So Play We All News

Ledger

Net Profit: $158.12

Hourly Earnings: $3.24

$19.12 - 2011-09-06 - Rule Violation Payment (Peter)
$19.92 - 2011-08-31 - Rule Violation Payment (Luke)
$28.83 - 2011-08-30 - Rule Violation Payment (Peter)
$10.00 - 2011-08-14 - Rule Violation Payment (Luke)
$30.00 - 2011-08-05 - Rule Violation Payment (Peter)
$10.00 - 2011-07-18 - Rule Violation Payment (Peter)
$10.20 - 2011-07-16 - Rule Violation Payment (Luke)
$20.30 - 2011-07-14 - Rule Violation Payment (Luke)
$29.91 - 2011-06-11 - Rule Violation Payment (Luke)
-$10.16 - 2011-05-18 - Domain Name

Deity of Development

published by JGadrow on Thu, 06/23/2011 - 11:52

This week, we had some changes in how the rules work. Instead of having a number of hours each week that operate as both a minimum and maximum number of hours available to us each week, we now agree upon the minimum number of hours and we're allowed to spend up to three times that amount at our sole discretion. Basically, this fixes a problem that we had in that if one person (maybe like myself) is behind in terms of feature development because of the way their system works, it doesn't necessarily guarantee that they will always be behind. Basically, it allows the contest to remain a bit more fair for all involved and means we have the potential to make some serious progress on our games at any given time. As a prime example of this, our minimum this week was only two hours. However, I spent my full six hours this week; and am I ever glad that I did! But before I talk about that, I have something interesting that I want to share.

Deities

Some of you may have noticed that there is a new tab at the top of this page: Deities. These are the initial Gods and Goddesses that your armies will champion in Allabrilyn. I decided to provide a brief "in-character" glimpse for each deity as the first paragraph. I'll have to talk a bit more about Makari Verslund one of these days (he's a notable entertainer / storyteller in the world), but I'm quite satisfied with what I produced given a very short amount of time to focus on generating the content. We had decided at an earlier point that the hours each week only encompass code work and that the fluffy stuff of the game is stuff that we should be working on in addition to our normal hours. As such, expect me to churn out more of this fluffy content at random intervals!

I've allowed commenting on the deity pages so please tell me what you think! Which brings me to another thing I wanted to talk about...

Why I Require You To Have An Account

Yes, I know it's a hassle. I do have plans in the future to make this easier on you but, for now, I'm mainly using it as rudimentary means of defense against spam. I can tell you that when the game launches, I will be allowing you to create a user account for the game directly from your blog account. This site will eventually become the community site portion of the finished game and anyone creating a game account will have a corresponding account created for this site. I'm also thinking I'm going to allow people to link their accounts with a facebook account to further reduce the number of accounts you have to keep track of. However, I dislike the fact that facebook can arbitrarily decide to remove your account and then you'd be screwed here... so, the Facebook integration (and anything else that you might suggest to me, if it makes your life easier, I'll at least consider it) will have to wait until I'm sure I can provide a nice solution for those who may end up without an account for reasons outside of their control (or if they decide to delete their facebook account).

I apologize for the inconvenience, but despite what people may be saying about me, I actively work to limit my scope so that only things that matter are budgeted for at this time. Of course, sometimes things happen that you didn't budget for... and that brings me to:

Why Top-Down Development Rules

Well, I didn't make as much progress this week towards becoming playable as I wanted to this week. Don't get me wrong, there's tons of progress that's been made (all of which you're privy to, more on that later), it's just not the progress that I wanted to make.

It started innocently enough. I decided that I wanted to take a few quick minutes to generate my API's documentation so that anyone interested could follow along and see how the framework is evolving. I generated a couple of different styles and started to click through it to make sure the documentation was more-or-less complete. The first thing I clicked? The Todo list, of course! I noticed that, essentially, I had two things to work on: BaseCache and CommandRouter. The CommandRouter is where I had intended to focus my efforts this week so I could make progress towards being able to output HTML again (you know, through the system, not a hacky "echo" solution like I used to have up). Well, I had a nagging feeling that I should dig into the BaseCache and make it functional... mostly because it's the very first high-level object that's used in the framework. Illogically, I had left it as something I had intended to work on towards the end of the project. Apparently, my years of experience kept telling me that was a "bad idea" and they were right!

So, I decided to assuage the voices in my head and start working on the BaseCache. Honestly, it's just flat-file storage with md5 identifiers. I really didn't think it would take me more than 30 minutes, all this is meant to do is allow the framework to skip the initialization routines for the very high-level objects that don't really change very often between requests. Prime stuff for caching. Well, after my first pass of implementation (which was about what I had expected), I was testing and noticed something: The Application was going to be stored one time for every single url combination. Pretty ineffective caching if you ask me! The reason this was happening was because the same Application can house multiple site installations. That's right, once I'm finished with this sucker, I'll be able to have several games all running side-by-side with the same code base (meaning when I upgrade one, I'm simultaneously upgrading them all). Unfortunately, there was no real way to determine which site was being run at the time that the Application was being initialized. The site settings were also the biggest performance gain for caching though. In order to determine what site is running, there's a recursive iteration process that determines from the URL which installed site most closely matches. Right now, it's not a very time-intensive process, but it's a process that will grow longer the more sites are added.

So, I tried a couple different methods to do things so that I could cache the Application without this information and nothing really worked (I also hit a nasty little error that forced me to move my handler method registration to a private method and also include it inside of a __wakeup() implementation since my cached Application wasn't being used to autoload class and interface definitions anymore). Finally, at the end of the night last night, I hit upon the idea that would solve the problem. I had to move the site stuff into its own class which would (during initialization) read all possible settings into a giant class. This class would be injected into the Application and the Application would "select" which site was running when it was executed. I had to write that all down because the girlfriend was getting annoyed that I hadn't spent any time with her.

So, I had to change a lot of my high-level objects around to accommodate this new process. It took 3 hours of my morning (which, if you recall, is half of my maximum time allotment for the week) but is now running like a dream! Now my high-level objects are 100% implemented and working beautifully! Imagine how much I'd have had to change if I'd have kept putting this off until later! It would have been a nightmare! Anything that called Application::GetSite () would have to be corrected!

Anyways, in the last few minutes that I had available to me, I did manage to get my documentation online for you to look over. I didn't get a chance to actually go through it and make sure everything is correct, but it should *mostly* be correct and the source code for every file is also posted in the documentation so if something is wrong, the source is still correct. Effectively, I have open sourced the framework right now (though you have to jump through some hoops to get things in the correct directory structure)... I won't promise you that I'm going to keep it as open source, but you can enjoy it while it is.

Hope you enjoyed this week's update!

Comments

Coach Outlet Online

Wed, 12/21/2011 - 15:40

Main menu

You are here

Search form