[Hamara-devel] hamara lists, xapian/omega open-source engine library and query parsing.

[shirish]

Hi all,
While I will be most probably putting the FAQ (tomorrow as today's day 
is over anyhow) which would cover terms and conditions for the site, 
this could be further refined. While I would share/tell generic and 
common-sense rules and guidelines, if needed our IP lawyer or anybody 
really could add terms and conditions which protect us (the company) but 
is also fair and mindful of any external contributors.

Another thing which has been bugging me since day 1 is we have no 
search-engine to query emails. This is necessary both for us as well as 
health of any community.

To make that job easier, I propose using either 'Xapian' or 'Omega' for 
indexing and searching of mails.

http://xapian.org/

The feature-list makes it quite worthwhile, especially that it has 
bindings for many languages :-

http://xapian.org/features

The scalability page takes most of the concern out though :-

http://xapian.org/docs/scalability.html

so once we get this at our end, for the next few years we need to do 
nothing,

Even http://xapian.org/users tell a lot about where it is used.

and lastly the queryparser itself

http://xapian.org/docs/queryparser.html

Looking forward to feedback.

Here's the implementation of the search page in Debian.

https://lists.debian.org/search.html

While the search page there is very basic, we could make it much-more 
user-friendly using all the options shared in the queryparser page.

We could use multitude of check boxes, dates, from: and to: user fields 
  with subject to content in email to have much more precise query and 
answers for the same.

This would definitely help with controlling 'information overload'.

I do have some more ideas that would will make perhaps more sense once 
we have some external users/contributors in our mailing lists and we 
need to be responsive and still be able to remain sane but that will be 
later.

-- 
Regards,
Shirish Agarwal,
Community Lead,
Hamaralinux.org