CoinOpSpace.com - Classic Arcade Game Forum

Classic Arcade, Pinball and Console Collecting Gaming Community

Raw Browser to Server Communications Code. (See full image.)
(a.k.a. HOWTO create a transcript from Ning Chat.)

Some of you wanted to understand a little better how I captured the Jeff Kulczycki chat. Here is an overview.

I used a web programmers diagnostic tool named Fiddler. This program logs all traffic sent to and from your web browser. This helps programmers understand the communications between the browser and the server.

Since Ning Chat is hosted in a browser, I found I could use Fiddler to log all of the traffic sent from my browser to Ning's IM server (ningim.com). Lo and behold, each chat message is transmitted from ningim.com back to the browser on line at a time!

Unfortunately, there is a whole lot of information included with each message from the server that we don't need. (See Figure 1.) Wading through all that information, we get to the information we want. (See Figure 2.)

Now, I export the raw text of all of the chat message package to a text file. (See Figure 3, and sample raw browser to server communications log.)

I tried manually stripping out the bulk of the junk, after an hour I was maybe 5% of the way through the log. So, I decided to try using a magical search and replace trick called Regular Expressions. Using Regular Expressions I can find text that "looks like a date" and replace it with "nothing". Using tricks like that get 80-90% of the nonsense stripped.

At this point I am left with a much more manageable log file. The last 10% is a lot of manual formatting.

I think next chat I will set my computer up to log the chat. That way I only have to read the chat while I am editing it. ;-)

The first time took me 2.5 hours after the chat ended; plus the duration of the chat; plus 3-4 hours of research to find a way to log the chat. All told I probably spent 7-8 hours.

However, I am going to try to provide the transcripts for at least one to two more chats. It is my hope that I refine my technique down to where I can take a chat log and clean it up in 1.5 hours. I think 1.5 hours is a reasonable amount of time to spend on a chat transcript. If all goes well, I may even be able to get it down to an hour by writing an AppleScript to do some of the initial log cleanup for me. :-)

Thank you for your support and encouragement everyone. I hope I didn't bore you too much. :-)

Brian Hanifin

Tags: capture, chat, howto, log, transcript

Comment

You need to be a member of CoinOpSpace.com - Classic Arcade Game Forum to add comments!

Join CoinOpSpace.com - Classic Arcade Game Forum

Brian Hanifin Comment by Brian Hanifin on April 16, 2009 at 7:49pm
@Alan: Thank you very much for the offer. I have just enough experience with Regular Expressions that I was able to create a set of patterns that stripped 85% of the "garbage" within 15 minutes. :-)

Obviously the Steve Wiebe transcript will be important content for the site. So it will be reviewed by the site management and a higher layer of polish applied before it is posted.
Alan Comment by Alan on April 16, 2009 at 5:32pm
Any chance you caught the Steve Wiebe chat? If so, I'm pretty handy with perl so I might be able to lend a hand in formatting it.
Brian Hanifin Comment by Brian Hanifin on March 16, 2009 at 4:55pm
I posted a Raw barely formatted version of the transcript on my CoinOpSpace Blog.

www.coinopspace.com/profiles/blogs/jeff-kulczycki-creator-chat
Chris A (driph) Comment by Chris A (driph) on March 15, 2009 at 11:02pm
Thanks, I appreciate the time you've put into doing these... can't always get out of the office in time for the chats, so the transcripts are the next best thing.

Has the Jeff K transcript been posted? Transcripts of all the chat sessions would be great to keep for historical purposes, and it'd be worth building out an area dedicated solely to their preservation (like say, a transcript forum where each new chat session is posted once it's been cleaned up).

© 2010   Created by Jeff Rothe.   Powered by .

Badges  |  Report an Issue  |  Terms of Service