world's online information and to make
it universally accessible and useful.

Google rocketed to the top of the web portal industry by making extensive
use of Python throughout its web crawler and search engine technologies
(later optimized for further performance in C). Google's web crawler now
spans the entire public Internet, with over 1 billion indexed URL's. The
total number of searches served by Google exceeds 13 million per day.


 Helix Code
http://web.archive.org/web/20001119052700/http://www.helixcode.com/
Helix Code is a startup building a free Internet desktop based on the GNOME
desktop platform, for Linux and Unix. The company is headed by the leading
devevelopers of the GNOME project. Helix Code ships a variety of Python
programming interfaces for the GNOME libraries and its associated graphical
widget set GTK+.


 Hewlett Packard e-speak
http://web.archive.org/web/20001119052700/http://www.e-speak.net/
HP used Python to build a truly open Market Making Broker with XML/HTTP
interfaces at all major interface points. This broker was far ahead of its
time, working across a large space of market makers. Initially, Python was
used only to produce the first prototype, with plans for a production
version to be done in Java. After the Java-intensive product proved less
stable with even fewer features than the Python prototype, Java work was
discontinued and the Python prototype was reworked in order to meet critical
deadlines.


 IBM
http://web.archive.org/web/20001119052700/http://www.ibm.com/
IBM uses Python on OS/390 UNIX System Services to facilitate automated
testing of Internet tools.

IBM is using Python to create the business practice logic for the factory
tool control applications that manage a semiconductor plant. Code that had
been previously written in an internal, proprietary scripting language has
been given new life via its port to Python.


 Infoseek (acquired by Go Network)
http://web.archive.org/web/20001119052700/http://www.infoseek.com/
Ultraseek Server, Infoseek's commercial search engine product, is
implemented as an elaborate multi-threaded Python program, with the
primitive indexing and search operations performed by a built-in module.
Most of the program is written in Python, and both a built-in spider and
HTTP server can be customized with additional Python code. The program
contains over 11,000 lines of Python code while the user interface contains
over 17,000 lines of Python-scripted HTML templates.


 ITI
http://web.archive.org/web/20001119052700/http://www.iti-oh.com/pdi
ITI's Product Data Interoperability (PDI) business provides software tools
and services enabling companies to effectively exchange, reuse, and share
CAD/CAM/CAE/PDM and related engineering data among diverse software
applications used throughout a manufacturing enterprise and it's supply
chain. ITI has had amazing success with Python.


 Lawrence Livermore National Laboratories
Los Alamos National Laboratory
http://web.archive.org/web/20001119052700/http://www.llnl.gov/
http://web.archive.org/web/20001119052700/http://www.lanl.gov/
A group at the Lawrence Livermore National Laboratories based its new
numerical engineering environment, CDAT, on Python, replacing a home-grown
scripting language of ten-year standing. CDAT used Python's exceptionally
versatile design to integrate C, C++, and FORTRAN subroutines, affording
climate research scientists an effortless analysis of global climate model
data sets.


 MCI Worldcom
http://web.archive.org/web/20001119052700/http://www.mciworldcom.com/
The Data Subnet Manager, MCI Worldcom's internal frame relay configuration
and provisioning suite, uses JPython for regression testing, interactive
debugging, and rapid prototyping of its Java Client. They also use Python in
their automated build processes and CGI scripts.


 Microsoft eShop
http://web.archive.org/web/20001119052700/http://www.eshop.com/
eShop, Inc. was originally formed in December 1993 to produce desktop
electronic commerce software. By November 1995, Pandora's Box had been
opened -- the Internet was something that had to be addressed directly,
rather than simply accounted for, and eShop had to move quickly or lose its
competitive advantages. By the time eShop was acquired by Microsoft, less
than a half-dozen engineers had built an entire Web-based shopping system
with Python in about four months. The resulting server was stable,
maintainable, robust, and scalable showing no compromise in quality for such
a brief development period.


 NASA
http://web.archive.org/web/20001119052700/http://www.nasa.gov/
Johnson Space Center uses Python as the standard scripting language in its
Integrated Planning System. Efforts are underway to develop a modular
collection of tools for assisting shuttle pre-mission planning and replacing
older tools written in PERL and shell dialects. Python will also be
installed in the new Mission Control Center to perform auxiliary processing
integrated with a user interface shell. Ongoing developments include an
automated grammar based system, whereby C++ libraries may be interfaced
directly to Python via compiler techniques. This technology can be extended
to other languages in the future.


 Penguin Computing
http://web.archive.org/web/20001119052700/http://www.penguincomputing.com/
Penguin Computing is a leading manufacturer of Linux servers for web,
database, and other enterprise uses. Much of the software developed at
Penguin for internal use and for shipping with their servers is developed in
Python -- after other languages proved too limiting and difficult to
maintain.


 PyBiz
http://web.archive.org/web/20001119052700/http://www.pybiz.com/
Python enabled a small team of senior architects at PyBiz to produce and
deliver both XDisect and eContentMgr on short schedule. It was estimated
that it would have required 3 times the build time and 1.5 times the number
of developers to deliver comparable functionality in Java.

Building in Python allowed delivery of a working product early in the life
cycle, providing more time to evolve the product towards customers' needs as
well as optimize performance of critical areas.


 Real Networks
http://web.archive.org/web/20001119052700/http://www.real.com/
RealNetworks has developed a Python binding for their RealMedia client. It
is used extensively in load testing and feature testing of both their server
and client on all of their supported platforms. Additionally, RealNetworks'
build system and bug tracking system have very significant Python
components.


 Red Hat
http://web.archive.org/web/20001119052700/http://www.redhat.com/
In the highly competitive Linux market, easy-to-use and flexible
installation technologies have become a key business differentiator. Red
Hat, a leader in Linux and Open Source technologies, identified Python as a
unique and powerful tool for building their installer, Anaconda.


 Sapient
http://web.archive.org/web/20001119052700/http://www.shn.net/
Sapient Information Systems built the Sapient Health Network for WebMD with
Python. The service provides chronically and terminally ill people with
information related to their illness. All of the backend web programming
uses Python, as do a large number of internal tools.


 Sonics
http://web.archive.org/web/20001119052700/http://www.sonicsinc.com/
Sonics engineers use Python exclusively to construct graphical and
configuration tools (the FastForward Development Environment) which provide
the layout of micro-network-based system-on-chips. Their experiences over
the years with many languages and systems led them to conclude that, with
the available time and resources, they could never have developed the
technology without Python.


 Yahoo! Mail and Four11
http://web.archive.org/web/20001119052700/http://www.yahoo.com/
Yahoo! acquired a company called Four11, a technology providing both an
Internet White Pages and a web-based e-mail reading system (renamed from
Rocketmail to Yahoo! Mail). Both of these services were written in Python to
obtain strong performance and flexibility. The Internet White Pages holds a
large number of email address records and the complete US residential
telephone directory -- so searching performance is key. Having acquired both
411 and eGroups, Yahoo! now relies on a wide range of Python-driven
technologies to provide the highly interactive experience its users expect.


 80-20 Software
http://web.archive.org/web/20001119052700/http://www.80-20.com/
80-20 Software used Python to rapidly develop a scalable and extensible
document management system that sits on top of the popular Microsoft
Exchange messaging technology. By using Python, they were able to develop
and release to market a feature rich client/server solution in under a year
with only 4 developers. Python provided an excellent prototyping tool for
the server technology, which originally was supposed to be converted to C++
for production use. 80-20 found during its testing that only a few
components needed to be in C++ for greater speed and that more than 90% of
the prototype code could be used unchanged without any significant
performance impact.


> Alternatively, if such a list already exists, let me know where, and
> I'll volunteer to help update and promote it.
>
> Domenic's
>    ERP system   >100      --

   Let's not call it that. :-)  And I will check upstairs when we get back
from our vacation to see if it's okay to release the number of lines in the
code.  Don't be surprised if it's treated like a National Security issue.
(Shh!  If they know how many lines are in the code, they can CRACK US!)


--
Domenic R. Merenda
Editor, The Edge Report
http://www.edgereport.com