Archive for February, 2010

My Arrow is about to leave the Quiver!

February 26th, 2010 7 comments

In my last post, I made mention of the mistakes I’d made around scope-creep and attempting too much in one go. Those curious can read the previous post ‘My Personal Arrow Project’.

I made a firm resolution to rectify that, and within 18 days, I kept my word.

The beta version of my product (to be announced here in a few days) is hereby complete. My original ambitious goal was to build a recommender system that could inform users which items in a steady information river flowing past them are actually useful (to them). It was a hard problem (to say the least). I have now taken a simple step in that direction.

For those (like me) with tons of useful RSS feeds to follow, but which we cannot possibly completely digest, I’ve create a tool that will automatically archive the blogs that you care about and raise alerts when topics of interest (to you) are discussed. The archive stretches back to when the blogs were initiated, so you never have to worry about missing a post again. Voila, your bookmarks are now so much more useful to you.

Watch this space for more action.

At the heart of this product is sophisticated crawlers, duplicate detectors, data cleaning and classification algorithms. Very complex stuff. At the front-end is a simple interface that a 5 years old could us.
It is designed to make a user’s experiences with their everyday information tools better, reduce anxiety, and enhance value.

It’s been a good two weeks. :-)

– Shahzad

My Personal Arrow Project

February 8th, 2010 1 comment

I’m a self-confessed history buff. I not only try and learn from my own mistakes (which we conveniently label as ‘experience’), but also from the blunders and achievements of others.

For every engineer out there, the ‘Arrow’ case-study should be required reading. If you have not read it, and you have a bit of free time, please follow the link at the end of this article to learn more.

Essentially, if you ask my opinions, this project was canceled due to poor feature management. You cannot have an all-200X technology solution if you plan to launch in 200X. It takes some time to put all the pieces in, and by the time you’re done… well, there are new pieces available, without which you don’t have a 200X solution! Hence, the hamster wheel.. hence feature creep galore!

Mea Culpa. Forgive me, I have sinned.

I know of this lesson intimately. However, that is exactly what I have been doing for the past few months.

I would never have done this on a consultancy project. I’m all about trying to get things as explicit, and agreed-upon, and ‘done’ as possible when consulting. From day 2, my personal philosophy is that there should be something ready to plug in, ready to deliver, so the client is never left lacking when they need the widget, module, or analysis that I am providing. We improve from the baseline that is established as early as possible.

However, since this was a personal project, I felt that it had to be something a cut above. I broke the rules, and enjoyed the luxury of enough slack in the rope to hang myself by.

However, it’s time to move on.

I’ve got a new plan… and guess what, my product should be ready for launch in two weeks. It’s not going to be cutting-edge, all bells’n’whistles with an IQ 10 times that of Einstein… but it will be ready for use today, and will make life better for people who want to track a subsection of the web they care about, and are not served by the free Google Search and Google Alert Services, and cannot afford the specialist tools offered by the social media monitoring firms.


Some comfort here: at least, I am in good company. The engineers in these case studies were all brilliant successful people, who build world-class stuff.