THE OPEN PACKAGE REVOLUTION

Before starting Reading, please remember this is being written by a X-generation guy. Someone who has been into the software development world for at least 25/30 years and has seen lots of stuff going on. And I mean having a hard time on the good old times. Well, I confess they were more “old than good”.

And I mean it by thinking about the hectic machine learning world, all those packages and abbreviations and new languages that come to our minds. Yo”u open your favorite software development environment or tool, R Studio, Anaconda, sPyder, on your Linux distribution that you have just downloaded and installed, using the Hadoop environment tools, MySQL or NoSQL databases, and put your stream of ideas and concepts down to code. It all happens too fast.

So fast that you barely think about money : you have started working and produced lots of code lines, using tools that you didn’t have to pay a single nickel for them, and they are all completely legal. Yes. “Why do you mean by ‘all legal’”? Millions of years ago people used to get software they hadn’t paid for, because just everything used to have prices, and they were all very, very high. Student edition? Home edition? No, there used to be just a single edition for everything. If you were a corporation with lots of money, you could pay for that. If you were a loner, a voice in the desert working on bits late in the evening on your sleeping room’s desktop PC XT, you had to get your tools the best way you could. Or use poor alternative versions built by fighters like you.

This brings us to what we have today. If I don’t like an operating system, I will look for a better one. A new programming language or environment. A database or file management system. We can dive into internet, through Google or not – there are even lots of search engines for coders. We have github. Stack overflow.

And, above all, we have them, the wonders of today. The packages. Paramount components of whatever you build in Data Science, Machine Learning, Deep Learning, Artificial Intelligence, Data Analysis.

Each package encloses lots of hard work. It contains pure science. Hard-working nerds who understood some math, concepts, and designed them so it turned into code. Then they published it and, that’s it. A package available, for free, to whoever wants to use it.

If you produced something like that on the Mesozoic age, you would SELL it, send the object code (NEVER the source) and charge a fat fee for it. Nothing more different from today’s packages, where every single line of code is provided – which makes each package a living and evolving entity. There’s not another reason for the amazing development speed : everyone, anyone with minimum coding skills can contribute, not only those weirdos who got locked into their rooms with nothing but a mug of coffee, donuts and a desktop. Today you see that guy with a heavy metal t-shirt, headphones (probably listening to Linkin Park, Slipknot or Avenged Sevenfold) on a shopping mall, in front of Starbucks, working on the last Keras or TensorFlow version.

It’s not possible at all missing the old times. Life used to be too hard, then, contrary to what our grandparents used to say regarding their old times. Things today are fast, and the most important of all, knowledge is open and being shared from all sorts of channels. You have video tutorials, Facebook + Twitter + Instagram coding pages, besides resources already referred above. You start coding, you check your colleagues on instant messengers, you go to coding communities (mostly open). And, yes, you can earn a living with that. You can build great products, sell them, and get good profit, even including the source code as part of the product. These are open times. This is the real evolution.