programming.log

Lucid Dream Journal Project

2008-02-24T00:52:00.008-05:00

In my free time, which seems to be dwindling lately, I have been working on a cross platform dream journal application to aid people with lucid dreaming. If you don't know what lucid dreaming is, head over to ld4all or dreamviews and get yourself caught up. You'll be glad you did. This post is not about the application specifically, as only the core is finished and the user interface is just starting to shape up. The details are for another time.
I'm going to talk about the general structure of the application in case anyone is interested.

I have taken a route that most, if not all software design classes (at least the ones I am familiar with) stress, similar to many other projects. The core functionality of the application is built as a library, and is as coupled as loosely as possible to the user interface. This is of course, partially because I am as of yet undecided of which toolkit to use to implement the interface. However let's pretend that it is all for the sake of good software design ;). I've narrowed it down to the FOX Toolkit, or Ultimate++. Both are well established, free and open source. If anyone has any thoughts of why one is preferable, feel free to share your thoughts.

Distributed Code Management with Mercurial

2008-02-20T00:27:00.006-05:00

I've been pretty busy the last few weeks and it will probably remain that way for the foreseeable future. Unfortunately that means my more detailed reviews of different revision control systems will have to be delayed. In the meantime for those interested I'll explain the basics of a workflow I've adopted on a few projects using Mercurial. Please keep in mind that you can use this method with most, if not all, of the current distributed revision control tools.

Since I have more experience with revision control tools than others in my team ("more" being "some"), I decided to set up a repository to help us manage our code and documentation. However, I'm not a fan of leaving my computer on all the time due to little details such as...the electric bill. The workflow we have been using is pretty simple. There is a master repository on my desktop, which is turned on in the evenings. When people make a change to the repository, they create a "bundle" file, and e-mail it to me with a certain subject.
The command in Mercurial is:

hg bundle --base [base revision] -r [end revision] [bundle file]

The desktop monitors the e-mail account for messages with the appropriate subject from a list of approved senders, and checks for the attached bundle file. If it exists, it is downloaded and applied to the repository. Merge conflicts are handled manually as usual. Once the new changes are successfully merged, a new bundle file is created and e-mailed to everyone so they have an up-to-date version of the repository.

While this may sound overly complicated, from a usage standpoint it is pretty simple and works well in practice for small groups of people. There is also the added bonus of enabling people to exchange changes with each other directly without going through the main repository, thus improving the quality of the main code. I agree that giving everyone SSH access to push/pull changes would simplify the process, however that is not an option for us, plus it would be boring ;).

Distributed Version Control Comparison

2008-01-27T03:12:00.001-05:00

After reading about several new distributed revision control tools over the past several months, I’ve been testing some out (actually, most of them), and a few have caught my interest: Mercurial, Bazaar, Git, and Darcs.

I seem to have settled on Mercurial for now, as I’ve been using it for both school and personal projects over the past few weeks. However, all of these tools have facilities to make conversions between them possible, and even easy (NO WAY!), in some cases. In addition, they also all have support for reading from Subversion repositories. If I happen to change my mind next week I can convert my repositories to another format and use another VCS, which is pretty cool. I’ve broken down the main differences between these tools below, however don’t worry, I will post a more detailed review sometime next week.

Mercurial: Written in Python and works equally well on both Windows and Linux (apparently under other systems as well), and is used by projects such as NetBeans, Mozilla and OpenSolaris. The UI is well designed and the commands are easy to get used too, especially if you have used Subversion or CVS (if you are still using CVS for the LOVE OF GOD upgrade to Subversion!) before. The output is completely customizable, which makes it perfect for scripting, and it includes all the repository hooks bells and whistles. My personal favorite.

Bazaar: Smaller market share than Mercurial or Git, but I have very few complaints. It is used by Canonical Ltd. to manage the Ubuntu linux distribution, and works great under both Windows and Linux. It does seem a tad less responsive, but not by much, and the performance improves with every release. Bazaar also has direct support for the “centralized” version control model (a la Subversion/CVS/Perforce, etc), which saves you an extra command if you use that model for a project. (In Mercurial/Git/Darcs you need to do an extra “push/pull” to communicate with the “central” repository). The conversion tools are also excellent.

Git: Originally made of up several C programs stuck together with bash and Perl scripts, this tool adheres to the UNIX philosophy of using several small tools, with each one doing one task really well. Git has since evolved to be a great VCS, and is most notably used to manage the Linux kernel. The main shortcoming is its lack of Windows support. It will run with Cygwin, but that is not an option for the average user. If everyone involved in the project uses Linux, Git is definitely worth a look. There is also a great interface for communicating with Subversion repositories, git-svn.

Darcs: Written in Haskell, and uses a “patch algebra” for changesets to make merging easier. Performance is on par with the rest, although I’ve heard of a few cases where a merge can take hours where it should have taken seconds. However, Darcs is still in its early stages and has some interesting ideas. Like Git, it does not work well in Windows, but for some people that is not an issue. I wouldn’t trust it with a corporate project, but definitely worth a look.

While you’re waiting for the more in-depth reviews, check out these other comparisons:

http://www.dribin.org/dave/blog/archives/2007/12/28/dvcs/

http://www.jukie.net/~bart/blog/git-vs-hg

Enjoy!

I'm back!

2008-01-16T23:29:00.000-05:00

So, as you can see from the title, I'm back. Hopefully with regular updates for a while now. I've got some side projects in the initial planning stages, and some interesting classes this semester. Stay tuned. Or Don't. Your Choice.

Projects that took off (in C++ and Haskell)

2007-07-22T02:04:00.000-04:00

I originally was going to work on three projects, but I've only been working on two, because...it's summer ;)

Anyway, here is what is actually going on. My development rate increased significantly when I started using a revision control system (Mercurial for now):

Name: Not Yet
Description: A journaling program that specializes in recording Lucid Dreams (look it up if you don't know what those are), written in C++ using the FOX GUI toolkit (which I very highly recommend). I'd say this project is about 30% complete, but I'm going to have more time to work on it n the next month or so before school starts, so hopefully I'll get something working.

Name: Doesn't need one
Description: An XML parser written in Haskell. Not much to describe here, except that Haskell is awesome for writing parsers. More on this later.

Summer! (again)

2007-06-04T21:23:00.000-04:00

So it's summer, which is cool. I have a new list of things to accomplish, which will hopefully go better than last year. I'm sure it will though, as I've already started on a few of these. Anyway, stay tuned for any stuff I end up posting regarding any of these projects. I mean, I am involved in them, so they must be awesome ;).

Name: Doesn't have One Yet
Language: C++, using the FOX GUI toolkit
Description: Journal/Organizer software, will store files in an XML format.

Name: N/A
Language: Haskell
Description: Write an XML (or possibly other type) of parser in Haskell to learn how it works. So far it's been awesome. Defining and using a binary search tree in 3 lines is pretty sweet.

Name: Doesn't have One Yet
Language: Javascript, PHP
Description: Web-based organization app (because of course there aren't enough of these yet). Just a basic idea at this point.

Looking back at this, I need to come up with some names for stuff. I'll work on it. Stay tuned for developments.

OpenGL Programming

2007-01-23T23:41:00.000-05:00

So I'm currently taking a graphics programming course, and its pretty intense even though we just started. Apparently it really helps to learn Linear Algebra even though it's not technically a prerequisite. So I'm giving myself a crash course in Matrix operations. WOOOHOOO.

Anyway, you should check out OpenGL. It's not object oriented, but it is awesome. Their site has some really good documentation as well. A helpful library to have when using OpenGL is glut. If you use Windows, here is a pre-compiled DLL of glut for you. Otherwise you probably know how to compile it yourself, or can at least find instructions on how to do so much better than I could write. If you use Visual Studio (which seems to work pretty well for some strange reason), you can check out this site which includes instructions on setting up OpenGL and glut in Visual Studio 2003, but it can be adapted to work with 2005 quite easily.

So when you have some free time, try some OpenGL programming. It's easier than you think and a lot of fun! I'll be posting some of my projects up here in the near future.

"Prey" - An Awesome Game

2007-01-04T23:44:00.001-05:00

If you haven't played Prey, you should. It's seriously one of the best games I've ever played. It's got a lot of stuff that makes it unique such as wallwalking, gravity switching, "spirit walking" and "death walking". I wouldn't do these features justice by talking about them here, so why not check out the official website. It has screenshots and in-depth descriptions of all these things.

The game also as an awesome story, which I like. Multiplayer isn't everything. So if you have some time, go check it out.

OpenDarwin Project Closing

2006-07-26T23:16:00.000-04:00

Yes, OpenDarwin is shutting down. This doesn't surprise me too much, as Apple has been known to have problems divulging any type of information. Apparently this caused a problem when they tried to work with open source developers.

Fortunately, everyone is trying to move the projects to new locations, so hopefully no one will lose anything. Check out the announcement here.

Happy Day Out of Time!

2006-07-26T01:06:00.000-04:00

Happy Day Out of Time! http://www.dreamofpeace.net/index.php
Check out the link to find out more.

In other news...
With the release of Slackware 11 coming soon (who really knows? ;), I've decided to delay my review of Slackware until 11 comes out. As I don't have a broadband connection I cannot keep up with current, so I cannot comment on new packages.

I may do a general Slackware review however. Stay tuned.

Linux Symmetric Multiprocessing

2006-06-21T23:16:00.000-04:00

If you, like me, happen to have more than one (or a dual-core) processor, consider yourself a priveleged and distinguished person. If those multiple CPU's, like mine, happen to be Celeron 400Mhz or less, then ignore this article and set your motherboard on fire. Seriously.

If you're still reading this because:
a) You think your CPU's are totally badass, or
b) You have 400 Mhz Celerons but want them to be useful despite my warnings,
then be my guest.

To take advantage of the Symmetric Multiprocessing features in the Linux Kernel (much improved in the 2.4 and 2.6 series), you'll have to recompile your kernel. If this scares you, don't worry, several guides exist online to help you with this. Try a search on Google. Just remember to always keep a backup kernel that you can boot into so you can mess around until you compile a working kernel.

Anyway, when you recompile the kernel, enable Symmetric Multiprocessing and the features included with it. Now once you have support for multiple CPU's up and running compiling will go a lot faster. You can check your support by typing dmesg at a terminal and reading the output, making sure it recognized all your CPU's starting with CPU#0.

One really cool thing you can do with multiple CPU's when compiling programs is run make with the arguments -jX, where X is number of CPU's + 1. For example make -j3 for a 2 CPU system. Make will run 3 processes at once, taking advantage of all the CPU's. This actually increases performance in single CPU machines too, run make -j2. Enjoy!

Rebirth

2006-06-06T23:41:00.000-04:00

I've decided to start writing here on a regular basis again. This summer should be full of experiences, as I'm setting up a server so I can mess around with mySQL, Apache, PHP, and IRCD. Should be fun ;)

I'm also planning on writing some software, such as a journal program and possibly an IRC client (insert "ooohhhh, aaahhhh" here). So stay tuned for some really cool and hopefully helpful programming/linux tips.

JAVA Scanner's nextLine()

2006-02-09T16:32:00.000-05:00

This is a simple concept, but some people may not be familiar with it since it's not used that often. When using JAVA's Scanner class to do standard input/output from the keyboard, make sure you use the nextLine() method of Scanner instead of next(), when you want the program to do something when the user just hits enter at the console without typing anything.

Of course, this is only true if you're checking that the input is blank. If you use next(), you have to match the input string with the newline character. For example, to throw an exception (or do whatever you want), when the user hits enter without entering any input:

Scanner stdin = new Scanner(System.in);

String inputString = stdin.nextLine();
if(inputString.equals(""))
{
//do whatever, mabye throw an exception
//or display a message
}

That little code snippet will work, doing whatever is in the brackets of the if statement when the user hits enter without typing anything. You can also use inputString == "", inputString.charAt(0) == '', inputString.length == 0, etc to test the content of the string.

This was a real problem for me, I spent 20 extra minutes on a lab because of it. If you want to implement the input this way, be sure to use nextLine(), not next().
Enjoy, and stay tuned for more programming tips!

Python. It's good, real good.

2005-11-09T22:27:00.000-05:00

At first glance Python may resemble an archaic language, but it’s not. Those of you who think that, well, I respect your opinion, but you’re wrong. Python is extremely powerful, especially for a scripting language, and rivals the functionality of classic compiled languages such as C, C++ and Java. To get a basic introduction to Python check out the links at the official Python site:

Http://www.python.org

Python contains some cool basic data structures that make coding with less actual code possible and coding in general much easier. The two data structures I’ll talk about are lists, and dictionaries. If you’ve programmed before, think of a list as an array, because it can function like an array, but it so much more. A list is a data structure that can hold a variety of data types, all in one list, which is pretty badass. (Come on, admit it.) For example: list1 = [“ice”, 9, I, “Kids”, “Joe”] would be a valid declaration. Items in lists are referenced by index (0,1,2,...etc).

A dictionary contains items bound to keys. For example, dict1 = {“Cheese” : “good”, “Bob” : 92}
declares a dictionary. To references the object “good”, you would use dict1[“Cheese”]. Almost any data type can be an item in a dictionary, (integers, string, lists, other dictionaries, tuples, etc), but the types of items that can be keys are limited. It's too complicated to explain here, and you'll probably only want to use basic data types as keys anyway.

Another Note: Python can be extremely fast because many of the modules in the Python library are coded in C. In fact, it's really easy to write your own modules in C or C++ and use them in Python (if you write C or C++ of course).

Stay tuned for more tips on Python and other languages. Code On.

Transformation

2005-10-11T00:34:00.000-04:00

I’m in the process of transforming this into a source of interesting tech news, so just chill out while it happens. I have classes and mid-terms to worry about, so it might take a week or so. Don’t get all excited and make porn movies.

The Cockfight

2005-02-16T21:38:00.000-05:00

Today was normal. Kind of.
Not that you care about my day. I’m sure you don’t, just like I don’t care about yours. But I bet you’d care about it if it were funny, right? Don't lie to me.

Anyway, I saw a cockfight at this gas station on the way home. Yes, my car uses gas, just like everyone else’s. I was standing there, pumping gas, trying to look cool like everyone does at gas stations (don’t deny it), and I saw a bunch of chickens come running out from behind the gas station. Two of the males started pecking at each other.

So let’s review: I’m trying to look cool pumping gas while a bunch of chickens peck the crap out of each other. Some other guy getting gas was watching too. We should've made some bets. Anyway, I got gas and a free cockfight at no extra charge. Life rules.