Re: Chandler, Python, speed

2009-03-08 Thread Tim Wintle
On Sat, 2009-03-07 at 22:05 +, Ville M. Vainio wrote:
 Alan G Isaac wrote:
 
  3. Chandler is not really an email client.  So specifically,
  which of its functionalities is it slow, and what evidence
  if any is there that Python is causing this?
 
 I remember reading somewhere that the cause of slowness is/was
 architectural - perhaps it was that chandler was persisting too much stuff
 to disk, or something. In any case, this might help you google for more
 detail.

I've been using it, and the only major issue I've got is the long
shutdown times, which is caused by backing up a copy of _Everything_ to
disk, so that upgrades can happen cleanly - sure that's going to be
fixed though.

Startup time is a bit slow too, but it's designed to be left open all
the time, and it's fairly zippy once it's open IMO.

Tim Wintle


--
http://mail.python.org/mailman/listinfo/python-list


Re: Chandler, Python, speed

2009-03-08 Thread Ville M. Vainio
Ville M. Vainio wrote:

 Alan G Isaac wrote:
 
 3. Chandler is not really an email client.  So specifically,
 which of its functionalities is it slow, and what evidence
 if any is there that Python is causing this?
 
 I remember reading somewhere that the cause of slowness is/was
 architectural - perhaps it was that chandler was persisting too much stuff
 to disk, or something. In any case, this might help you google for more
 detail.

Here's the somewhere:

http://blog.chandlerproject.org/2008/05/21/rick-rawson-on-why-i-use-chandler-and-what-would-make-me-stop/

QQQ

Alex, you’re right that VM usage is a big factor, although there are cases
where the CPU is doing too much, too. So far as why this is, there are many
contributions. The biggest one is that architectural decisions taken early
on in the project meant that way too much data is being persisted. For
example, the app persists all schema and GUI layout information, leading to
on-disk and mapped memory bloat. (There is also redundancy in the data
stored on disk that leads to more of this kind of bloat).

QQQ


--
http://mail.python.org/mailman/listinfo/python-list


Re: Chandler, Python, speed

2009-03-07 Thread Ville M. Vainio
Alan G Isaac wrote:

 3. Chandler is not really an email client.  So specifically,
 which of its functionalities is it slow, and what evidence
 if any is there that Python is causing this?

I remember reading somewhere that the cause of slowness is/was
architectural - perhaps it was that chandler was persisting too much stuff
to disk, or something. In any case, this might help you google for more
detail.
--
http://mail.python.org/mailman/listinfo/python-list


Re: Chandler, Python, speed

2009-03-07 Thread alex23
On Mar 8, 8:05 am, Ville M. Vainio vivai...@gmail.com wrote:
 I remember reading somewhere that the cause of slowness is/was
 architectural - perhaps it was that chandler was persisting too much stuff
 to disk, or something. In any case, this might help you google for more
 detail.

My understanding was the Philip Eby's oft-quoted Python Is Not Java[1]
post was a direct response to how he saw the Chandler project being
developed:

I was recently looking at the source of a wxPython-based GUI
application, about 45.5KLOC in size, not counting the libraries used
(e.g. Twisted). The code was written by Java developers who are
relatively new to Python, and it suffers from some performance issues
(like a 30-second startup time). In examining the code, I found that
they had done lots of things that make sense in Java, but which suck
terribly in Python. Not because Python is slower than Java, but
because there are easier ways to accomplish the same goals in Python,
that wouldn't even be possible in Java.

[1]: http://dirtsimple.org/2004/12/python-is-not-java.html

--
http://mail.python.org/mailman/listinfo/python-list


Re: Chandler, Python, speed

2009-03-03 Thread Alan G Isaac

On Mar 2, 1:11 am, Paul Rubin wrote:
Mitch Kapor (of Lotus 1-2-3 fame) spent a lot of money hiring 
very sharp Python programmers to write an email client called 
Chandler, but from what I understand, progress so far has been 
disappointing, at least in part for performance reasons. 



Paul McGuire wrote:
Yipes, have you read the book (Dreaming in Code)?  
Chandler's problems were much more organizational than 
technical.



This discussion makes me curious about a couple things.

1. It is hard to see how Python could make a slow
email client.  Is that the actual claim?  Where is the
demand for computational speed coming from?

2. This is especially true for IMAP, where it seems any
computationally intensive work (e.g., sorting very large
mailboxes) will be done on the server.

3. Chandler is not really an email client.  So specifically,
which of its functionalities is it slow, and what evidence
if any is there that Python is causing this?

Thanks,
Alan Isaac

--
http://mail.python.org/mailman/listinfo/python-list