Resetting habit defaults


We know adding good habits is hard. Stopping bad habits is harder.. though Sam Thomas Davies will have you believe it’s not impossible.

Exercise every day? Good.

Smoking? Bad.

Eat vegetables? Good.

Daily drinking? Bad.

Sleep 7-8 hours daily? Good.

What about the vast spectrum of decisions in between? Uber or Ola? Flipkart or Amazon? iPhone or Android? Mac or Windows? Google. Gmail. Chrome Browser. I guess I’m largely talking about technology usage habits and the last three may not even have credible alternatives.

A recent incident at work was the trigger for this line of thinking. A colleague (Ash) had recently started cycling (and running) and innocuously mentioned Google Fit as his choice of app. He got an earful from me on why he should use Strava. “But Google Fit is a perfectly good app..”, he argued.

He didn’t need to say that it was *Google* Fit. Indeed, what could possibly go wrong with a Google product decision? Seemingly nothing yet I proceeded to convince him otherwise. There’s more to tracking your workouts than ‘tracking’, I hectored. There’s *community* also. Anybody who is even half serious is on Strava. I pointed to 2 other colleagues who were already on Strava. That eventually clinched the deal for Ash.

The whole truth is that I haven’t really wanted Google to conquer any more markets and collect any more data than they already have. Owning search, video and Android is plenty. I did not want Google Plus to succeed and I certainly don’t want their deep learning algorithms to tell me that if I change my diet/lifestyle to Plan Qwerty123, it will extend my healthy lifespan by 2.7 years.

Ok Google.

Sorry Google.

You are not getting my fitness data.

I’ve been using the Nike Run app (no social features – hallelujah!) this year and I hope to sunset that app too in the not so distant future.

My wife has observed this about me: I’ll start to do something new (and let me qualify that as ‘good’ new) and massively ratchet it up.

On that ratcheting note, I naturally moved to Could I extract and toss away any more Google tentacles?

Could I switch from Chrome to Firefox? Once a savior for millions of frustrated Internet Explorer users, Firefox struggles for relevance in a Chrome-dominated world. However, their most recent release got me excited. The universe seems to conspiratorially egg me on because in the past 2 months, I’ve had to restart Chrome way too often for my liking.

What about Google Search? Could I stop using it? Giving up Google Search would be harder than Chrome but it adds to the appeal. Years ago, I tried a similar experiment with Bing and it was a middling success. But I didn’t persist.

This time I’m heart’ing on the private search engine DuckDuckGo so I rate my chances higher.

Going beyond the world of Google, are there other habits one can change? Maybe the Uber vs Ola and Amazon vs Flipkart is more of a preference rather than a default, a temporary switch to the ‘other’ driven more by availability/supply side. 

In closing, I’d say three factors seem to be influencing my choice of technology products:

  • underdog status
  • purity of purpose
  • commitment to privacy

Sometimes the post is finished and, despite multiple edits, you are still scratching your head wondering “why on earth did I write this?” This is one such post.


Amazon Webservices or dedicated server?


If your website needs are anything more than a blog (or corporate website), chances are you’d have researched options like “cloud hosting”, “dedicated hosting”, virtualization or the big daddy – Amazon WebServices (aka “AWS”). Lately I’ve been busy converting my 7 year old Macbook into a developer-friendly machine — installing all manners of CLI developer tools and tinkering with Amazon Webservices. Why I’m getting back in touch with my “developer side” is the subject of a separate post.

So anyways.. I’ve been impressed with the quality of forum discussions and finally learning what the fuss is all about Stackoverflow – is there any installation/error code that DOES NOT have a fix on Stackoverflow?

This morning I read an AWS forums posting which succinctly answers the Why AWS question. The thread starts with how Amazon measures an ‘instance hour’.

User Allen writes about the instance hour:

See “Paying for What You Use”, the 2nd to last paragraph at .  That should explain it.  You are charged for every running instance, whether its in use or just sitting idle.  The rounding up happens for each instance, i.e., for each instance, you are charged an integral number of hours (1 hour, 2 hours, 3 hours, etc.), never a fractional number of hours.

User aschildbach is miffed about EC2’s unfair /uncompetitive policy..

Hmmm, this makes EC2 very un-competitve in my eyes. I am currently paying 20 Euros a month for a dedicated machine with roughly an EC2-Unit of prozessor, but only 512 MB RAM. This price is including tax, several terabytes of traffic (more than enough) and some backup space.

With EC2, I would pay 30 days * 24 hours * $0.1 * 1.19 (tax) = $85. And that does not take into account backup space (S3), traffic and true persistent storage (EBS) yet.

Sure, the “Cloud” has it’s advantages (Firewall, private Subnet, easy scaling). But is this really worth 4 times the price?

I would prefer a pricing model that would actually charge me for usage.

At which point, user lenny delivers the checkmate argument – basically the value of AWS over a dedicated server.

Sure, the “Cloud” has it’s advantages (Firewall, private Subnet, easy scaling). But is this really worth 4 times the price?
EC2 does not provide you with a private subnet.  <img src=”” border=”0″ alt=”” />

In order to see why EC2 is special, pay particular attention to the word “Elastic”.  There are a lot of real-world applications whose resource needs vary throughout the day/month/year (nightly processing/monthly processing/Christmas season, respectively).

Does your hosting provider let you spin up a server every night at midnight for batch processing and then shut it down when the processing is done?  If they do, do they charge more than $0.10 for the privilege?

Does your hosting provider let you add servers and remove servers from your clusters dynamically and programmatically based on current load,however you decide to measure load?  If they do, do they charge you by the hour?  Probably not.

Does your hosting provider let you spin up a testing environment for a day of testing before going live, and only charge you for the day’s usage?

Does your hosting provider let you conjure up a 400 CPU server farm with 140GB of RAM in 5 minutes and for $16.00/hour?  I’m thinking probably not.

If you’re honest with yourself about what you’re getting for your 20Euros/mo, what do you think would happen to your application in the event of a software or hardware failure?  Do you have remote reboot to recover from a kernel panic?  Do you have access to console output? How long would those take?  What would happen in case of a hard disk failure?  How much downtime would that cause in your 20 Euro/mo plan? More than 5 minutes?  What if an unskilled backhoe operator accidentally severed your provider’s internet connection?  Could you immediately launch your application in a different datacenter, or would your downtime be measured in days?

If your computing needs are static and downtime in your application is acceptable, Amazon’s Elastic Compute Cloud is probably not the lowest-cost tool for the job.  Indeed, I have a family blog and photo album… do you think host it on EC2 or on a $5.00/mo webhost?  Sure, I experience downtime, and my site’s performance is abysmal, but that is acceptable to me.  If your customers can’t get to your e-commerce site to spend money and they go to your competitors instead, is that acceptable to you?



The Fifth Elephant coming to Bangalore.. bearing Big Data Goodies



Pic: courtesy HasGeek

Q: What is The Fifth Elephant?

A: It’s  a community powered conference on the Big Data ecosystem in India, organized by HasGeek.

Q: What does ‘community powered’ mean?

A: For all the conferences that HasGeek organizes, they use a crowd-sourced speaker-selection funnel where potential attendees vote up or down which talks they want to listen at the upcoming conference. The final program list (across 3 tracks – Infrastructure, Analytics, and Visualization) can be viewed here.

Q: Errr… what is Big Data?

A: No shame in asking that question.. especially if you’ve not been watching the tech ecosystem for a while. A decade ago, I turned to Tim O’Reilly’s seminal article (Web 2.0 Design Patterns) to understand all the nuances of Web 2.0 – I wasn’t disappointed. So I shall point you to this O’Reilly article What is big data.

Q: Where is it being held? Can I still register?

A: It’s a two-day conference (Jul 27-28) being held at the NIMHANS convention center.You can still register here but be quick!

A detailed press release from HasGeek below….

Big data and analytics are buzzwords. People are aware that something called big data exists but do not know how it applies to their lives, businesses and futures. Mystery also surrounds the very nature of the different kinds of data that are being collated and processed. Yet, data is on everyone’s minds and lips … … …

The Fifth Elephant is a community-powered conference organized by HasGeek to demystify the big data ecosystem in India. It brings together speakers who have worked extensively with technology, statistical and mathematical tools, and user interfaces for mining, analyzing and representing data. The conference attempts to showcase the technology landscape which today makes it possible to work with data and leverage on its predictive powers. On the first day of the conference, participants will be able to get an understanding of the technologies available for managing big data, analytics and visualization. On the second day, we will showcase talks that explain how technology is being applied in retail, medicine, IT, finance and governance to manage and analyze data.

The Fifth Elephant has been organized for two explicit purposes: first, to understand the state of society and economy in the context of pervasiveness of data and technology. Second, the conference intends to provide a platform for data enthusiasts, developers, scientists, analysts, entrepreneurs, firms developing data products and services, platform providers, CTOs, CIOs, researchers and journalists to interact with each other and work towards building a community around big data in India.

The Fifth Elephant features speakers from Red Hat, Splunk, Google, Strand Life Sciences, IBM, Flipkart, HP Analytics, Mu-Sigma, Amazon, PayPal, MediaNama, among others. Keynote talks will cover topics such as how Google crunches big data at Google scale, community architecture and big data at Red Hat, the genomics project and challenges of big data in personalized medicine, and the process of building Watson and natural language processing.

The conference also features talks on building search engines, messaging architectures, use of Hadoop in mining and managing data collected under the Aadhar/UID project, data explosion in medical imaging, open data in life sciences, financial market and economic data, to name a few.