понедельник, 28 января 2013 г.

Software Bucket

I have recently come across a short blog post devoted to having a personal list of software pieces, that one wants to develop - a software bucket. Of course it was interesting for me to think over the possible contents of my own bucket. The first lines of my list are occupied by the things I am currently working on. The topmost is definitely a simple tool for arranging personal tasks. I know that there are plenty of such applications, nevertheless I have decided to create one more - mainly because I want a tool, which is as convenient for me as possible, and hope that someone else might also find it useful.
I have originally come to the idea of creating such an application almost a year ago, but it was only in the end of October when I have finally made up a deadline for the first version of the tool. Being fascinated by the new Windows 8 OS I have chosen to deliver my app through the Windows Store. The key advantage of this approach is that it will provide me with both the means of distributing my creation and the feedback channel - in case someone downloads the application. Beside that, Windows Store applications development feels quite suitable for me because I have got reasonable experience with .NET development and, particularly, with Windows Presentation Foundation.
Despite having done a significant share of work, I have missed the initial deadline, but feeling ashamed for this I have already set up the new one. I feel determined not break one more promise given to myself. After delivering the first version, on which I am working now, I plan to continue expanding this project and add some more features to help me and  potential users plan activities in even more effective manner. I will struggle to avoid making the app overloaded with details and inconvenient, so that other Windows 8 users will be able to consider it helpful and easy to use.
Another thing in the bucket is my current job project - I am developing a little toolset for generating manufacturing-related reports from the data provided by a complex PDM-system. Although the task may seem not very interesting, I feel quite absorbed with it, because it offers me a great chance to practice TDD and refactoring techniques to build and deliver an efficient and reliable piece of software, which is capable of giving valuable help to our customers. Additionally, I believe that the need to continuously align the thing to fresh portions of requirements will build me up towards multiple challenges that I will face in future.
On the other side, I was stupid enough not to test drive the application from the first steps, so it developed in a way which is far from perfect. However, I feel surprisingly glad about that, because for me this is an opportunity to feel all the pain caused by my own code and design decisions - now I am refactoring the project and supporting it with automated tests. Lots of interesting details of this piece of software and decisions made while working on it, as well as the desire to clean and develop it further, make it an important part of my software bucket.
One more item on the list, which is easy to come by, is my diploma project. This one is going to be a data analysis tool for monitoring students' progress data in my university and making predictions based upon it. I am not sure whether there is a substantial chance for it to be adopted and used by the university staff, nevertheless it provides me with important experience and offers a lot of interesting opportunities. First of all, I would like to explore existing data analysis tools and techniques and to adopt some of these to real data. Besides, while working on the project I have to use an external system, which is aggregating and providing the data consumed by my application. The fact that I have been working with it for a considerable period has taught me to put significant effort into shielding my own creations from the fluctuations in the consumed systems' behavior - positively, that's an important lesson. In this sense this looks a bit like my job project, where my tools retrieve data from an external system. My diploma and job feel even more similar because both have something to do with reporting activities, although the projects focus on different problems. The key aspect of the former is data analysis, while the latter revolves around providing flexible data retrieving and processing tools to bridge the gap between the PDM-system, that our customers use, and the reporting engines. My work on the diploma project, as I see it, will involve facing multiple software development problems, learning new techniques and, hopefully, obtaining valuable results - the project is worth the effort.
The rest of my software bucket looks much less clear. Being fascinated by the world of software development and its numerous applications, I cant say where exactly I want to direct my work. Moreover, I love the process of designing, writing and testing programs so much, that it is very easy for me to build things merely for the sake of building things. I feel that's a problem and I have to fight this passion sometimes, because it prevents me from focusing on the issues that are truly important, and not only is it harmful to me, but it can also become a source of pain for my employers and customers. On the other hand, this same passion for software allows me to make out a couple of things, which are attracting me as a young developer.
Among these things one can find code analysis tools. I have come across multiple mentions of such software recently and it so happened, that the idea to contribute to some open-source project to taste the idea of open-source myself has come to me almost at the same time. That's why I plan to search for a project in Clojure, which will allow me to practice understanding others' code and design, to make contributions to a system that may feel alien to me and to better understand functional programming and Clojure itself. I don't really limit the search to code analysis projects only - I acknowledge these may be absent - but I feel eager to work on a project of this particular kind, as this may help both in understanding the ideas behind code analysis and writing better programs.
All the items listed above are united by the concept of tool - they are designed to assist users with their work and provide them with some solutions to the problems, which they face in some area of their life. At the same time any tool is marked by the lack of autonomy and relies heavily on the signals from its consumer. No hammer drives a nail into wall itself - it always needs a person who will pound it on the nail. Definitely, that's not the only way software can function, and I really wish to participate in building an entity, that is capable of analyzing situation on it's own, making sensible decisions and acting correspondingly. I cant say that I am fully determined to work on, say, self-driving car, but I suppose that the area of autonomous (and autonomously developing) systems - either robotics or less physical applications of AI - may turn out quite suitable for me. Anyway, it won't be easy for me to reject an opportunity in the field.
The key aspect of such a 'self-driving' system is the ability to function in a self-directed manner without a flow of instructions from human operator, who serves as a link between the thing and the world around it. The idea that 'reality' of this world may be varied - in the sense that the world of the thing need not be the same as the world where I write these paragraphs - combined with my affection for video games leads to another piece in my software bucket: creating an AI for a video game is a challenging task. While I can hardly imagine earning my living in the game development industry, I could sooner or later participate in some open-source game project. Although the game development field may seem not serious, I still think that it can sometimes serve as a source of inspiration and insight for other, more serious areas. Furthermore, my minor attempts to create some simplistic computer games showed me that this kind of work is highly paying in emotional sense, because it is always amazing to play with your own creations.
So these projects and ideas constitute my software bucket today, although due to the reasons mentioned above the growth and expansion of the list are barely constrained. This said, I will have to practice saying 'no' a lot in the near future to avoid getting torn apart by my own curiosity and desire to build things. Still, I hope that willingness to say 'yes' will guide me towards doing what I can do best and therefore must do.

3 комментария:

  1. Game's AI is usually very rudimentary and is often intentionally crippled to make it more predictable and manageable — it's clearly inferior to game design and serves it's very limited purpose. It's full of hacks and crutches. I'm not supporting your will to practice AI development in the field of games ;)

    Something interesting could however arise from the placement of the AI into typical "human" situations of playing non-computer games — like the one I mentioned to you: of the game of crocodile.

  2. Yeah, you told me that AI for games isn't very cool stuff usually - I remember that. Still, I suppose the situation will change some day)

    Anyways, I've said that's not a thing that I would like to devote my entire life to. The idea just sounds interesting and, with this attitude, 'traditions' of game design hardly bother me.

    As for the game of crocodile, that fairly suits my idea of 'AI for games'. You could name it either 'computer participating in typically non-computer activity' or you could merely say that that's just another kind of computer game.

    1. That's not a computer game by any means.

      The thing is, usual computer game AI can cheat — it can use information, not available to it as a player, but available to it as a part of computer system.

      In this game, it can't (in case it's implemented in a robot).

      In case the game itself is computerized and held in virtual space via some UI, it fits your definition.