When i look at my progress as a software developer who came into this world from a different engineering background, i clearly see few major stages of learning process and i think many of the engineers who came into software from a background other than CS would agree to it... First i had to learn the concepts of programming language... And mind it, had it been procedural languages of earlier days, it would have been one story... But with the advent of OOAD the whole ballgame of learning programming concept had changed... To work independently i not only had to master the syntactical parts of the languages like C++ or Java but also the concepts of OOPS... Here i had to learn from varied parts of a language starting from VTBLE in C++ to inheritance concepts to alignment in memory of an user defined data type and data structure... And believe me its not easy... It took years to master only this part of the learning curve...Then i had to master subjects like UML and Design Pattern to wisely apply my OOAD knowledge for a practical purpose... And again it took a lot of perseverance to do this... Then i had to know about a OOPS framework and figure out where my OOPS knowledge may fit... For me i started my software journey with the Microsoft MFC/VC++ framework...
However this was just the beginning... After learning the nitty gritty of a language, and by nitty gritty i mean even understanding the compilation error, i had to delve into the world of OS and had to pick up the real CS areas like threads, process, multithreaded application, thread synchronization, scheduler and other concepts... And then i became a little useful in the real software developement...
In the meantime, i wanted to foray into a specific domain like mobile domain, and for that i had to constantly search for what is going on in this space... it was not easy to mix up the telecommunication part of the industry with the software that drives that telecommunication world... i never got a chance to go into the core telecom side of the software and their management but i was lucky enough to be a part of the mobile apps world... And even after coming into the mobile world i had to switch over from a Symbian C++ engineer to an Android/Java engineer... And to learn Android at a basic minimal level i had to burn lot of midnight oil in front of Google to decipher the Android framework code like the Android service internals, Android mediaplayer framework or Android Asynctask nitty gritty etc... It took lot of patience and perseverance to master only a little portion of Android...
I have learnt lot of things by standing on the shoulders of Giant like Google. I always suggest my students that to learn real software, they need to dig into the source code of different frameworks. For example, when i started going through the GoF design pattern book back in 2004, i used to delve into the MFC/VC++ source code to figure out the implementation of different design pattern. Thus i learnt how the Doc-View architecture resembles to the Observer Pattern. Or when i dug into the MFC's command routing source code and initially thought that its an example of Command Pattern because the term Command Routing, but then i figured out that its just an example of Chain of Responsibility design pattern. And believe me, when a software engineer understands these design sides of a framework, his joy is no less than that of a scientist.
Let me give you another example. When i started my journey in Symbian C++/S60 framework, all the Symbian books used to describe about the Active Object. But i never got the idea exactly how this whole thing is working although i was aware of Asynchronous ans Synchronous programming concept. Then in USA at NOKIA i got some time to dig into this part of Symbian S60 framework and figured out that in the Client-Server architecture, when the server finishes the task handled over to it by the client, it notifies the client through an InterProcess communication message passing system(actually the function is called Complete and this belongs to the RMessage2 class of Symbian/S60 framework), the whole thing became absolutely clear to me.
And in recent times (not exactly recent but in 2013) when i dug into the source code of Asynctask of Android and analysed the SerialExecutor function, the whole concept of Asynctask became clear to me. I am saying all these things to convey the message that to learn real software one has to look into the source code and analyze it.
And these days when i am upskilling myself with the knowledge of Aritificial Intelligence, Machine Learning, Computer Vision and related subjects, i am again going through the old engineering concepts of mathematics and statistics like Matrices, bell curve, Gaussian noise, Gaussian filter, Gradient Descent Algorithm and so many things... However this is just the tip of the iceberg vis-a-vis Computer Vision and AI... i know i have to go a long way to be useful in this world of computation... But frankly speaking the real joy of learning has started... i no longer worry about the programming syntax, but the real picture of the end technology...
However this was just the beginning... After learning the nitty gritty of a language, and by nitty gritty i mean even understanding the compilation error, i had to delve into the world of OS and had to pick up the real CS areas like threads, process, multithreaded application, thread synchronization, scheduler and other concepts... And then i became a little useful in the real software developement...
In the meantime, i wanted to foray into a specific domain like mobile domain, and for that i had to constantly search for what is going on in this space... it was not easy to mix up the telecommunication part of the industry with the software that drives that telecommunication world... i never got a chance to go into the core telecom side of the software and their management but i was lucky enough to be a part of the mobile apps world... And even after coming into the mobile world i had to switch over from a Symbian C++ engineer to an Android/Java engineer... And to learn Android at a basic minimal level i had to burn lot of midnight oil in front of Google to decipher the Android framework code like the Android service internals, Android mediaplayer framework or Android Asynctask nitty gritty etc... It took lot of patience and perseverance to master only a little portion of Android...
I have learnt lot of things by standing on the shoulders of Giant like Google. I always suggest my students that to learn real software, they need to dig into the source code of different frameworks. For example, when i started going through the GoF design pattern book back in 2004, i used to delve into the MFC/VC++ source code to figure out the implementation of different design pattern. Thus i learnt how the Doc-View architecture resembles to the Observer Pattern. Or when i dug into the MFC's command routing source code and initially thought that its an example of Command Pattern because the term Command Routing, but then i figured out that its just an example of Chain of Responsibility design pattern. And believe me, when a software engineer understands these design sides of a framework, his joy is no less than that of a scientist.
Let me give you another example. When i started my journey in Symbian C++/S60 framework, all the Symbian books used to describe about the Active Object. But i never got the idea exactly how this whole thing is working although i was aware of Asynchronous ans Synchronous programming concept. Then in USA at NOKIA i got some time to dig into this part of Symbian S60 framework and figured out that in the Client-Server architecture, when the server finishes the task handled over to it by the client, it notifies the client through an InterProcess communication message passing system(actually the function is called Complete and this belongs to the RMessage2 class of Symbian/S60 framework), the whole thing became absolutely clear to me.
And in recent times (not exactly recent but in 2013) when i dug into the source code of Asynctask of Android and analysed the SerialExecutor function, the whole concept of Asynctask became clear to me. I am saying all these things to convey the message that to learn real software one has to look into the source code and analyze it.
And these days when i am upskilling myself with the knowledge of Aritificial Intelligence, Machine Learning, Computer Vision and related subjects, i am again going through the old engineering concepts of mathematics and statistics like Matrices, bell curve, Gaussian noise, Gaussian filter, Gradient Descent Algorithm and so many things... However this is just the tip of the iceberg vis-a-vis Computer Vision and AI... i know i have to go a long way to be useful in this world of computation... But frankly speaking the real joy of learning has started... i no longer worry about the programming syntax, but the real picture of the end technology...
