Lessons Learnt

Mom's Guide: Privacy on Facebook

2012-01-17T06:11:00.000-08:00

Mom: Hey dear! Are you up on Facebook?
Son: Yes Mom! Thanks to the idea of faking year of birth; In Facebook I am 20 years old and it allowed me in. Look here I have shared a message as well.

Mom: When you share a status message, photo, link or a video (collectively known as posts), have you wondered about visibility of the post i.e. who can see that post?
Son: Obviously. Facebook is about friends so my posts are also visible to my friends only.

Mom: Are you sure? Have you changed the settings?
Son: What settings?
Mom: Then you just check your posts again. As I know, if you are new to Facebook and you have never changed your settings, your posts are public i.e. visible to everyone in the world even without logging in to Facebook.

Son: Oops.... but I want to share only with my friends. Isn't there a way? You were talking about some settings.
Mom: Yep! whenever you create a post, you have to pay attention to bottom-write area of the publisher. Right previous to the post button, privacy setting tool for the post is provided. You must choose who all you want to show this post.

Son: Oh great! But wait .... do I have to do this for every post? That will be too much. I may forget about it some time.
Mom: Oh No my boy .... you can setup the default privacy for all future posts from the privacy settings page.

Son: Yess.... now its makes sense. And the custom option is also available. So I can choose to share with a list or specific people and also can hide my post from specific people. This is really fantastic.
Mom: Yes the control is there but never hide anything from me :) and you know if you are on a custom list, the default privacy for the post is automatically set to that list only. So remember to add me on all your lists.

Son: But Mom I can take care of privacy for my new posts. What about the old posts? They will remain visible to everyone.
Mom: Nop!!! Go to your wall. You get the options to set the privacy of any post that is already submitted.

And again on privacy settings page, you can change the visibility of all previous posts in one shot.

Son: That's very good. Let me do it now.
Mom: Okay dear .... Enjoy safe posting on Facebook.

Kids' Night Out

2012-01-09T09:10:00.000-08:00

Here I am talking about kids of 5-10 years of age.

We as a perant alwase try to keep our kids happy and far from any hardship. We, the parents, always take care that they, the kids, get everything that thay need or may be more than that they need. We do not want them to strugle for anything, at least at this age.

However, we forget that at some time thay have to go away from us and face the world independently. We forget that as parents, it is our responsibility to train them so that they can face and concoqer any situations of their life. We forget that WE need to make them independent. It has been established again and again that excessively pampered kids fail to complete challanges of life at grown up age and the life rags them throughout their lifetime.

Typically, in the age of 5+, a kid should be able to work for himself. Means the kids should eat, drink, dress-up etc on their own. But, generally, parents do all this for the kid even if the kid knows how to do it. So I think, we should feed them small training capsules in a fun way so that they get ready to become independent.

While talking to a few friends, I felt that most scary and unimaginable thing for them is the kid's night out alone. However, if you think, the kid is going to sleep most of the time during the night out as oppose to a usual day-care where most of the time the kid remains awake. So night out should be simpler than the day out and the kid gets to learn more things for the life e.g. taking care of the belongings and getting ready themselves.

I also found that a few people understand the importance and send there kid for a night-out to a friend's home. I found a few summer camps taking the kids for night out but they are typically at a camping location far from the city. Many mothers expressed concern in sending the kid to a far off place as they can not quickly reach to the kid in case of any hypothetical emergency.

So I thought of the idea that similar to the day care service, there can be a night-care service. However, the reason/rashnale for sending a kid to night-care will be completely different from that of day-care. Of course, sending a kid for night-out is going to be a 'once in a while' activity so service will also not be available all the time. Typically, these night outs should mostly happen on weekends/holidays so that kids do not have a tight schedule lined up for next day. The span of time of a night out activity shuld be such that that kids have enough time for some learning activity; for example, how about 7 PM to 9 AM?

If you have read so far, please drop your comment. If not more, you can copy-paste from
- I do agree and want to send my kid
- I do not feel like sending

Those whose response is negative, may please let us know the reason so that we know if we are missing out some very obvious reason for parents not being able to send the kids for such activity.

Learning History with Time-lines

2011-11-21T04:57:00.000-08:00

Read full Article

The Problem:
History as a subject forms a very important aspect of academic curriculum. However
due to the way it is taught in the educational institutions, it does not enhance the
interest of students. The methods employed by the history teachers encourages
student to mug up the dates of historical events. While doing so the students
generally fail to appreciate the significance of an event and its corresponding causes/
outcomes. It is therefore important to look history from a fresh perspective.

Simply said, history is just a bunch of events spread over time. If one looks to
memorize the history in the context of dates, it becomes very cumbersome to
remember it. Moreover such an approach only focuses on passing exams.

The main problem is that we need to remember the events with which we cannot
relate. Typically, we do not even have visual clues about the events happened in
past. This remembrance of events starts getting mixed when volume increases and
we need to remember history of different places, people, objects and subjects e.g.
history of India, Mugal dynasty, earth, literature, astronomy or mathematics.

History becomes more complicated when one tries to relate two events which
occurred around the same time. Since the current teaching methodology focuses on
learning events separately, there is no straightforward way for a student to find out
the impact of one event on another. For example, there is no simple way to explain
how Marathas and Mugals impacted each other.

The Solution:
As they say that a Picture speaks a thousand words. It is a proven fact that visual aids help a lot for remembrance. If historical events can be presented visually on a time-line, it will help people for remembering them in sequence. Each event on the time-line has images and links to other images, movie clips and documents.

Each item has have its own time-line. Zooming-in on a time-line shows more granular events. Time-line of more than one items can be studied together. This helps understanding what was happening with those items in a given time frame.


Time-line of ClubHack conference

Time-line of ClubHack conference Zoomed-in for 2010

Translatting to multiple languages

2010-08-17T05:11:00.000-07:00

On Quora, I saw a question

Where can I translate one word/expression into multiple languages
simultaneously?

5 people were following the question but there was no answer for more than 2 months :(
Hence thought of creating a simple HTML/JS app that can translate text to many languages without taking pains of selecting name of the languages one by one in Google's interface.

Thankfully I got a jquery plugin for translation that internally uses Google translate APIs.
Please use the app and let ne know your comments.

NOTE: Google has withdrawn the free translate APIs (made it paid) so this app stopped working. Let me figure out other ways to put this service back.

Pricing your product

2010-08-02T10:34:00.000-07:00

On Sunday I attended Product Camp. The topic was Product Management and as many of attendees were from software development field, the discussion was mainly on the duties, skills and challenges of a software product manager and hence typically people talked about the trade-offs that a software product manager need to take for defining and prioritizing features .

However, there were physical product designers as well and during the after-talks it came out that they wanted to understand about how to do product pricing. Though I am not a product manager and have never done the product pricing myself but working with startups, right from their conceptualization, I observed product managers coming up with the price. I hope penning down my observations will help in some way.

Initial internal thought process:
It was a hosted (SaaS) version of an on-premises product. The on-premise product was priced at $400, after doing the math, initially everyone thought that $100 is a really good price. It is just 25% of the price of the actual product plus the customer comes out of all the pain of installing the hardware etc. so everyone is going to welcome this price and the sale is going to break the roof in one day :).

Nothing like that happend :(
Problem: Everything in the idea seems okay. Actually 75% off is a great deal. But no-one came up for purchase. Here are the reasons:

Existing customers were aware of the $400 pricing of the product. New offer was really good for them but they have already spent the money. There was no point for them buying the product.
For new customers, it was a new product. They were not aware of the old pricing. And they feel that for a SaaS product, $100 is a big money.

Market Survey:
The word 'Market Survey' gives an impression that you need to have a big questioner and have an army of people going around and interviewing people followed by really heavy statistical calculations :).

Nothing like that happened. We just had a few talks with a few people. These people were either potential customers or others who can provide us a sense of spending pattern of potential customers. During the talks, along with so many other things we asked, directly or indirectly, about how people feel they can pay for the product. Following are some of the findings:

No-one was willing to purchase before a trail
No-one was willing to pay for a trail
General feel was to pay only $25 for the product

The Challenge:
Bridging the gap between $25 expectation and $100 pricing was the key challenge. We really need to come up with the ideas how to recover the costs (and make profit) while keeping the customers feel good about the price. Here are a few points that helped:

Free limited time trail
Batch licensing: This really depend on the product. Our product was such that typically people need more than one license. We imposed a limit of minimum 3 licenses per order
Free Limited features trail
Pitch the product for a such a need so that people attach more value to it. For example, a monitoring product suddenly becomes more valuable if it helps you maintain compliance with the law.

Conclusion:
While choosing the price for the product one should always keep the end customer in mind.
Mostly it is required to make large volume sale for a break-even.
Innovative marketing ideas always help keeping your price point attractive.

Give a few free clicks

2009-08-17T23:20:00.000-07:00

Clicks are always free...who charges money for clicking a link?

Yes, no one charges direct money for clicking a link. But what about links to documents having required registration before you can read them? Startups generally tend to ask for personal information when someone wants to look at a whitepaper or case study or something similar.

Filling out personal information is not direct cash but it does cost. Most of the companies (not companies like Jigsaw who trade contact information) look the user as a potential customer and try contacting back. Such calls/emails really bother me so I would like to avoid filling such forms. Does that means I am losing some interesting and useful information...may be...but Okay. And BTW who knows that the info behind the form is really useful?

When I talk to some friends having such registration on their website, I heard following:

"We want to know who is interested in our stuff so that we can approach her back for bizdev"
"If we shall make everything available via website, what we shall do in pre-sales process"
"The document is our hard work; we do not want anyone to copy our stuff"

All the points above look genuine, but are they really? Let’s zoom in.....

Answer to last point first….you never refuse sending the document to anyone and once I have it, I can copy it as well. You, having just my email address, cannot protect the document. Think of something else :)

Here are two scenarios:

As there is no scarcity of email addresses in the world; Manoj has 5. He has dedicated one of his email accounts for filling out forms. So many companies have Manoj’s email address, and are wasting their resources sending emails to Manoj.

Shailja does not have so many email addresses and she actually looks at the emails sent by the companies, says something like "#@!*&" and deletes them.

What the companies get?

Bottom line is you can only waste your cycles and resources having my email address. If your stuff is interesting and I have the paying capacity for your product/services, I shall contact you anyways. If I am not in position to pay, I shall find something else.

Next, I ask so many questions during pre-sales because they are unanswered. If your online documents answer all my questions without asking them, I am ready to sign-up. Who has time and energy to waste?

So if you are really an expert in some area, it is important to demonstrate your expertise to as many people as possible rather than hiding it. You never know who becomes interested in you. Showing your expertise constantly also gets you people’s mindshare and you are always considered in requirement.

Theory is fine...does it really happen?
Yep, here is an example:

Jai Rawat created a comprehensive presentation about Display Ad Landscape and made it available for free via slideshare. In just 2 days about 300 people viewed the presentation, digged it and it came on the top of Google search. This also increased people’s interest in Jai’s company Velocita.

Do you know any other way to top the Google and gain visibility overnight?

Marketing: Branding: Name

2009-04-02T21:18:00.000-07:00

Why Branding
When you are creating a business (company and/or product), you need to market it. (I shall again emphasize here that the marketing should start before you start your product/service.) Marketing is nothing but is to create awareness of your business in the market. Basically, people should be able to relate to your business, remember it and feel so good about it that they become ready to pay for your offering.

What is Branding
Branding is nothing but identification of your business. A brand registered in people's mind, help a lot at the time of sales. Some of the important items related to branding are:
1. Name of company
2. Name of product
3. Logo
4. Tag line
5. Tag line Audio clip (not everyone has that)
6. Corporate color theme
...
... (many other things can be included here)

Human Psychology
As the branding has to be registered in people's mind, it has to do a lot with human psychology. Hence while deciding on branding, target audience has to be kept in mind. For example, name of a product that is going to be used by school students will be different from the name of a product (may be with similar features) used by senior business executives. This is even more applicable for tag line as tag line is a phrase that acts as a bridge between the person's feeling and the product.

Color to be used has also to do a lot with the psychology. You must have seen most of faminine products in pink, natural products in green etc.

Problem with Multiple Names
Human mind recalls similar things quickly. If you have no similarity amongst the names you are using for your business (in other words you have multiple brands), you need to put effort separately for popularizing each of them.
Unlike that, if you have same name everywhere, you can put all the money, time and effort just for one brand. This suggests that you choose a name such that it represents your business theme and can be given to everything associated (of course with a little variations) viz. company, products, website, services etc. Examples are www.iscaleup.com, iScaleUp Pvt Ltd, iScaleUp Tester, iScaleUp Diagnosticks, iScaleUp Practice, iScaleUp Speed etc. (I hope you realized that first name is web address, second is company name and rest are product/service offerings)

Finalizing a Name
Amongst all to whom you need to give a name, website is one where you have limited choices. You need to see if the domain name for your desired name is freely available. Mostly you will find that the domain name is not free. Actually some people do that business; they purchase domain names and resale them when you are in need of them. The price in that case depends on your need. If you are starting a business, you may not afford the price and you may want to go with something else that is available freely (not for free but in open market and for normal price).

In this case you may like to add a character to your desired name. You have option to put a dash (-) of underscore (_) in between words. Think again before you do that .... Why you wanted to have a short and sweet name .... just because one can communicate it verbally ..... Right? A dash or underscore make it hard to speak, understand and remember. So this option is ruled out. You can mis-spell, but again when someone speaks other one is going to go to the site with correct spelling. Do you really want to do publicity for someone else?

Using a letter as prefix to your desired name works mostly. The most sounding letter in this case (for software) is 'i' and next comes 'e'. For example, if you take 'talk' as an example of desired name, maximum acceptance will be for 'iTalk' and then for 'eTalk' as people will be able to relate to the name intuitively without reading much of the text. However, if some brand is already very popular, letter from the brand can be borrowed e.g. Google can have any name starting with 'G'.

What really works
All said and done, success of marketing greatly depends on the social behavior and uncertainty attached to all the laws of social sciences prevail. You never know what really clicks (or does not click) unless you try it and have actual results with you.
Confused ..... that's marketing ..... :)

Creating a successfull product

2008-12-01T21:07:00.000-08:00

Product sale happens by marketing not by engg (features).
So marketing should start even before you come up with the product.
Sale a concept not the product. People look for the solution to their pain rather then having a product first and then looking at what it can do for them.

At initial stage, put more focus on online marketing rather then traditional one. Get some articles on the web regularly. When your product mature, your marketing material will also be accumulated till that time (and of course competitor as well). When the sales team will pitch the product to some customer, they will not be able to show the product in reference to competitor's product because of availability and time constraints. At that time it is the material accumulated over the time on the web that helps.

Your product may be useful for everybody in the world. but do not try to sale to everyone. Identify just one vertical and one market segment at a time for sale. Focus all your energies to make the product successful for the chosen ones. Though at technical level your product may satisfy everyone's need but to pitch to different verticals (and market segments), you need to create different stories. Having too many stories confuse everyone including yourself.

You may be having 100 features in your mind. Do not try to put all of them into the 1.0 product especially not the half backed ones. Go to market early with less but complete features. Again remember nobody can compare 2 products feature by feature all the time. If a feature does not work properly, competition will make a big deal out of it and you will loose in the market. Keep your road-map for missing features ready.

Programmers care least about looks of the product. Take the services of a user interface and interaction designer to carefully design UI and work-flow. Whatever story you are pitching in your marking material should come out of the product UI automatically. Users should not need to receive training or read help.

If you are creating something new, you can never be in pull market. So need to push your product; nobody will come to you to purchase. People wold like to try the product also before purchasing. Let people play with the product freely. The product trial has to be really quick. Nobody comes back after a day to try the product. People even do not like to provide their personal info (including email address) just to look at something. Try to come up with a process such that he customer is able to get a feel of your product without even giving email address and is able to try immediately providing minimum contact info.

Let some basic features of the product be available for free all the time for trial. This captures customers for you. Charge for the premium features and/or volumes. Having more customers on-board will also help you quickly shape your product and product road-map.

Your big prospects will always come up with a list of missing features as 'required' to avoid your sales person. Make sure that you listen to customers more than prospects. If you change your road-map and development plan based on these inputs, you will end up producing bad things. Such prospects always deny for purchase telling some other reason when you go to them after creating their list of features. Focus more on the people who are already using your product; keep them happy, they will do free publicity for you.

Starting a service project

2008-12-01T20:25:00.000-08:00

While being in a service company, you come across the situation where you need start a new project for a new client. Many times when the client is also a small startup company, starting the project with the state of flux is okay. But if you are dealing with a big client, you need to take ultimate caution. Client helps lees. expects too much and scare you to scrap the project.

So IMO following should be taken care of before starting the project:

Have the project lead on board: You may want to move someone from other project or share somebodies time to minimize the cost but that comes with trouble. The project lead should be able to propare himself to spend significant amount of time on the project.

Project ramp-up plan: At the time of agreement itself, propose a very high level project ram-up plan to set the client's expectation right. Ramp-up plan should have time allocated for understanding the existing system, requirements/expectation, team building, trails and making a plan for actual project execution. Remember that people at client's end know much more about their product/work and unless explicitly stated, they will not understand (or be willing to understand) the time taken in above mentioned activities. This plan should be made by the project manager (who is talking to client initially) along with project lead and project lead should start working on this plan.

Do not club the project with other project: Unless the project is really being managed in a clubbed manner at client's end also, you should not try to club the project with other at your end for ease of management. you can always do this later when the project is running smoothly. In initial phase, project requires a lot of attention and hence has to be handled separately.

Assemble initial team quickly: Quickly get at least one person in each function that you are starting the project so that the initial work can be shown asap. The project lead has to identify these functions and has to prepare a very high level plan (may be shared with the client if required) for each function to start. These team members are required to understand actual work, try it and start planning for execution.

What is in the blog

2008-07-24T22:25:00.000-07:00

The idea behind the blog is to put together the events and works done by me.
It is actually intended to extract and list learnings from the past for a better future.
Currently I am posting personal history; I shall post the learnings later.

Why This blog

2008-07-24T22:02:00.000-07:00

So many things happened in my life. ......Actually happen in everybody's life.....nothing special about me. But I feel special about them and hence this blog......

Complicated??

Events in the life happen good as well as bad. At least some events look bad at the time they happen but later we can find good part of those bad events also. So the idea is that one should not only face the hard events firmly but also try to learn from those events and try to find good part of those events. That's why I named the blog like this.

Things happened in past but I will write them now ..... what will happen to chronology and sequence? I am not talented like Gulzar to link events from different times (remember Ijazat) so I try to maintain chronology using other means for sequential and easy reading.

Managing Release process for SpectraGaurd Suite

2008-05-24T23:13:00.000-07:00

Continuing focus of quality development and delivery, I identified areas in software development process that can be improved for better information flow between product development and testing team, strengthen impact analysis and provide better visibility to engineering and testing managers into ongoing activities. To enforce quality, I linked source code version control system to bug tracking system and bug tracking to test case management system. To facilitate information flow, visibility and analysis of the situation, I developed custom web interfaces on top of these tools. I have seen challenges and inefficiencies coming from our own branching strategies so I have studied standard processes around branching and merging and suggested different ways of branching and merging for different types of development scenarios.
Because of breadth of my knowledge across products and impact analysis skills, management has also given me responsibility of software configuration management and release process management. Additionally, because of troubleshooting skills, customer support group always looked at me for technical help.

Headed Release group with complete ownership of release process for multiple product lines.

Coordinated with Engineering, Product Management, Documentation, Web development teams ensure timely GA of the new releases
Handled 16 major releases in 2.5 years for 5 product lines

Implemented development process improvements

Rolling out configuration management practices and monitoring tools
Integrating and/or customizing CVS, Bugzilla and Testlink to accelerate development and testing processes
Revamping prevalent CVS branching strategy to allow flexibility, greater efficiency and reduced complexity

Challenge

Getting buy-in from multiple stakeholders to implement process changes without affecting turn-around time

Other Activities

Technical guidance to customer support team to address customer queries
Mentoring and coaching of newcomers
Management of multiple small teams
Product, technology and process training to partners as well as new hires in the company

Database Design and Performance Optimization

2007-07-24T23:24:00.000-07:00

During the implementation of new upgrade architecture, I suggested change to many product development and testing processes and management accepted and implemented them all. I found that the way of upgrade process directly impacts the way of implementation of database, build & packaging process and installation process. Over the time management has given responsibility of all these functions to me and I overhauled all these functions. During this course I designed tools to simplify and automate build process. I also designed a tool to sanity check of database implementation to enable high quality bug free coding efficiently. As result of all this, the server upgrade of product never failed in field. Having responsibility of database implementation, I took part in database design process and improved it significantly.

Formed specialized Database group
Reviewed database entity model and suggested design changes to improve performance and reduce complexity
Created a test harness to compare sql query response times
Changed Linux and database server configuration to optimize database performance

Analytics Support to Web-based Marketing Campaign

2007-07-24T22:54:00.000-07:00

* Management of Google Adword program

AirTight has the business in the area having very low or no search volume. But many areas relate to AirTight®’s business. Earlier AirTight was bidding on the keywords but not getting the traction. Keywords were inactive in spite of high bid, there were very less impressions/clicks and bounce rate was high. I analyzed the state based on with help of Google Adword help and realized that we need to put Ad on the keywords having traffic and then escort the person to the solutions offered by AirTight by providing the missing piece of information. Here I suggested a 3 page model. The first page is talking about exactly what the user was looking for and offers what the Ad offers to the user. The second page describes, how the area in which AirTight works is related to the term the user was looking for. Third page is the AirTight’s solution page that gives the marketing pitch.

* Design of tools for analysis of Google Search results and placement of Google Ads

The idea was to understand where AirTight and its competitor stand in Google and yahoo search results when a user searches for a related keyword. This info helps in 2 ways: 1. AirTight can stop bidding on the keyword for which it is in the top of search results. 2. Apply focused energy where AirTight is lacking. I designed a combination of a bash script and a web application in php to record the position information from search result at a given time. After collecting the data over a period of time, the data can be plotted as graph to analyze the trend.

Another idea was to find out the places where AirTight should put Ads on content network. The objective of this tool was to find out which websites that show-up high in the search results of Google and Yahoo have the Ads by Google on the landing page.

* Design of tools for analysis of click-stream data coming from website

AirTight was using Stacounter to collect activities of the website-visitors. Earlier marketing team was using Statcounter’s web interface for analyzing the data and figuring out the potential leads. This was a time and effort consuming task and the person analyzing was getting limited time info about the users. I designed a tool to put the raw data into a in-house database so that the info of all the visits remains recorded and we can do better analysis. First phase of analysis-automation was to provide info about the organization from where the visitor came along with all the historical activities of a visitor. Second phase is to do the profiling of visits and qualify a visitor as a potential lead based on the pages visited by him. Other extensions of this tool includes understanding the intention and interest of an organization even if the visits are from different IP address and identifying the pages on the website that should be optimized/better written.

* SE optimization of corporate website

Suggested changes in web pages to make them web page friendly. Most of the knowledge came from Google help. This activity includes changing title, meta keywords, meta description, headings, alt text of images and content in a way such that they remain focus on a single topic and come in the top of search results.

* Creating strategies for online marketing (awareness generation)

Suggested different methods of making the technology and the product popular in the global marketplace using web2.0 tools such as blogs, forums, social bookmarking, feeds etc.

This area of work involved:

Gathering relevant data about

· Organic search results leading to relevant pages

· Google Ad placement and ROI

· Page visit analysis

Analysis and recommendations for

· Content modifications for improved organic search outcome

· Google Ad placement decisions (considering location of Ad and ROI)

· Content reorganization to improve visit time and pages viewed

· Content modification to ensure clear relationship between Ad content and page content

Development of analysis tools for

· Gathering click-stream data from multiple sources like Google Analytics, Statcounter etc.

· Filtering of data to gather relevant information

· Visualizing search results to help make right recommendations

Challenges:

Ad placement decision for ROI maximization (in presence of low search volume)
Non-intuitive content reorganization based on search volume and page visit analysis

Technologies and Tools:

Google Analytics, Statcounter, Market2Lead, Google Adwords, Google Trends, Google Keyword Tool, Google Traffic Estimator, Google Advance Search, Yahoo Site Explorer, Alexa, Technorati, Google Webmaster’s Tools, Keyword Density Tool, HTTP header viewer
HTML, PHP, ASP, Javascript, CSS, Linux shell scripting, awk, sed
Content Management System (Typo3)

Role and Responsibilities:

Collect data about website from different sources and send alerts to top management
Identify improvement areas and present them (with solutions) to the US marketing team
Coordinate with marketing, web development, technical writing and R&D teams to implement the changes recommended

Solving upgrade issue of SpectraGuard and other development process improvements

2005-07-24T22:38:00.000-07:00

In 2005, AirTight released the product to very small set of customers. I was primarily responsible for product verification (software testing) at that time and also looking after customer support. Almost all the customers were reporting data loss during the upgrade. We have already tested the product properly and found it working. However, upgrade bundle was prepared late in the cycle and the assumption was if the freshly installed product works fine, upgraded one will also work. But that was not happening in the field.

I analyzed the process and found that AirTight’s product has a different relationship with data and hence upgrade of the data has to be dealt differently. I proposed the new architecture of upgrade solving both the problems of data loss and timeliness. Senior management of AirTight liked the idea and gave me the responsibility to get that implemented. This architecture got very tough challenges from the branching strategy and parallel development prevailing at that point in time.

Founded CLI group taking full responsibility of product installation, upgrade

· Definition of long lasting upgrade architecture to support version dependent, automated data migration and operating environment upgrades during installation of new version

· Management of disk partitions to utilize space available on newer drives as well as operating system and product software upgrades in a seamless manner

Created in-house tools for build and sanity test execution

· This involved development and maintenance of build scripts and creation of database sanity tools to ensure build sanity

Challenges

· AirTight ships its main product as an appliance and set of sensor boxes

· The boxes have no interface for external media; only Ethernet is available for communication

· The system stores different types of data viz. information about wireless air-space, policy and other configurations defined by administrators, metadata and environment configuration parameters etc.

· The boxes run customized Linux as base OS and users have access to a restricted shell only which require handling runtime environment changes along with seamless data migration, during upgrade

Upgrade is carried out by customer in absence of AirTight engineer while the customer might have skipped last few releases

Software Testing for Wibhu Technologies (now AirTight Network)

2004-07-24T22:35:00.000-07:00

I worked with the founder of AirTight for creating (both development and testing) a prototype of the idea.

After he started the company at Pune, I used to do software testing assignments from Kanpur on part time basis. After successful completion of work at DGP project, I decided to join AirTight at Pune on full time testing assignment. Knowing the evolution of the product right from its prototype, domain expertise gained during wi-fi deployment and experience of software testing were advantages for me. At that time, company was in its initial stage, software release cycles were short, delivery was not packaged, testing processes were not being followed and testers used to do manual exploratory testing only. Over the time I started demanding packaged installable software and the information of what is coming next to facilitate useful test case writing.

I started load and longevity testing of the product and written a few tools to collect data about responsiveness of the product over time under different conditions. While testing team was focusing on black-box testing I analyzed the delays in response further and found that the delays are happening in database calls. Getting this info, I worked on multiple fronts to solve this issue. First I suggested change in the style of database design to improve performance and reduce complexity. Next I started query optimization exercise with development team members and then did tweaks in database server configuration to optimize the performance. For this task I created a test harness to compare performance of two given queries eliminating the statistical variation.

When AirTight decided to have automation testing framework in place, I was given responsibility to work with testing automation company and identify parts of the product for which testing can be automated. Later when company decided to outsource testing of another product for a release, I was appointed as point person for all technical matters.

When AirTight started shipping the product to customers, engineering team faced the challenge to work on multiple releases at the same time. The challenge was there because of high number of small usability issues and small change requests were being done from the field and AirTight was handling them in ad-hoc manner. I suggested a parallel development and micro-release system for better handling of requests and keeping the main course of development intact.

I also did manual exploratory testing of both the products of AirTight, wrote test cases, did planning and estimation and managed a small team of testers.

Long Haul Wi-Fi Deployment at IITK

2003-07-24T22:33:00.000-07:00

* Configured and used Wi-Fi Equipment
* Demonstrated long range wireless network
* Setup multi-hop wireless network between IIT Kanpur and Lucknow
* PC software development using Visual Basic 6.0 and Microsoft Access

The Digital Gangetic Plain (DGP) Project of IIT Kanpur is intended to connect rural-India using the 802.11b wireless LAN technology (a.k.a. Wi-Fi). I am one of the people who are involved in project since starting of the work. We have successfully proved the concept that Wi-Fi technology which is primarily meant for indoor use, can be used in outdoor for long range connectivity and virtually any distance can be achieved using multi-hopping. Currently I am involved in developing a suitable techno-economic model
Along with this I developed PC software using VB to keep track of the items purchased by the project. Some Linux-shell scripts are also developed to monitor the network.

Work at PDDUSDV

2001-07-24T22:32:00.000-07:00

* Established and managed LAN having 23 Windows98 PCs, 1 Windows2000 Advanced Server and 1 Linux Server.
* Software development for school admission process using Visual Basic 6.0 and Microsoft Access
* Linux based portal for UP industries database using PHP and MySQL.
* Coordinated development of animated educational movies using Macromedia Flash 5.0
* Teaching computers for class 11

Apart from managing the LAN and servers of the schools lab, I some standalone and network based PC software using VB6 as front end and MS Access as beck end. Those are meant to automate some processes related to entrance examination and admissions.
A portal www.iitk.ac.in/aicte was developed under a project of Institute Industries partnership Cell, IIT Kanpur to web enable the database of Industries of UP and Uttranchal. PHP is used through Apache web server on Linux to present the appropriate data requested by user after extracting it from MySQL database.
Online examination software was also developed by me firstly using VB and MS Access and later using PHP and MySQL on Linux.

PC Software Development at UV Export

1999-07-24T21:59:00.000-07:00

In this organization I, as a team of two, developed some application software using VB and MS Access. Those software are primarily intended to keep data from user input and to generate reports and alerts periodically. Custom ActivX controls and DLLs are also developed to facilitate some tasks.
A web application was also developed to run on intranet using Microsoft Technologies. That application was intended to show the data as per users’ requirement and to process users’ input.