Vivek Ravisankar's Web » Uncategorized

Estimating Programming Tasks

bytingme — Mon, 30 Aug 2010 18:37:17 +0000

Sleeping under the chair, working long hours are almost common traits of a programmer – it has become so common that you wouldn’t be accepted into the geek community if you haven’t done the above. But I asked the geek I am close to, on whether it’s actually essential and found an interesting answer/method that I have begun to follow. The credit to the following points goes to him who is lazy to write a blog post.

So, how do you estimate programming tasks at work ? A simple answer – DON’T ESTIMATE .

Lets say you have a task at hand – Build a codechecker which would evaluate a code snippet against a test case and return the result.

Phase 1: Study

Break down the task into atomic sub-tasks. There is absolutely no coding involved here. It’s simply breaking down tasks, viz. designing table structure for code submissions, desired prototypes of the function, communication should be done over XML/SOAP/JSON,etc.
Get down to the research mode to identify which protocol/database-type (couch/mongo/MySQL) would be ideal for your component
Just as a website/page can’t be designed without a wire-frame, similarly the structure or a pseudo-code would go a long way in clearing out the blockers you might face later. Get the structure and pseudo-code ready.

You still can’t tell your manager how long it would take for you to complete the task, you just completed your research .

Phase 2: Code

Now that your pseudo-code is ready and you have done enough research, get down to programming 1 sub-task in the list. Lets say it takes x days. As said, the best way to estimate the task is by doing it!

By definition of our sub-task, each one is atomic and is indivisible. Hence the total time that would be required to complete the task = n * x where n is the number of sub-tasks present.

That’s just the 70% of the total time required ,T days (i.e) 0.7 * T = n * x. The remaining 30% goes for bug-fixing, “unexpected errors”, etc. And hence, you arrive at the total time required to complete the task.

In workplaces you would be required to take ownership of an entire component and this helps a lot in arriving at a good estimate of the time required. Whatever role I play, I think programming would definitely be a part of me and happy to have learnt a way to estimate them. Let me know your thoughts/feedback and how it worked for you.

Image src: http://farm3.static.flickr.com/2416/2522802875_f095c0a40d_m.jpg

5 things I do before a demo

bytingme — Thu, 12 Aug 2010 13:18:13 +0000

The beautiful thing about starting a startup is you would love almost anything that you do – sales/marketing/programming/PR/blogging, the feeling of ownership is unmatched. Our product is in it’s stabilized version 1 form and I do a lot of marketing & sales these days – on an average about 1-2 demos/day.

Few things I learnt before demo-ing your product to a client. These are the things I do before I demo and has proved helpful -this is not to be treated like a TODO list,etc. – after all every startupper has his own way of getting a customer.

Research
With the advent of so many social networks, it’s possible to get almost any activity/data of an organization/person. Three things I research on
| The company – read about it’s history, interesting trivia, past techcrunch articles, etc. This gives an idea of what the company is about and goes a long way in understanding their needs and making appropriate customizations to your demo
| Founders – Most likely, it’s going to be an interesting/inspiring story on how they started. If you get an opportunity to meet the founders, nothing like it and it’s extremely obvious that you need to do your homework about them, else it serves as a good read
| Person you are going to meet – Not always would you get a chance to meet the founders, it’s good to know the background, past work experience of the person you meet – Linkedin, twitter, his blogs importantly would help a lot in understanding him.
Getting ready
“Never waste a single second in the meeting” – this has always been my thought before a demo. You hardly get 15-20 minutes with a senior person, try to make the maximum use of it. Before I demo my product,

| Hibernate mode – Booting your laptop before the demo is such a waste of time, you are losing 20s if you are using Ubuntu and anywhere between 2000-3000s if you are running Windows. There shouldn’t be an idle time, put it in hibernate mode when you start.

| Keep your browser/app ready – I have a separate firefox profile for each demo containing 3 tabs – the product, the report & a view of codechecker. Remember you only have 15 minutes and need to explain the pain point and demo the product. This quickens the process if you are already ready.

| Customization – People love it when you make your product/service customized to them. Before each demo, I look at the careers page, check the openings and prepare the questions of the test accordingly.
Why it helps? – the company can start using my product for that role instantly. Only rarely is anyone going to pay at the first attempt, first make him use the product and get him hooked to it.
Punctuality (Too early is bad)
Being punctual is good, but being too early might go against you. When you are going to meet the decision maker, his calendar is going to be full! He/She would have given you a 30 minute slot in the middle of other meetings.

Even if I am a little early (say 20 minutes prior), I tell the receptionist to inform the guy only 5 minutes before the scheduled meeting time – Reason? 90% of the time, he/she is going to ask you to wait in the lounge till it’s time, but sometimes your presence acts as an interruption to their current meeting/work. If this happens, they would like to finish off your meeting quickly and get back to their work – you can be rest assured that he/she isn’t going to listen to your pitch/demo, but just wants to oblige the meeting request. This happened to me once and yes the chances of him getting converted to a customer became slim.
Have a goal – shouldn’t be a coffee meeting
It helps to have some kind of a goal/outcome of the demo – maybe the company should use the product for free, run a pilot program with them for a month, etc. This helps your demo to move in the right way and proceed towards something concrete instead of just getting to know him/her.
Being fresh
This might be clichéd, dressed neat, pressed formals, appear fresh, etc. But as startups, we really can’t afford the luxury of an A/C car daily – even if it’s 30 KM from my place, I still travel by bus, standing in the middle of 50 people. It makes you a little tired and weary before you demo which is obviously a bad sign.
To avoid this and be economical, I travel by bus, get down a couple of stops before the destination and take an auto for 20 bucks. The 10-15 minute drive in auto with a quick deo spray makes you look much better than traveling by bus fully – economical and best.

All the above are small (maybe obvious) stuff but happens in the 1st 5 minutes of the demo which is a very important phase! A lot of my posts mostly originate from conversations, learnings, experiences – I am no expert, but the reason I blog is, if the points make sense it would accelerate the growth of a startup by a good amount which is critical.You have zero money, little resources, no time but still need to beat the biggies – acceleration is key.

Who is your mentor?

bytingme — Tue, 27 Jul 2010 17:16:21 +0000

Who is your mentor? – I honestly didn’t think this question would trigger a blog post. I met this guy X, a couple of days back who is a wannabe entrepreneur (I hate this term, but that’s how he introduced himself). As we were chatting about his idea, he interrupted, asking who were the guys mentoring me and their qualification (NOTE: He was interested in their qualification more than their background).

He was shocked to know that none of them were either an MBA/PhD/Masters/any-other-extra-degree. And he proudly claimed that his mentor was a VP in a top MNC with an MBA from ISB, Hyderabad (with no previous startup experience).

I have been mentored by a group of serial entrepreneurs for almost a year and this led me to tell something about what I feel mentoring to a startup should be about.

Aligning yourself in startup framework
To me, the biggest value add a mentor should bring to the table is to help place the founders in a startup framework. A startup life is tough, very tough,especially in the early stages. Your business model can even take a 180 degree turn, to understand this and provide clean advice is the biggest value add!

Only if your mentor has started a couple of startups, will he/she be able to place you in this kind of a framework since the journey is emotionally tough.

I personally think it would be very tough for the VP to associate with any of the above. He might have never been in a position where there were resource constraints (or) his product being rejected by other companies, after all his MNC is such a big name. And even if that’s the case, he would have pulled up a team of engineers to build one more. Where’s the money in a startup?!
A mentor isn’t going to run your business
He was overjoyed to tell me how his “mentor” frames the B-Plans, the financial charts and even talks to people and converts them to customers.

Firstly, I am not a great believer of writing B-Plans without having a product with a few paying customers – only if you have paying customers, does a B-Plan actually make sense. That aside,a mentor shouldn’t really run your business, he should definitely not be your boss but more like a co-founder to juggle ideas, brainstorm and take inputs.

The basic tenet of an entrepreneur is expecting freedom and if that’s curbed, the startup gets a hit.
The VP has a huge network
Having a big connected network is definitely a huge add-on, but that’s not the only reason why you choose someone as your mentor. It’s good if your mentor is connected with a lot of people – helps in getting introductions, etc. But choosing someone because he has a wide network with top VP’s and isn’t a sufficient condition.

Introductions/networking isn’t costly – most of them would be glad to introduce you to people whom they know closely provided you have a good enough reason, latch onto a couple of big guys, they would definitely do the intros (talking out of experience).

I tried telling him this, but beyond a point I stop talking to people who don’t understand the bigger picture and thankfully because of that I was able to write a blog post.

Bottomline being if he/she hasn’t founded a couple of startups (doesn’t matter even if it’s a failure as long as they hung on to it), I would not bother bringing him/her on the mentoring panel however impressive his qualifications are.

Image courtesy: http://www.flickr.com/photos/chrisandrin/3056695077/

6 obvious things I learnt the hard way

bytingme — Sun, 27 Jun 2010 05:37:59 +0000

I have always cribbed about the fact that a lot of entrepreneurs in India don’t blog much. Only a handful of them actually share their experiences, struggle, etc – a look into your google reader would tell you the number of indi-startup feeds (vs) valley-startup feeds. To contribute a little towards the Indian startup ecosystem, here’s an attempt to share what I have learnt in the startup journey till now.

We started off 10 months back with a portal that would enable candidates to attend mock interviews [skype/telephone] from industry professionals. Some of the things I learnt about online businesses focusing on the Indian market and more so with the student population.

PS: When I mean people, it could be interchangeably used to refer to general indian market (or) student crowd since that’s the segment I interacted with the most.

1. Paying for a service

We have been used to paying for a book, dress, movie ticket that suddenly paying for an online service becomes something totally unheard of. The notion/mind-set amongst people is ‘Anything that’s online is free of cost‘.

People were ready to buy a book containing a list of interview questions & answers priced at Rs.400 rather than for an online personalized 1-1 mock interview priced at Rs.350 since they could hold onto something tangible for the money invested.

Very very very crudely speaking, they paid, for talking to someone over the phone! – unfortunately the value behind this remains in the dark.

2. Offline Model

However web-savvy you are, you need to do a good amount of offline marketing to increase the reach. This is especially true if your product could be consumed offline.For eg: to attend a mock interview on our portal, all you needed to have was a telephone. 70% of interviews that happened were through an offline medium.

| I used to stick posters in colleges saying “If you are interested in mock interview, SMS your dream job to 98408….“. This poster simplified the process to a great extent – the moment you saw the poster, all you needed to do was just send an SMS. A good share of interviews came via this route.

| We had a wonderful set of campus ambassadors who would evangelize our product in their colleges – we got quite a lot of interviews via this channel

| We put up stalls in college fests and students made on-spot payments for interviews – again an offline route.

| We conducted a lot of workshops/ppts in different colleges to market our brand.They loved our workshops and wanted us to conduct more, but rarely did it convert to getting people apply to mock interviews through the online channel.

As much of inbound marketing, SEO techniques you would be implementing, to me, it’s not wise to ignore the offline channel – people still need an SMS confirmation, call on their mobiles, etc

3. Online indulgence

Lets talk about the majority here. 80% of the time you spend online apart from your normal work goes off on this, this & this. Either the problem you are solving has to be the topmost and you need to fare well on the search results (or) you need to hop onto one of the existing channels. Like every early stage startup, it’s wise to do the latter.

We decided to code apps for orkut, GTalk bot wherein you could get random interview questions, puzzles, etc.We missed the basic rule in all these things – it needs to be FUN, it’s okay even if it’s funny, but it can’t be serious.

The golden rule of orkut apps is – at the end, you definitely needed to slap someone, kiss a friend, kick someone’s butt, else it fails to qualify the criteria to get published. So, if you need to get onto those channels to spread the word, fortunately/unfortunately you would need to abide by the above rules.

4. No trial version eh?

This is something that everyone of us inherently are used to – but I realized it only after getting this far, better late than never! Be it buying a t-shirt, book, audio CD, you would always want to have a trial of the *same* product before buying it – trying out the same tee, reading the first 20 pages of the book, watching the trailer of the movie before going, etc. Similar was our case – people wanted a trial of the mock interviews.

We built a few by-products (free obviously) which we thought would get us some conversions to the main product – the logic was spot-on but the numbers were no way related to our expectations. We did give off free interviews which pulled in a lot of students to attend, but there were 2 problems

| We set it up for 30 mins (since it was a trial) and the full effect wasn’t visible in that duration

| We couldn’t have a trial like this all through, since the same student could take this 30 minute free-interview again – we needed to compensate the interviewer for every interview.

Ofcourse, the concept of trying out something before I buy has been prevalent all through, just that I realized the importance only now.

5. ‘Ssshh.. I can’t tell my friends about this website!‘

This one should be the best! This was from a student from a reputed college who took up a mock interview. He was very happy with our service and I thought he would atleast put it up as his GTalk status message if not blog about our service. I was wrong!

He said (verbatim) ‘I need a competitive edge against my friends for tomorrow’s interview and I shall maintain this website as a secret only to me’ OMG! :O I really didn’t know whether to laugh/feel-upset. But, bottom-line, I couldn’t reach out to that college through him.

6. E-mail campaigns don’t work

The information overload is so much these days, that at times even personal e-mails get the stick! You are out of connectivity for 3 days, your GMail/Twitter is spammed! In such a scenario, spending on e-mail marketing is such a waste!

As most of us would do, an e-mail from an unknown source/website is archived without even reading the contents of it. Blog posts dealing with ‘10 Tips to better e-mail marketing‘, etc are absolute nonsense. Those things work only if someone opens the mail, but that itself is a rarity.

However reliable services you use for e-mail marketing to send thousands/lacs of e-mail, there is a *good* chance that your domain would be blacklisted. Now, that’s a very dangerous thing to happen – and it did happen to us, we couldn’t send any e-mail from “interviewstreet.com” for 2 days. We used a “reliable” service for the marketing experiment but still got hit. If at all you have money, you could throw to Google Adwords and get some traffic.

The above are some of my learnings. I am not criticizing/offending anyone, but it’s more like a realization on how the market is and how I should position my product. The reason I wrote this post is in the hope that some guy who is planning to start-off (or) is in the early stages could change his positioning a bit if the points makes sense. Thank you.

GCJ to Sun Java

bytingme — Sun, 06 Jun 2010 11:35:33 +0000

It’s always a problem when you have more than one version of a particular product/package and more so if you have both of them co-existing in your system. I was playing around with Selenium yesterday when I had to execute the selenium server jar. It repeatedly kept giving errors and crashing whenever I tried executing the Selenium tests.

Like how a naive person would think, I thought executing “java -jar selenium.jar” would always invoke sun’s version of java since the open-source version invocation to execute is generally
“gij classname ” and “gcj class-name.java” for compilation.

Apparently, it isn’t so! And it was using Open-source’s version of java to execute the jar because of which so many exceptions were thrown! I always hated setting classpaths, modifying profile information, etc. – most of the cases it has never worked for me! Googling a little got me to a solution where you could easily switch from gcj to sun-java and vice-versa.

sudo update-alternatives – - config java

That’s it! All you need to do is just choose which version you want and it uses that as a default value. Isn’t this simply amazing instead of some nasty classpath changes, exporting SHELL variables, etc..! And yes, if you are a web-developer, I suggest you checkout selenium – amazing tool!

‘Gift’ed friends

bytingme — Wed, 19 May 2010 06:06:51 +0000

The last 3 months has been yet another roller-coaster ride with only a few things looking positive, results of which I would probably know in a couple of weeks. In the moments of lowness,a.r.rahman’s compositions come to the rescue, but a couple of interesting events happened recently which also put me on a high…

I have always believed that if you need something really really badly, sooner or later, you would get it. With my cellphone picked at Bangalore railway station and my iPod losing it’s life, there was no way I could listen to those rahman’s tunes on my travel – and I used to travel a *lot*!

Just as it would sound like a tamil movie’s script, a month later, I got this beautiful mp3 player from Arun along with Inbound Marketing book from Balaji – a book that is a *must-read* for anyone starting up, which is not available in India! It came to be as a biggg surprise and was really touched given the fact that both the guys didn’t have a full-time job, but still decided to get something! I almost never get out without the MP3 player in my pocket, my mobile comes only next.

The second incident was even more dramatic! I was just flipping through the initial pages of the book ‘Viral Loop‘ at Odyssey. The first 10 pages just caught me on and I just didn’t want to leave without the book – a battle between mind [which always thinks about money] & conscience ensued and I came out not buying it. The 1st 10 pages dealt with an interesting story of hotornot, and I tweeted saying “Checkout hotornot.com” [PS: There is no reference to the book in the tweet]

In what would NOT even happen in Tamil cinema, that very night Rashmi shipped the book to my place!! Would you believe it? She had no clue whether I read the book (or) if I even knew that such a book existed! I am still yet to come out of the surprise/shock, there is a height to co-incidence also right? I so wanted the book in the afternoon and 2 days later I am reading it. Thank you soooo much!

Whenever I talk to someone about some of my genuine friends, these guys have always been there in the list – no, I am not saying because they got me a gift , honestly I have mentioned their names everytime I talk about genuine buddies [you can even ask my mom!]. The beautiful thing about these events was that there was no special occasion associated with the gift, not my birthday, not a new year, absolutely nothing at all – and to still remember me and send something is absolutely wonderful!

For now [and forever] my love and linux troubleshooting is always there for you guys . With respect to gifts, just wait for a little more time, I will be making money soon . Thank you soo much!

Lessons from Dropbox

bytingme — Wed, 05 May 2010 18:12:28 +0000

I am a big fan of Drew Houston, founder of dropbox and here he (along with Xobni) tells their startup story. Very insightful and must-read for product/technology based startups!

From Zero to a Million Users – Dropbox and Xobni lessons learned

View more presentations from Adam Smith.

I mailed Drew on whether his presentation was applicable only to US startups [obviously not expecting a reply back, since I am used to people not getting back to me ] – but thoroughly, totally surprised, he got back to me!! saying….

And definitely needs to be agreed! It’s a great feeling when someone whom you regard as a source of inspiration takes some time off [even a couple of minutes] to get back to you!

Signup

bytingme — Sat, 01 May 2010 10:30:39 +0000

The importance of a sign-up (or) login page is often underestimated. A recent survey indicates that having the text ‘Sign up‘ instead of Register increases the conversion by over 20% ! – isn’t that amazing?

The old ways of signing-up are slowly getting deprecated, users want to try the product before they sign-up. Instead of Sign-up -> Try the product, users want to follow ‘Try the product -> like it? -> signup‘
I already have to remember so many e-mail/password combinations, don’t want one more! – this might also be an apprehension amongst a lot of users. Fear not, for the introduction of openId, Facebook/twitter connect, etc. has made life simpler to so many people – the user can use the same credentials to login. Infact the popularity of these services is rising so much, that a few companies are making money by providing a way just to plug this to your site
Some websites totally eliminate the concept of sign-up! Kudos – collect user data only if you need it. For eg: redbus doesn’t have a register/signup form – you possibly can’t do anything but book tickets
Ok, you are hell bent on having a sign-up form – please don’t ask for my birthdate, gender (unless you are a dating site ofcourse!), my sister’s name, etc. Keep it simple with just 3 things – e-mail, password, confirm password
Geeky world: Most of the designers/coders live in a geeky world! When the form contains e-mail, password – a lot of users think you want the password of their e-mail address! It’s quirky, yes! but come out of the geeky world, a lot of them are like that! You could consider renaming it as New password
Activation e-mail: This is a common practice among a lot of websites where they send an e-mail to confirm whether you are really *you*. The user needs to come back to his mail account, click on that long lousy activation link, and then login again! (few websites provide your *password* also in the activation mail – atrocious!) This is an unnecessary hop – what you could do, is to set a cookie for 24 hours and log him in for that session.

The bottom-line being to make users use your product with minimum hassles. If your product sucks, it’s a different problem, but atleast let people use it Like most of my earlier posts, I have made all the mistakes above, but wait till the end of this month – we are reincarnating

Invest in Technology – Start(i)up

bytingme — Tue, 13 Apr 2010 19:55:06 +0000

Being frugal is essential in a startup life, but there is a thin line between being cheap (not spending at all) and being frugal. The thought of a startupper is always to get the best by spending the least. Some of the things I felt startups shouldn’t mind spending on …

Designers are needed, esp if you are a geek:
Hope all geeks dont take me wrong on this, but most of the geeks suck at aesthetic sense and designs. Given a mock-up, there can be none better than them to get it live on the site, but designing a mock-up requires a different (I would even say complementary) skill. If you are running a e-commerce business where most of your transactions and business is going to be online, better get someone to design the pages – even the sign-up button, it’s position, size, everything matters!
Even though, I don’t use wufoo much, I simply love their site, their colors, the dotted underline when you hover over a button, everything is so cutely made that I was forced to use their product!
Interns are costly – hire them
There are tens, no hundreds of internship sites where you could post your description and some of them say ‘Rs. 1 per hour‘ – I understand you are a startup, but come on … If you really need to get your job done perfectly, you need to shell out a little. The ultimate goal should be to get the feature done (or) accomplish the sales target you have set – don’t compromise on quality
Hosting space – get a dedicated one
We initially hosted interviewstreet.com in MediaTemple (shared hosting space) and faced so many issues! Yes, it was cheap, much cheaper than others existing in the market, but not having a dedicated one implied you don’t have root access (you don’t have root access on your own startup?!), limit on memory limit to run django process and many more.
We wasted 2-3 weeks of solid development effort trying to tackle all problems posed by a shared server and finally unable to take it anymore, moved on to linode
It’s so much peaceful when your local development machine and the remote server are of the same configuration & OS – no nasty hacks on the server to get your app running.
Accessories – laptops, netbooks, keyboards, etc
Again, I am sure there would be a team (or 1 person) whose job role is to code. Invest in a good laptop (not netbook!), a comfortable keyboard, good internet connection – never be frugal/cheap in these. It’s extremely frustrating to work on a bad keyboard with slow internet speeds – this frustration can make the developer stop coding!
Stay Connected
Customer service is a key element in a startup – replying to e-mails spontaneously brings so much joy in a customer. A data card (or) a phone with GPRS facility to keep a tab on customer activities and getting back to them tells a lot about your startup. Invest in them

There are so many other ways of saving money (work-from-home), economical travelling, etc – when you are running an e-commerce business, don’t hesitate much to spend on technology if it’s going to improve the productivity of the overall company. To be honest, I didn’t do anything of what I stated above initially , but slowly realized and have almost done everything.

MBBS & Donatello

bytingme — Sat, 03 Apr 2010 21:49:18 +0000

Thank you so much for the Bday wishes! Really overwhelmed I share my birthday along with Aditya, my classmate at NIT, Trichy. Not many of you know about our similarities, so here’s an attempt like John.F.Kennedy & Lincoln

Both of them were born on April 4th 1987 in Bangalore in hospitals which were just 2 km away
Both of them did their schooling (major part) in Chennai
Both of them joined Computer Science at NIT, Trichy
Both of them secured a rank in 100’s in AIEEE and just landed in Computer Science
Both of them were in the same section (‘B‘) in the 1st year
Both had the weirdest nicks in the geeky forum – vivmbbs, donatello which had no relation to their name unlike myprasanna, nkarthiks, etc.
Both of them loved Linux and were in a way Linux evangelists, though one definitely knew more Linux than the other
Being GLUG-T members and FOSS evangelists, both shamelessly appeared for Microsoft placements!
Finally both of them got an offer from Amazon.com, one took it up and one didn’t

Now, it’s diverged a bit, let’s see if we meet sometime! Ok, actress Simran also is celebrating her birthday today – more about the similarities with her later

Thank you so much! It was absolutely great talking to you all after a long time that I wished it was 4/4 everyday! My phone was ringing all night, it might have been in waiting for a few people, I tried to call everyone back, if I have missed someone, am sorry, please let me know (or) you can call