Rahul gives unsolicited advice

 I just finished listening to this amazing special series produced by the folks at HBR IdeaCast podcast on "The Rise and Fall of Carlos Ghosn". If you didn't know, Carlos Ghosn was the CEO of three automotive giants at one time - Renault + Nissan + Mitsubishi. This was before he was jailed for hiding his income from the Japanese authorities. He subsequently escaped from Japan in a crate meant for musical equipment in a private plane to a country which doesn't have an extradition treaty with Japan. Here are the 4 parts to the special series - Part 1 , Part 2 , Part 3 and Part 4 . Fun-fact : I met Carlos Ghosn. Seriously. Back in college, I was part of the Extra Mural Lecture team at my alma mater, Indian Institute of Technology Madras (IIT Madras) and one of the speakers who we invited was Mr. Carlos Ghosn. Here's a blogpost by students at the Department of Management Studies (DoMS) at IIT Madras from that day. I can't find better resources at the moment.

For those of you who don't already know, I'm a big fan of the podcast Planet Money by NPR. I haven't listened to them in a while now because I wasn't making time in my daily routine - until a couple of days back. And today, I came across this gem of an episode. Oh, by the way. I've also become a big fan of the US Television Show Jeopardy over the past year. I hope that helps you understand why I loved this episode.

I've been thinking about Product Management in FOSS (Free and Open Source Software) for a while now and I just spent most of my morning reading up on it. Here are a few good articles I came across :  Why Product Management is Open Source’s Fatal Flaw - An interesting article (from 2008!) that used an example from the Pidgin FOSS IM Client (I've heard about this in college) to discuss where Product Managers fit into FOSS projects and why there is a need for more Product Management in FOSS. No Such a Thing as Open Source Product Manager - Here's an article from 2005 that questions whether or not Product Managers add value to the underlying product. The anecdote about a FOSS competitor beating out software from Microsoft (where Product/Project Managers are involved) feels like an outlier rather than the norm. Rules for product managers at open source companies - Here is a more recent article from 2020 that tries to layout exactly where Product Management can help an open s...

At work, I am part of a team which maintains the Enthought Tool Suite (ETS). ETS includes a number of open source Python packages which are the foundation of the desktop applications we build at work. One of those packages is enable , a 2D drawing library. Yesterday, I came across this issue on enable regarding a python module ( kiva.fonttools.sstruct ) which wasn't used anywhere in the enable codebase at the moment. "... at the moment" is important here because the enable GitHub repository has been evolving for over 10 years now and we don't know if at some point in the past, kiva.fonttools.sstruct was used somewhere in enable. Before we decide on removing the python module from the codebase, we need to understand when it was added, who was using it and how and when the dependency on enable.fonttools.struct got removed from enable. Only then can we confidently remove the module without worrying about unexpected repurcussions. So, I started digging. The first thing ...

It's a good thing I didn't just blindly close the hundreds of tabs that I have open because I just read this beautiful, amazing and thought-provoking piece by Esquire (which has some amazing long reads BTW) on a wall street hedge fund betting that american shopping malls were going towards a crash . It was amazing and fun to read but at the same time there's a deeper and sadder issue of the impact this has on the people. I'm not talking about the people impact of the bet itself but the fact that a bet can be made in the first place. Two paragraphs nicely sum up the issue. “A well-documented historical pattern is that fraud thrives in boom periods and is revealed in busts,” the university researchers wrote, adding that end investors were unaware of this hidden risk, a deception akin to buying a Ferrari secretly outfitted with a rusted-out Kia engine. It could be argued that CMBS had been a magic trick all along, with big banks one step ahead, luring investors to pick a c...

A month or so back I remember coming across the Big Bang AR app by CERN . I ended up searching for and installing a bunch (A BUNCH) of different apps made by CERN and NASA. I am not too happy with the HiLumi3D app ( on the app store and on the play store ). It was slightly buggy and not exactly information but the Big Bang AR app blew my mind. It was made in collaboration with Google so understandbly, the app itself works and looks much better than the HiLumi3D app. I also loved the content itself in the app, along with the small amount of additional reading. I'm in awe of what people are doing with AR for science education and communication. One day, I hope to be able to join the party.

In the previous blogpost, I linked to the code that Guido used to visualize the PEG parser now used in CPython 3.9. The code lives in the GitHub repository - pegen . I use Windows as my daily driver and right off the bat, this is a problem because the visualizer uses the curses standard library package - which only works on Linux/Mac OS by default. There are apparently Python packages I can install to get it to work on Windows but I didn't try that approach. A brief search on PyPI led me to the windows-curses package, which might be the solution. I switched to the Ubuntu WSL I have setup on my Windows machine. Once on Ubuntu, I needed to install Python. Specifically, I wanted to install Python 3.9. For that, I had to add the deadsnakes ppa to the apt repository list. Once I did that, I needed to install Python3.9-dev and Python3.9-venv. venv to be able to create a virtual environment and install the necessary packages. Python3.9-dev because one of the requirements needed the de...

For reasons I don't fully understand, I started reading the PEP 617 -- New PEG parser for CPython today morning. Note that the PEP is accepted and the latest Python release (Python v 3.9.0) uses the new PEG parser by default and in Python 3.10, the old LL(1) parser will be removed. I remember looking at the Python 2.7 grammer when I started using CPython professionally four years back. None of it made sense whatsoever to me so I mostly ignored the Language Reference section of the Python docs. To cut the story short, see the difference in CPython grammer for yourselves - grammer for version 3.8 and grammer for version 3.9 . I ended up searching for, and coming across the above linked talk by Guido because I wanted to see if anyone had already put together a talk around the new PEG based parser for CPython. I also came across this interesting looking series of blogposts by Guido regarding the work he did to implement the PEG parser for CPython. I say interesting looking, not in...

I just read this profile of Matt Levine on the NYTimes . You need Matt Levine in your daily life . That's all really.

I keep rewatching this talk by Randall Koutnik on Developer Career Path every so often, just to reevaluate my own developer career. I think the "Solution Implementer", "Problem Solver", "Problem Finder" career path that Randall talks about makes a lot of sense, especially in the career path I am currently on. A highlight of his talk is the discussion around shapes and sizes of problems and that the developer career path should ideally expose the developer to a large number of shapes and progressively increasing sizes of problems. Thinking about my own career path in terms of shapes and sizes of problems makes me realize that there are still a lot a lot of shapes that I haven't worked with. I think I'm a "Problem Solver" at this point, while also doing a bunch of "Solution Implementer" work but I'm far from being a good "Problem Finder".