Garvit Khatri

How to learn a language? (it's not programing language :P)

Fri, 20 Apr 2018 11:00:00 +0000

In this blog, I would try to relate the process of speaking a language and brain. This blog post is purely fictional, based on my imagination, and I don’t have any biological/experimental proof to verify the things written on this blog.

How do we think? How do we speak? Why are we good at one language while stammering in another?

These were the questions in the back of my mind. I sat one night and tried my best to answer them with the knowledge I possess. I was looking for a logical answer which would fit well in the real world, though may not reflect the way things work. I started with asking myself questions like “Do I know all the grammar I need for my native language?” The answer was no. Then how come I mostly use the right tense and grammar while speaking? Then I asked myself “Do I know the grammar of the foreign language?” Yes, my mind answered. Well, then why do I stammer and use fillers in the foreign language and not in my native language?

With these questions in mind, I thought of breaking the problem into smaller parts. First part is the process of figuring out the content/keywords that I want to deliver. Next is knitting them into meaningful sentences. Finally speaking out to express my views to the world.

To think and come up with the content I want to speak is processed by my conscious mind. Now, think of the conscious mind as CPU of my brain. It processes any random request sent to my brain and is good at it, but the downside is that it is a little bit slow and not very good at repetitive tasks like phrasing keywords into sentences. While on the other hand the stitching of those content into sentences is done by our subconscious mind, which is like a GPU, hence is very fast in performing repetitive tasks. The subconscious mind works on a machine learning model, which has to be trained with examples. Thus to make it quicker and better you need to teach your brain with examples.

Now with the above considerations, let us try to answer the question Why do we speak our native language so quickly? Even though we may not have a formal education in our native language, we speak it fluently. The reason is that we have trained our subconscious mind in our native language. We have been listening and speaking in this language since our birth. Once our subconscious mind is prepared, it is fast in knitting the keywords together. Hence most of the time is taken only in thinking what we want to speak and not on making the sentence.

Now let us consider the case when we try to speak a foreign language. Now our subconscious mind is not trained in it, and our conscious mind has to perform both the tasks of thinking what to talk and then forming a sentence. Since our conscious mind is not trained for such jobs, it achieves them at a very slow speed compared to subconscious mind. Hence we think a lot before while we have to speak in a foreign language.

Why do we stammer? Why do we use fillers? So now there are two possibilities which could happen. First is that you are thinking about the content you have to speak, but in the meantime there is silence, and you come under the pressure of talking something as soon as possible. Now what happens is that your mind knowing that it would need time to construct the sentence as well, it thinks half the content, process it into a sentence and you speak that out to fill the awkward silence. Once done you construct rest of the part and continue the process. The second possibility is that you have a limited capability to process and retain. Now you would have thought of all the content you have to speak. Your mind starts to transform the material into sentences, while doing so it knows that the content you have thought is important and must be stored in brain unless you finish what you want to convey. Hence due to the shortage of memory in your mind, your mind processes the sentence half and then sends it out, and you speak it making space for the next part of the sentence to be prepared and hence in both the cases, we use fillers, or we stammer while we are speaking a foreign language.

Knowing all this, How can we improve our grasp over a foreign language? The solution is to train our subconscious mind, but how? The easiest and the best way is to speak the foreign language more in our day to day life. The more we use and talk, the more our subconscious mind would be trained, and keep on getting better. Yes, we would stammer and take time in the beginning, but unless we teach our subconscious mind, we won’t be able to speak appropriately ever. Also converse more in the foreign language with a native speaker or the one who is already fluent. Talking would help us in two ways. First, we would listen to the correct way of speaking, which would assist in training our subconscious mind. Secondly, ask them to point out few silly mistakes you make. Aforementioned is another excellent way to receive feedback and improve your machine learning model in the brain.

So next time if you are learning a new language, keep these things in mind ;)

I hope it was a good read for you.

Editing thanks to Arushi Mittal

Happy Learning 🤩🤩

Zomato - my first love

Thu, 12 Apr 2018 11:00:00 +0000

I believe a person has a special place for the first company he/she works for. Zomato being my first company acquired that place in my heart. As a graduate straight out of college, I was excited and enthusiastic to welcome this change in my life. My journey began on 4th of July 2016. On the day of orientation, HR told me that I would be working with the Web team at Zomato, whose primary focus was to work under the domain of “zomato.com”. This felt special and alarming, for the responsibilities I would have for a consumer-facing product. This experience led me to believe how much faith this company shows towards their employees. I won’t shy away from saying that though I was excited to work, I was also a little depressed to see lack of proper frameworks, coding guidelines, tests etc. Nevertheless, I was interested in facing the challenges that would come my way.

Team

I along with Areeb, Ankit, Prabhanshu and Aniket joined the web team and were introduced to Sparshith, Nidhi, Ashish Kanwar and Gunjan (CTO Zomato). It was an immense pleasure to be working under Gunjan. He is one of the most grounded people I have ever met; his dedication towards his work inspired me. His small gestures like to take out his time to listen to my most straightforward queries made me his fan. The best thing about the team was that they all were passionate about software engineering and work was not work but play for them. They would be in an energised state when faced with a challenging problem. I soon become a part of this loving family and never felt like an outsider.

My first project was to work on lazy loading images on our collections page. I worked along with Ashish Kumar Kanwar on this project. Working with him, I learned the art of never giving up and keep on pushing my boundaries to achieve more. Moving forward Ashish and I become terrific friends, and I hope we would stay the same for the rest of our lives. Unfortunately, I enjoyed a short period working with Sparshith as he decided to move on in his career. Though in this short time span, I learnt the importance of managing time from him. Nidhi to me is a living example of how to keep on learning and exploring. From working on HTML, CSS, js, to working on nodejs, PHP, go, she knew it all. She never said no to a challenge posed in front of her and kept on growing in the process. While I learnt a lot from my seniors, I also learnt a lot from the people who joined in with me. One of them was Ankit; his experience helped me in understanding the importance of teamwork. How essential it is to take the team together and have a feeling of ownership of the project you are working on.

Since I said it was more of a family than a team, we not only worked together but rather played, partied and shared personal as well as professional learnings with each other. Any person was just a walk away.

culture

To me, Zomato meant open-ness, freedom, rush and 1% done. People at Zomato are given the freedom to choose the technology that they feel is best for the task in hand. People were open to discussions regarding the new technology and ways to do a certain task. This freedom not only led me to add linters in our PHP code base but also introduce a minimalistic slim framework while I was working on PWA Apis. During the journey we adopted Reactjs, go and even started writing our first user service inside Java. From a monolith to breaking zomato into services and UI kits, it was indeed a journey full of learnings. Discussions with teammates, learning Fridays, sharing things you find interesting, etc. all will be missed. The benchmark is set and high, and I hope my future endeavours don’t disappoint me.

learnings

What did I learn at Zomato? A lot. I present the things I feel are most important in no particular order.

Ask the question why? You may already have a solution to it. As a developer/engineer, your work is not only to do the stuff that management wants instead engineer’s job is to listen to the problem at hand and then find the best possible solution.
TEST! Yes, it’s not just testers job, it’s yours too! Make sure you have thought off all the edge cases your self and then tested them. Your goal should be to make sure testers are not able to find a bug in your code.
As I worked on UI as well, pay attention to the details. It is essential and makes a lot of difference between a 2px border radius, 4px border-radius and no border radius at all. Have a good look at the design, zoom in and even suggest if you think designers missed out on something.
Grid in the mind of a designer is so much different from the row-col based grid we use in HTML. The grid in designer’s mind is the global grid across the site while in HTML we know bootstrap way grid is nested. When you start working on a new design, confirm the grid size used by the designer and make your grid decisions accordingly.
Think of all future possibilities, it’s better to spend an hour first thinking about the options than to code asap and hack the future upcoming into the system.

It was a great experience working with Zomato; I wish Zomato best for their future. For me, I have moved on and joined Adobe. I am taking all these learnings with me, let’s see what do they have in store for me ;)

#happyHacking

Amit Kumar Jha

Mon, 01 Jan 2018 04:00:00 +0000

It all started way back in May of 2016, a guy wearing helmet, on a bike stopped me in the middle of the road. This guy didn’t even care removing his helmet and out of nowhere asked “Hey, are you Garvit? Are you the one going to Singapore?”. While I couldn’t understand what was happening, I politely replied yes. That is how I met Amit Kumar Jha.

My second encounter with him was in Singapore where he hopped from one place to another. This guy stayed at different place every day during his stay in Singapore. At first, I found this guy to be the unorganized and careless. He even forgot to book his ticket to Singapore where he was speaking at FOSSASIA conference. I met a person with such an attitude in life for the first time. In Singapore, I got to know he is a core contributor of Sympy Developer and was a GSoC student with Python Sofware Foundation for his project with Sympy.

About six months later we met again at Zomato office. He was at the office to give an interview for the Software Development Engineer post. He got the position and joined about a week later. He asked if he could move in with me for a month. I was living all alone and had no problem in sharing my room with him.

That was the start of our friendship; we stayed together for a month and a half. With him, joined another person Suhaib who lived near to our flat. We often spent time at Suhaib’s flat. There we played cards, enjoyed songs on youtube, danced and did all the crazy things you would expect out of newly graduated students. I still have a vivid memory of an incident. He had to prepare a talk on GIL lock in Python language. We all stayed up till late night and figured out what exactly GIL lock is and why do we need it. Suhaib and I dozed off once we acquired a substantial knowledge, but this guy stayed almost all night and prepared his presentation. This incident made me realize that he takes his presentations very seriously. He had to present this at a regional meetup of Python Delhi group. He told me that he prepares a presentation, submit it at a regional group, then at the national level and finally internationally. While after every presentation he would take feedback and improve his content. This is what made him an elegant international speaker. He speaks at various international conferences. I did use his material and the knowledge that we gained that night to present a talk on GIL in another regional conference of my organization KDE.

Following six months down the lane, me, Suhaib, and Amit decided to take a flat together. We found a perfect home for three of us, and we named it “Haveli.” We were joined by Amit kushwaha after we moved to our new haveli and a frequent visitor Anu Mittal. Haveli became the host of unlimited fun, memories and joyful events. We decorated our haveli with bean bags, PlayStation, and TV. Though Amit and I never played it :P, but TV was utilized to the best of its capabilities. We watched movies, listen to tons of qawwali by Nusrat Fateh Ali Khan and songs by Momina Mustehsan, Atif Aslam, and others.

All three of us used to live in Delhi and used to go back to our house every weekend. Amit used to drive us all back home and pick us every Sunday night back to the haveli. This showed his selflessness and how much he cared for all of us. He had to many time take the long route back to his home to drop both of us to our respective house. I never saw him whining about such events. We even used to call him our Uber driver. Such are the memories associated with him. I am omitting out many of the other memories because we can’t share many of those in public ;).

We were enjoying our daily course of events when he told us that he got a job offer from a company based out of London. He told us that he would apply for exceptional talent visa for the UK. This news got all of us excited. After about a month of his research and collecting the relevant documents he submitted all the related documents needed. About three weeks from this he got a mail stating that he has cleared his first round and need to send his medicals and other necessary documents. This made sure that his selection for exceptional talent visa was almost confirmed and it was mere formalities that are necessary to be completed. Eventually, he got his exceptional talent visa. This news brought joy to all of us and made us proud of him.

Today, I am writing this blog as we waved him off at the airport. He to me is a great friend, teacher, and inspiration. I will miss him; this Haveli will miss him. Those sleepless nights would not be same without him. But more than that, it is the joy of having a friend having exceptional talent :D. I wish him all the luck in his life, hope he will rock at his new job in London too.

All the best Amit (#AKTECH) :)

LinuxChix Meetup

Mon, 18 Sep 2017 04:00:00 +0000

Yesterday, I gave a talk on Linux file systems at LinuxChix Delhi meetup group. I have to firstly apologize to all the participants and the organizers, Shivani Bhardwaj and Priyal Trivedi for reaching the venue late. It is a great effort by both of them to enlighten the youth and female community about the advancements in Linux. They are trying their best to help them learn basics, benefits and enormous opportunities that Linux exposes. It is optimistic to see the female community taking such an initiative in the software industry. The best part of the group is that they welcome all, young, male, female, professionals and students to join and share their thoughts and experience.

Coming back to me being late, I had my talk scheduled at 5:00 pm and I reached 5:10 pm at the venue and sadly missed all the talks for that day. I heard that talk by Suhaib’s on “How SSH Works” was fantastic and so was Anu’s talk sharing her experience on Linux. You can find more about Anu’s talk and get the slides from her blog. Apparently, all of the speakers were from Zomato. Hence felt like a reunion to meet them at the meetup.

About my talk, it started with the history of why the extended file system (Ext) came into existence. I covered the file system with which Linux started, i.e., Minix (Mini-Unix) file system. Continuing to the presence of Ext and Ext2 along with Xia file system. Here is a snippet from my presentation which shows a quick comparison between all the filesystems mentioned above:

Further, I explained the basic elements of a file system, i.e., inodes, directories, and links. I explained the need for a virtual file system layer that was added by Chris Provenzano and how it helped abstracting out file systems for application layer while never wandering about the implementation of the actual file system on the hardware. It was much of learning for me too as I found many facts and stories while looking out for the content for this presentation. You can find my presentation hosted on the drive, with this link. I took most of my inspiration from the research paper published by Rémy Card, Theodore Ts’o, and Stephen Tweedie. You can find their research paper here.

To end this blog, as the organizers of LinuxChix say, they need more participation from the female community and they want them to come forward and share their experience. So from this blog, I am inviting all the readers to join the linuxChix group and be a part of this initiative. You can find them from meetup page and also join telegram group.

Adios till my next blog, happy hacking :)

5 Myths about reading

Sat, 02 Sep 2017 06:29:31 +0000

This blog post is not about giving you instructions on how to read; rather it illustrates my findings on how to read. Recently, I started reading about “how to read optimally?”. During my expedition, I came across many articles which helped me burst the common myths about reading.

The first question thrown at me was: What is the fundamental difference between a competent reader and an ineffective reader? An efficient reader has the right mindset towards reading. They know why they are reading and what they should expect out of the book. On the other hand, an ineffective reader has many misconceptions about reading. I will be discussing few of those misunderstandings in this blog.

Myth 1. You cannot read more than 1k words per minute

This is a misconception that many people have. Individuals, with practice, have attained this reading speed and still understand the text. Reading is a skill that can be improved over time. Thinking that you can improve indefinitely would be foolish. There is a limit till which humans can read and comprehend the text simultaneously. After a certain threshold, we start reading in our unconscious mind, and our comprehension becomes inadequate. Hence realizing your toppling point helps you improve your reading capabilities.

Myth 2. You can understand 100% of the text

You can never understand the complete text in one go. You may require multiple readings of the book. Even after multiple sessions, if you pick up the book after a period you could still get new insights from the book. This happens because over the period you have improved and haveehad many changes in your perceptions. Hence one should never focus on getting 100% out of a book, rather concentrate on getting the main gist of the book.

Myth 3. You must read slowly to get better comprehension

Everyone has his/her reading speed and comprehension level. Your reading speed doesn’t justify your capability of understanding the text. It is a misconception that reading faster can reduce your comprehension of the text. Yes, after you have reached your usual reading speed, comprehension level starts to decrease, but this doesn’t mean that you should always read slowly and it will always help you in understanding the text better. Many also have a misconception that one should read slowly when reading novels or poetry to appreciate it. This thought is not right; you would miss the rhythm of the text if you go too slow. Hence it is important that you understand your speed of reading does not relate with the comprehension of the text.

Myth 4. You should keep a dictionary with you while reading

Instant lookup to difficult words distracts you and breaks your reading flow. It acts as an obstruction in reading. Mostly you would understand the bigger meaning even if you don’t understand the significance of that particular word. Rather you should highlight the words that you find challenging and revisit them once that reading session is over. This would help both in developing the vocabulary and also improving overall reading experience.

Myth 5. Word-to-word reading is best

When you focus on word to word reading, you have a high tendency of losing the context of the bigger picture. Hence, one should read the complete phrase in one go rather than focusing on individual words. Many people believe that skipping words depicts your laziness. This is not true; rather humans do tend to skip words when focusing on the whole phrase and it is entirely reasonable to do so. Hence, you should not put an extra effort in reading every word of the sentence.

I find the above-listed points helpful and hope that they help you as well. I hope next time you pick up a book, these points will come to your head and will assist you with a better reading experience.

Adios till my next blog, happy reading :)

QtCon && Akademy 2016

Wed, 14 Sep 2016 14:34:25 +0000

This year Akademy took place along with QtCon in Berlin, Germany. 2016 is a special year for many FOSS projects: KDE has its 20th birthday while FSFE and VideoLAN both have their 15th anniversary. So, this year KDE, Qt, FSFE, VideoLAN and KDAB came together to bring QtCon, where attendees met, collaborate and got the latest news of all these projects at the bcc in Berlin from the 1st to the 4th of September 2016. Following QtCon was KDE’s Akademy BoF.

This was my first QtCon and Akademy. I was supposed to arrive on 3rd September, 9:00 am. Yeah, I know I was already two days behind the schedule but…. Anyways I still had two days and my talk on 3rd September. But, unfortunately, my flight got cancelled. Flight was rescheduled and reached Berlin at 9:00 pm instead. Sadly I also missed my talk on 3rd of September… :(

But I managed to attend the last day at QtCon, and it was incredible. I know I did miss a lot of things but still cherished the events I attended.

Firstly, I met Lydia, President of KDE Ev. This was my first meeting with her, and it was a delight to meet her in person.

Then I attended few of the talks which included a talk on ThreadWeaver by Mirko Boehm. He explained basics of ThreadWaver library by KDE and told about runnable, thread pools, and sequences. It was more of an introductory talk on how it works and where it can be used than diving into API. I guess this was much needed, I hope I will try and tweak some things around it soon :) Next was a talk by Jos van den Oever on how to use automated IO code generators and why writing your own IO read/write code is a dreadful idea. He also explained about his experience in reading PPT format file and how he wrote XML description for it, which was then translated into code using IO code generators for file formats. Another fascinating talk was on Open Web Services by Cornelius Schumacher. He talked about problems we face and What can we do to make a web service Open Source. He also mentioned maybe we could have a certification authority that makes sure that data on the web service is safe, secure and is not miss used, and uploader of the data have the complete ownership of that data. It still needs to have long discussions as many points were still not covered. Then I was video operator in the talk on GCompris by Sagar Chand Agrawal. He explained why they shifted their app to Qt and what do they share in common with KDE. Lastly, Ken Vermette told us about how we should change flow and design of our application. How it adversely affect the user’s mind if we change the complete design and keep flow unchanged or vice-versa. He also told us how we should change design first, let the user get acquainted with the new design and then gradually improve the workflow of the application.

Here are some of the pics

I spent some time with the FSFE guys and yes I got some stickers from them :D

Representatives from FOSSASIA were also part of the event. It was great to reunite with them; we had a long chat with them discussed problems existing in India and how we can solve them through open source technologies.

At last, it was a blasting experience, and I took a lot of memories from this trip. After QtCon I went to Toulouse, France to meet another KDE contributor Bruno. Here’s a small picture with him ;)

FOSSASIA 2016

Sat, 26 Mar 2016 14:34:25 +0000

Thank you FOSSASIA for inviting me to share my experience with different developers around the globe. It was an exciting place to meet and make new friends, share our experience and learn from other’s experience. I would also like to thank Google and KDE for helping me financially and sending me to this year’s FOSSASIA. The event was very well planned and executed. I hope the participants find all the talks and workshops effective and implement them someday.

All the speakers had something inspirational to the audience in their talk, who seem to be enthusiastic too. I was amazed to see small kids interested in software development. I helped Jigyasa Grover and Mohammad Umair in taking an Android session for school kids which excited me the most to help around small kids getting started with android development.

It was amazing to meet Stephanie and Catherine in person. They shared their GSOC experience and how such a big event is managed and successfully organised from past 12 years continuously.

I also met Mike McQuaid and had a healthy chat about homebrew and GitHub (where he is currently working)

It was a great experience hope to be in touch with all the new developers I met :)

conf.kde.in 2016

Sat, 12 Mar 2016 14:34:25 +0000

This year’s conf.kde.in was organised in Jaipur. I was super excited to be part of KDE India and conf.kde.in for the first time. I was taken back by the preparations that volunteers had done. I really want to take some moment to put forward my thanks to the whole LNMIIT team for such a great welcome and hospitality. Special thanks to “Sagar Chand Agarwal” who made his whole effort in making the conference a success. I took a lot from various speakers, each of them was a pioneer in what they were doing. It was an exceptional experience for me. The best part was the development sprints where we taught students on how they can build their first own Qt applications. Students showed keen interests and asked many questions, we tried our best to help them and solve as many problems as we could in the small time span we were given. Those two days gave me an experience of a lifetime of many speakers. I want to specially mention to “Pradeepto”, at first seeing his reply on emails made me curious to meet him in person. But my perception towards changed him when I met him personally, I got to know that he was the creator of KDE India, Season of KDE and conf.kde.in, and he shared his own experience of his journey in details. We even sat on the grass to listen to his experience and felt it should never end, that was his charisma which I guess attracted almost every person who attended conf.kde.in. After the event, I went on to explore Jaipur with Bruno and Timothee, which gave me insight about how they started their work on GCompris and Krita respectively. It was fascinating to know the motivation that has kept them working for so long years.

At the end, I would say it was awesome to be a part of such a great conference hope to continue contributing and be part of many such conferences in future :)

Pictures to end the blog ;)

Final evaluation

Tue, 18 Aug 2015 14:34:25 +0000

As we all know we have our final evaluation of our GSOC project next week. I have completed my project and would like to display how the integration between Cantor and LabPlot works. For this demonstration I will be using Python3 and numpy.

Let’s first create a Python3 session inside LabPlot.

I will be referring to the script from here. So let’s execute the script and transfer every variable to lists as LabPlot for now supports only lists and tuple data structure. The screenshot bellow shows the final script when written.

Every variable that is either list or tuple is converted to a column inside LabPlot which are child objects of the main CAS Session. These columns can then be used to generate plots as shown below:

Finally, we chose the columns/variables we want to use and plot the graph. The following graph is generated if we plot “t vs s1” and “t vs s2” columns from the above data:

I have created parser, that is used to parse the variables to columns, for every backend, but testing for Sage and R backend is left. I will be testing those two parsers extensively during this week.

Below are some more screenshots showing zoomed in and zoomed out plots.

Moreover, user can now save all the data of the session as well as the plots inside one LabPlot’s .lml file and it can be loaded to be used next time user want. I hope with integration of the two applications, user experience of the users will increase and they will have a richer experience while ploting and using CAS .

GSoC 2015 (Mid Term Evaluation)

Fri, 03 Jul 2015 14:34:25 +0000

The Midterm evaluation week has almost come to an end and midterm evaluation deadline ends today. This post will describe about what all I have achieved in my project “Integration of Cantor with LabPlot” and what I plan to do further.

Below are the screenshots of LabPlot.

As you can see in the above screenshots Cantor’s session is integrated. Variable manager, Print, Print Preview and all other relevant actions in Cantor has also been implemented into LabPlot. With implementation of all these I have successfully achieved my midterm evaluation target. I was working on improvising my code implemented so far and implementing my mentor’s suggestion to code. I will now move on to extract variables from the cantor’s session so that we can use them to create new plots inside LabPlot. After that I will work on saving Cantor’s data along with LabPlot’s data, so that user could save and load both the worksheets. I have learned a lot during this journey so far. I learn how duscussion plays an important role in the development of a software. I learn about some of the best practices that should be followed and their importance in real life code. That is not all! My upcoming weeks will see more of coding and hence I am prepared to learn more during that time.