Player Hardware and Software

While choosing which player hardware configurations and operating software will work with the game software may seem like shooting ducks in a barrel, it is actually one of the toughest issues to address, due to the speed with which PC technology changes. One almost has to have a crystal ball to know what the standards will be in 24 to 36 months.

Push the Envelope or Design for Today's RAM and Processor Standard Configurations?

There has been constant growth in personal computing power over the years. About every 18 months to 2 years , CPUs take a 50 “100% leap in processing speed and the standard amount of RAM pre-installed in new PCs seems to increase, although standard installed RAM has been slowing down of late and there are some early indications that processing speed is reaching a plateau as well.

In the late months of 2002, the standard configuration for new PCs seemed to be a 1-GHz Pentium processor, 128MB of RAM, and a video accelerator card with 32MB of video RAM. A "buffed out" game machine is a 1.2- to 2-GHz Pentium processor, 512MB to 1GB of RAM, and a video card with 64 “128MB of video RAM. If the trend continues as in the past, today's buffed out machine will become the standard configuration sometime in late 2003 or early 2004.

To date, the habit in the industry has been to push the technical envelope as far as it can be pushed and still have an audience with the computing power to buy and run the game. Up through 2001, almost every new online game required an upgrade of the processor, RAM, video card, or all three, to get the fullest measure of the game. This is the "hard- core " or early adopter model of gaming (see the player market definitions in Chapter 1, "The Market"), appealing to those with the money and inclination to spend on gaming.

So, Who Are You Designing For?

The problem with the early adopter model is that it tends to limit your potential audience to that 10% of the market; most people aren't willing to spend a lot of money upgrading their hardware just to play a game. This is why it is so important to select, early in the process, which market niche(s) you'll be appealing to with your online game. If you're going after the sweet spot of the hard-core market, then you'll definitely want to anticipate what the buffed out PC configuration will be as you enter public Beta testing and take every advantage of the capabilities of those machines.

If you want to spread subscriptions among two or more of the niches , however, you'll need to scale back your aspirations and design either for the buffed out machine of today or, if you hope to have a mass-market product, design for today's standard configuration.

Broadband or Narrowband Access?

Over 90% of the world is still using 56k dial-up for access, and that figure is not likely to change significantly in the next two to three years; why would you limit your market by designing now for broadband access only?

Designing for broadband will likely cost you more in bandwidth leasing after launch than you planned for. If you plan to scale your game to broadband needs, be prepared to open your checkbook a bit wider. Also consider what broadband scaling might do to the RAM and processor speed requirements of your game servers; with all that extra data flowing back and forth, you'll probably want to increase both to prevent as many bottlenecks as possible.

Never Forget: The Client Is in the Hands of the Enemy

There is a phrase used quite a bit in the online game development industry: The client is in the hands of the enemy. The first known public use of it was in 1989 by Kelton Flinn of Kesmai, after a cheat hack of the Air Warrior player client appeared on the GEnie online service.

The phrase is not exactly a secret among developers; we've all heard it multiple times at conventions or seen it used in email or in developer message groups. Before the popularity of Internet auction sites, cheat hacks such as this were done purely for the twisted need to win at all costs, or because players got tired of getting their heads handed to them by other players. These folks were containable for the most part because there were only a few of them, and gamer populations were small compared to today.

Now that use of the Internet has exploded and web auction services such as eBay exist, there is a whole new incentive to cheat. One can make a tidy sum auctioning off items or characters in an online game. Some buffed out characters with lots of loot have sold for $3,000 and more on eBay. One company started by a couple of enterprising young capitalists once claimed to be making $400,000 a year selling UO characters and items. This is not chump change. It has become such a part of the current gaming culture that one of these companies is unabashedly suing Mythic Entertainment over halting their eBay auctions of Dark Age of Camelot items and characters.

The "enemy" phrase was not meant as an insult to players, but as a warning to front-end client developers: Once you distribute the front-end client, the enemy has access to it, not just the good guys. It says, "If you leave important, modifiable information in the client, the modification of which would give an individual an advantage over other players, it will be modified." It is a clarion call to make sure you don't leave important information on the client, to protect the integrity of the game for the 99.9% of your players who will never attempt to hack you.

It's significantly worse in these Internet days, as it is pretty easy to acquire software that can sniff the data packets going in and out of a modem. All it takes is one pretty decent code mechanic to sniff the packets, decrypt them if necessary, write an application to use that information in some way, and voil! ! You have interesting little applications like ShowEQ or UO Extreme that give some players a huge advantage over everyone else.

This kind of thing isn't just a headache for developers. Such activities have a debilitating effect on legitimate players; if others are speed-hacking or exploiting holes and bugs to build characters or monopolize rare or necessary game items, denying others a fair chance to succeed and build their own legends, the honest players tend to get mad, then discouraged, and leave the game for more congenial ”and secure ”playgrounds. This is a hit directly on your pocketbook.

The prudent course of action, then, is to keep as much of the game as possible on the server side of the process and as little as necessary on the client. As we said previously, streamlining your code to lower the bit rate is good. Lowering the bit rate by putting important, modifiable information on a few hundred thousand PCs (remember, whatever dangers it comes with, wide distribution of the client is your goal) means that somewhere, someone is going to "have a little fun" with your code ”then they are going to tell their very best gamer buddies . The next thing you know, you are on the wrong end of an expensive lesson in exponential growth.

You also need to make sure your monitoring and security tools are in place at launch because you can't possibly catch every bug or exploit.

Localization: Talk to Me, Baby!

Developers and publishers are starting to realize that Internet gaming is a truly global market. The example of NCSoft's Lineage: The Bloodpledge in South Korea, which has some 5% of the country's total population playing that one game, opens whole new vistas for gaining new subscribers. It also opens up a whole new set of problems, not the least of which is the localization of language to allow conversations in the local tongue. Thankfully, there is a solution to make such localizations easier: Unicode.

What Unicode does is provide a unique number for every character in virtually every human language used on computers today, as well as some historic and classic languages. These numbers are platform- and program- agnostic , meaning they work on just about every computer and software application imaginable. The Unicode standard currently supports over 94,000 characters, and its implementation is controlled by the non-profit Unicode Consortium. ^[2]

^[2] See the Unicode Consortium's web site at www.unicode.org/ for membership details. Membership is not free, but it is low cost ($2,000 per year or less for most corporations and commercial enterprises ).

The beauty of Unicode is that when it's implemented on a client/server basis, it becomes much easier to implement new languages, especially if you keep as much display text as possible on the server (always a good idea for an online game, as it is much easier to add and modify text on the server than on the player's client, which requires a download patch).

There are other solutions available to use in place of or in conjunction with Unicode, such as the enterprise translation software provided by Systran Software, ^[3] but establishing a facility among your developers with Unicode seems to provide the most cost-effective and useful solution overall.

^[3] See www.systransoft.com/ for information.

Which Languages, Then?

Which languages to support at launch depends on the publisher's willingness to branch out into the territories . We recommend the following as a minimum set to be supported, as it covers the most active online gaming populations in the world (as well as most of the world's population):

• English

• French

• German

• Spanish

• Korean

• Chinese

• Japanese