Thejo Kote

As part of the IOLab course, we worked on a bunch of mini-projects related to themes like the semantic web, controlled vocabularies, distributed classification etc. We spent around 15-20 hours on each project. The Flickr Tagr, which Ljuba and I worked on, was one of the projects that was in a complete enough form. Here’s how we described it:

Flickr Tagr visualizes photographs based on social classifications (i.e. tags). It allows users to search for photographs tagged with particular words while showing them additional tags related to the ones selected so far. The goal is to allow a user to quickly visualize the photos available for an arbitrarily selected intersection of tags, while exposing them to other tags they might find interesting. Users can also click on a photograph to see a larger version in a custom-built light box.

We thought the faceted search interface to Flickr was fun to play around with!

Just Samachar had been pretty much abandoned since February when it was created. I had prepared a to-do list back then, and have finally been able to get to it during the last two weeks. So, a bunch of new features are available now.

Different Views
Apart from the normal view, where each category gets its own page with news from different sources, two other views are available –

• Tag View – A recurring problem I’ve faced is that, sometimes there just isn’t enough time to go through all the categories and the news from each source. In such cases I glanced through the ‘In the News‘ tagcloud and learnt what was in the news. But, that still required a page load for every category. The tag view is an attempt to fix that. It shows the news from all the categories on one page as a normal tag cloud (font size and colour decide how ‘hot’ the news is). Click on a tag which interests you and you get the most relevant headlines right there. The Week and Month tabs show what was in the news in those time periods. In short, the aim was to get a lot of news quickly on one page. Of course, it uses simple javascript AJAX.
• River of News – Not everyone preferred the normal view with news from all the sources in boxes on one page (Hi Amit!). So, the river of news view is also available now. If you’re not familiar with that phrase, the simplest way to describe it would be, news with an emphsis on time and not the source. Dave Winer, who coined the phrase describes it here. Though it originated with reference to RSS readers, everyone and his uncle has hijacked it now, including me.
  Actually, I created this view because it gave me the opportunity to learn some cool stuff I didn’t know anything about earlier (yeah, I’m a little selfish that way). This view also provides a ‘Top stories’ section, which displays obviously, the top stories, but with other relevant headlines clustered together. This whole attempt to identify relevant stories led me to try out some cosine similarity and term weight methods and such voodoo. Finally, after much head scratching and reading, I looked more closely at the Fulltext indexing feature of MySQL and realised that it does more or less what I want and definitely more efficiently. Becoming familiar with the mathematics behind Fulltext indexing and the scaling issues involved will definitely be useful. There are other similar interesting things that can be done around this river of news view. Hopefully sometime soon.
• Default View – Since different views are available, you can now set your preferred view as the default and have that view loaded every time you visit the site. Cookies need to be enabled for this to work.

Stats
I’ve been collecting a bunch of data since February and this section is the first step in sharing that. The top news topics are available by month and by category. Really interesting stuff if you ask me. If you’re a stats junkie like me and would like to use the data for anything cool you can think of, drop me a line.

Local News
Long time demand by users. To the people who asked, thanks for waiting patiently. I’ve added the states for which I could find RSS feeds. If you know of others, please let me know.

Search
A proper search function is now available. You can search in specific categories and time periods too. Wherever applicable, related search terms are also displayed to help determine the context in which the search term was in the news in the chosen time period. Boolean search as provided by MySQL is supported. More information on how to use boolean search is available here.

Archives
One month of archives is now available for all the news sources Just Samachar tracks. In the normal view, if you want to read more items from a particular source, just click on the icon at the top of the box. Source names are linked to the archives everywhere, so it can be easily accessed that way too.

Popular News
Earlier, only the day’s popular stories were displayed, now the popular stories for the week and month are also accessible instantly. Popular stories are based on what the visitors to Just Samachar are reading.

That’s about it for now. I started out with an aim to reduce the size of the to-do list, and even though I’ve drawn a line through a bunch of things, in the process of working on them, many more have been added. I want to learn more about character encodings, unicode, multi-lingual support and many other things which are a complete world unto themselves. Things have started getting hectic at work again, so I will get to them eventually though.
Another thing. If you’re reading this in Internet Explorer, please switch to Firefox. I’m sure thousands of developers will be very thankful.

As always, feedback and suggestions are welcome.

There’s a saying which goes something like, “If you want something done well, do it yourself”. The engineer’s version of that would be, “If you’re not happy with something, build it yourself”. Well, I’ve tried.