Mapping hyperkitty users

I am working on the reputation system for HyperKitty. The reputation system comes to play when the content is generated by the community. We know which user posted which answer and we give points to the user depending on the response of community to his / her answer.

In hyperkitty, we get the archives from mailman. We offer users the power to vote the message through the web interface. But the users who create content are not necessarily the same who generate the content. And, for the reputation system to work correctly, it is very important for us to map the users who actually posted the content to the users who are consuming the content.

Update

After talking to wacky, I found that we are actually planning to have user for every author in the mailing list. Hence problem solved!

Reputation system

I am basically going to implement the reputation system for HyperKitty. It will determine, to an extent, the amount of subject matter expertise you have and the level of respect peers have for you. This will enable us to highlight the most active contributor of the list on list overview page.

We will use it to give more features to users having high reputation which won’t be generally available for low reputation users by default. Giving access to certain features after someone meets a certain reputation criteria will help us to maintain the quality and fight spam.

I need your ideas about implementing the reputation system in HyperKitty.


Relevant links:

http://meta.stackoverflow.com/a/7238

https://help.launchpad.net/YourAccount/Karma/

deploying hyperkitty using apache wsgi

Step 1: install mod_wsgi,

sudo apt-get install libapache2-mod-wsgi (ubuntu) or,

sudo yum install mod_wsgi (fedora)


Step 2: Clone latest HyperKitty code,

git clone https://github.com/syst3mw0rm/hyperkitty


Step 3: Create logs & python-egg directory

mkdir -p logs

mkdir -p .python-egg

chmod -R 777 .python-egg (make sure .python-egg directory is writable by wsgi process, more info)


Step 4: Edit httpd.conf

Add the following line in httpd.conf (generally present at /etc/apache2/httpd.conf)

Include “/path/to/application/apache/apache_django_wsgi.conf”


Step 5: Edit the settings file

Edit apache/settings_production.py to include production database details.


Step 6: Restart apache

sudo service apache2 restart

unit testing hyperkitty

As I have written a decent amount of code for HyperKitty it’s time for me to write unit tests for it. I am planning to write unit tests for each model, view and forms. I have created a tests directory inside application code to store multiple tests file. Following is the directory structure,

tests

├── __init__.py

├── test_forms.py

├── test_models.py

└── test_views.py

I will write all the view related test functions inside test_views.py and so on.

hyperkitty - modern archiver

In this post I will try to explain some of the essential features an archiver should possess.

If you are an open source contributor you must have used archiver for sure. Open source projects extensively use mailing lists for discussions and archiver stores all the messages that are posted on the list. Most of the mailing lists are powered by GNU Mailman. Pipermail is one of the most popular archiver for Mailman till date. It has served the purpose for more than a decade and now it’s time to move on.

The major problems with pipermail are  : http://wiki.list.org/display/DEV/ModernArchiving

In my opinion an archiver should do the following jobs in outstanding way :

Search
The primary purpose of an archiver is to store messages to search through the messages and give result as quickly as possible. The part of which depends on the search index because most of the people just Google for the answer and having neat and descriptive URLs helps a lot.


promoting good content
Second most important feature is the archiver should be intelligent enough to promote the good posts and suppress the spamy ones. In hyperkitty we are giving our users the power to upvote or downvote a message which will automatically bring upvoted content on the surface.


User interface
I can see the time, when searching through archives is lightning fast and it’s easy to find good content on the site. But to effectively implement ratings..we need to have a intuitive UI. People should be willing to vote if they found the message helpful and without a good UI, I don’t think this is possible.


As, Mailman3 is under active development trying to fix some of the flaws Mailman2 had, HyperKitty is trying to fix major problems listed above.

Personas Connect : MIT100K YouPitch entry

Hi!

Have you realized that your friend circle is shrinking – friends getting married? Are you alone in the city? You don’t have enough friends to hang out with. We are here to help!

Personas Connect is a platform to bring people together at awesome events with the objective of meeting like minded people.  We help you grow your friend circle by organizing events like salsa, games, BBQ sessions and much more…

Liked the idea?  Help us win MIT100K YouPitch Competition by liking our Facebook Post & Youtube Pitch.

Feel free to mail us with any questions or feedback. If you just want to say hi, that’s cool too.



Tarun & Aamir
IIT Roorkee

HyperKitty - New RESTish URL structure

I’m looking for a reasonable way to represent HyperKitty URLs. The basic model is a mailing list.

In my opinion the new URL scheme should look like:

/mlist_fqdn/summary instead of /list/mlist_fqdn/ to display the list summary. 

/mlist_fqdn/2012/03 instead of /archives/mlist_fqdn/2012/03 to display mail archives for march 2012. For example, mail-archive uses http://www.mail-archive.com/firmware-testing-team@lists.launchpad.net/ to display the archives of list. They don’t support monthly view for archives.

/thread/devel@fp.o/threadId/action/like to mark as favorite discussion thread.

/thread/devel@fp.o/threadId/action/dislike to remove from favorite discussions.

/thread/devel@fp.o/threadId/action/addtag/newtag to add a tag to the thread.

I’m not sure about the current URL structure for a particular message. I’ll include the message action (like, dislike) URLs in next update.

For the search URLs,there is a confusion whether search parameters should be included in path info or query string. An advantage of including search parameters in path info is caching while query strings are standard and widely understood.

5 days challenge

I realized that I spend most of my time online going through random news articles and reading about the cool stuff built by other guys. I go through hacker news, quora questions and answers, twitter streams and these are all great source of information but it also takes up a fair amount of your mental cycles. I believe that reading too much of news started to hamper my productivity.

So, I decided to avoid reading the news for the next 5 days. This was motivated by Matt Cutts Ted Talk. Matt said that 

“The next 30 days are going to pass whether you like it or not, so why not think about something you have always wanted to try and give it a shot for the next 30 days?”

I’ll start with smaller duration challenges in the beginning and increase the duration gradually. Since its my first challenge, I’ll go for 5 day challenge.

Let’s see how it goes!

stay connected to IRC 24x7 using quassel

Recently, I was in need to remain connected to IRC all the time, so as not to miss any important discussions going on during my absence. After searching for a while I came across a IRC client named Quassel. Quassel runs on a server and keeps a backlog of all your chats. Quassel is client-core setup. Quassel core is to be installed on machine which has 100% uptime (server) and the client is used on local machine.

You can install the quassel core using (debian based),

# apt-get install quassel-core

I installed it on my Amazon EC2 instance.

Then, install quassel client on your local machine,

# apt-get install quassel-client


Next open up Quassel client on your local machine. Enter the host name/ IP address of the server machine where Quassel core is setup. Connect to core and follow the first time setup of the core.

Are you behind firewall ?

Quassel core runs on port 4242 by default. If you are behind corporate firewall, or firewall of your academic institution, it might be the case that port 4242 is blocked. You can confirm it by issuing a telnet command,

$ telnet ip-address 4242

If this connection times out, there are chances that 4242 port is blocked. You must run your Quassel core on some other port which is opened on your router. In my case I run my quassel core on port 8000 because port 4242 is blocked by my university.