The Angry Dome

I work at a small company. I support about 30-40 people, and keep the servers running.

This is also my first real job, where I've been given not only the freedom to work on my own, but also the respect of higher ups who will at times even follow my recommendations.

This company has some entrenched systems and procedures that can't be uprooted. But in my time here, I've learned of many open source tools that can be used to enhance the way people in a small businesses work.

So, what are the ingredients for having functional, easy-to-use systems that cost $0 in licensing fees? Well, here's a list.

1) Linux or *BSD operating systems. The more I've used both, the more I've come to love them. No CALs, no GUI to drag down your servers, no weird problems that aren't logged and no fighting with the registry to change things. I've come to truly enjoy working with CentOS (which is rebadged RHEL 3), and I fully recommend it for any system that has a static and unchanging role if the core distribution will fill your requirements. However, FreeBSD is quickly winning me over as well, and the cleanness of the system is truly admirable. Debian is also a good choice, if Sarge ever makes it to stable...

2) LDAP. If you have more than one server, you need LDAP. It enables single signon and a nice, single point of administration for all of your user accounts. It also plays nicely with...

3) Samba. Whether you're using Windows, Unix-like, or OS X for your workstations, Samba can provide every system with access to files. While you could use NFS with the latter two options (and even with Windows, if you're up for a little bit of work), Samba just works too well to pass up. You also need Samba if you want to get domain-controller-like options for unifying your Windows workstation logins - NT 4 style domain control can be had easily enough.

4) iptables or ipfw/natd. Your office probably only has a couple of IP addresses and a lot of workstations, so you'll need some sort of NAT router/firewall machine. Right now I use Debian for this because I knew iptables' syntax very well and it included...

5) OpenVPN and OpenVPN GUI. OpenVPN is a project that uses UDP SSL tunnels to send encrypted IP traffic across untrusted networks. Used along with tun/tap in Linux or BSD, you can 1) tunnel between seperate LANs and route traffic between them and 2) bridge individual machines (OS X, Windows, or Unix-like) so that they can appear to connect directly to the LAN. Using this from the start, you could even block all external access to things such as POP 3 or IMAP and instead only allow them from inside the firewall. Each user needs an SSL certificate, and on top of that you can require users to authenticate against PAM - so if you're using LDAP, that means you can get people to login using the same username and password they use for other systems.

6) Courier? Postfix? Sendmail? Exim? You need an MTA, and I'd say this is a matter of personal preference. Pick one. But be sure to use...

7) Clamav. Open source virus scanner, it works really well. It shouldn't be considered your only line of defense if you're using Windows (client machines will still want a commercial product), but it catches most viruses. It works with Samba nicely, as well as your MTA - of course, you also need something to deal with spam, so...

8) Spamassassin. The perl-based spam catcher with some bayesian learning capabilities. The default rules are pretty good, but if you use IMAP and teach your users what to do with spam, you can set cron jobs to feed the learning system to make it even better. Of course, you do need an IMAP server, maybe...

9) Cyrus? UW? Courier? You need an IMAP server. If at all possible, I avoid using POP - you really want to store your messages on the central server.

10) Apache, PHP, and MySQL. You probably want a web presense of some sort (though you could go with a hosting company, if needed). At the very least, you'll probably want some kind of web-enabled box on your network so you can write custom apps to handle tracking internal resources as needed.

11) SNMP/MRTG/ntop/ethereal. You probably want to keep an eye on your network. These tools should be good starting points.

6) Mozilla Firefox. It's open source, it's better than IE, and it doesn't have the same avenues for malware installation that IE does. Regardless of your OS choice, Firefox is a great option.

7) Mozilla Thunderbird. The premier open source email client. It works, and it does a great job.

I haven't been posting much lately. That much is probably clear, though the reason why may be missing.

To be honest, I've remained busy wasting my time on World of Warcraft. I thought I was out, and they pulled me back in - every time I think about cutting my playtime, something new catches my eye, and it's like becoming addicted for the first time all over again. Hmm.

The new PowerBook line was announced, and it's virtually identical to the previous one. I remain very happy with my purchase, though I do plan on replacing this system with a new model once the 'Books are equipped with G5 CPUs and/or OS X 10.4. Just how long I'll be waiting remains to be seen, but I need to start thinking about buying AppleCare before the stock warranty on my system expires this August.

I'm playing with FreeBSD again. Things have mostly remained slow at work lately (with fewer users come fewer chances for failure) and I've had the time to test and research many options to replace our existing systems. I think I'll make a full post on this later, so stay tuned...

Apple today released two products that, while not entirely surprising to people who read the rumor sites, promise to solidify Apple's position as... well... the only innovative PC company.

First you have probably the single coolest looking computer ever made, the Mac Mini. My friends, it's half the size of the Mac Cube and umpteen times more powerful, not to mention that at a price of $500 it's the cheapest Mac... ever?

Next is a USB flash memory based iPod, dubbed the iPod Shuffle. Since it's lacking a display and has been simplified to do just one thing - play through a list of songs, either in order or randomly - this device is, um, way cheap, with the 512 MB version going for only $99. Even the low end portable music market now has an iPod, and those of us who already have big daddy iPods will likely pick up one of these babies for when that "pack of playing cards" sized behemoth just isn't going to cut it (going jogging, anyone?)

So, you have two products that both smash the #1 anti-Apple sentiment: "I'd get an Apple product, but they're just too expensive!"

Now, these products should ROCK, assuming they aren't lemons for some reason. As of right now, I predict that Jobs has sealed Apple's fate as the personal computer company of the next decade. The timing just couldn't be any better, with the Windows XP/IE "one-two" punch leading to vast armies of zombied, spyware-infested machines that run like crap and are just impossible to deal with.

$500 for no viruses, no malware, for something that looks cool, is tiny, undoubtedly uses less energy, and matches perfectly that kickass iPod everybody's been talking about? Um, yeah?

Use the iPod as bait, get them on your line, and reel them in with the Mac Mini. Oh yes, it's going to happen. Jobs is a true genius, even if he himself is not behind the designs - he at least knows how to recognize genius, and he knows how and when to deliver a brilliant product.

My friends, it's the beginning of the end for Windows as the dominant personal computer OS, and I for one couldn't be happier.

Ah, at last, the nightly builds of Firefox have been fixed to allow for middle mouse click in OS X. It's good to see this 2+ year old Mozilla issue finally resolved, and I'm sure my friends on IRC will be glad as well - every few months, I'd point out how pissed off the bug made me.

Thanks go to the guys who put the effort into this, I'm really glad to have my favorite browser working properly in OS X.

I have enabled the authimage mod for WordPress, which requires you to input a sequence of characters that are displayed in an image when you wish to comment on a post.

This effectively prevents spambots from hosing the comment system, and has allowed me to remove moderation and re-enable comments on the weblog. Feel free to leave your mark, and go down in history as having contributed to something that's completely useless.

A testament to open source - if there's a problem, *somebody* will be able to adapt and overcome it without having to wait for a vendor to make things do what you want them to.

It's probably no wonder that it's such a frequent occurance for male persons to play female characters in MMORPGs. There's a sort of "Lara Croft" effect, in that it's nice to have a hot chick on your screen whom you can manipulate at will. Sure, it's only polygons, but they're shapely polygons all the same.

In my EQ days I would play several female characters, in part because that game's male models looked decidedly unimpressive and in part because I, at first at least, assumed it would be a roleplaying game in which people would maintain a degree of seperation between reality and characters (nevermind that the second assumption was utterly incorrect, this was my first MMORPG and such misconceptions should be understandable). However, there are undesirable side effects to playing females - in fact, there are so many that I've resolved as of late to never play a female character except as an "alt" or a diversion, never as a character I wish to play for the long haul.

WoW is probably better than most, due in part to the influx of bnet players who are used to forced crossdressing from Diablo2, and in part to people having a better understanding of demographics as the genre matures. But in EQ, you'd run into problems as I'll describe here:

1) Incorrect assumptions, and/or hopes. This is an odd occurance you would notice if you played a female character. Sure, a lot of people knew to go in and default their gender perceptions to "male," but there were a hell of a lot of people who were simply too dense to grasp the concept. They would either assume or hope that every female avatar was played by a female player.

2) Reality check. It would invariably come up that you need to expressly reveal your own gender, either due to people obviously operating under incorrect assumptions, or as a result of people directly asking you. At this point they get a classic reality check - no, that's no supermodel behind that hot elf chick, it's just some guy on the internet.

3) Liar. The people who make assumptions can feel betrayed by the reality check, when they learn that the reality of who you are does not match up with their own perceptions. Once you reveal to them that their assumptions are out of line, they may act miffed or disappointed.

4) Derision. This is another common reaction after the reality check, frequently made by some hyper-sensitive and/or homophobic individuals who are really freaked out/offended by seeing a female avatar and knowing there's a male player behind it.

A lot of this can be countered by choosing a leet, inappropriate, or otherwise bizarre name, but then you're just instantly viewed as a fucktard. I know when I see a name like "Tigolbitties" I'm given a generally negative impression of the player.

Now WoW is a bit different due to the things I mentioned earlier, but the fact of the matter is that crossdressing can still lead to confusion. Nobody is ever really sure of your gender until you expressly reveal it, either by obviously gender-specific discussion or by a simple statement of fact, things that may never come up in the course of a normal pickup group and may seem awkward to come out and state pre-emptively.

Women do play this game, after all, and while the assumption that a female character is played by a male player is certainly logical given the odds, you cannot be guaranteed of that fact without either asking or observing gender-specific behavior or discussion.

So, what's left? Give it up entirely, or at least go light. Otherwise it'll be a pain later, if you ever plan on playing a character seriously in the high end game. You'll be accepting social awkwardness in exchange for a digital pair to gawk at, and ultimately that's probably a bad deal.

I'm still playing WoW, albeit not with the same fervor as before. I spend a lot of time dancing on mailboxes and doing other stupid things, going out to fight only when Ed or somebody else is around to team up with.

In the process, I've begun to notice some of the problems that combat healing brings to game design, and have come to the conclusion that, if I ever design a large scale persistant multiplayer role playing game, it will not contain any combat healing at all.

Now, combat healing has been a staple of the genre since its inception, and it's become so engrained in how people play that it's difficult to imagine a game without it. However, if you can look past the "it's always been that way!" mindset and look at how it damages gameplay, you might just agree with me that it's time to get rid of it once and for all.

EQ is the anti-combat-healing poster child, so I'll poke at that first. This is a game that was built, from the ground up, with the "perfect group" in mind. EQ was warrior/priest/mage all the way, and for the longest time it suffered all the more by having so many classes with only one viable healer. With the game designed specifically around this group composition, most other groups were simply incapable of doing much of anything, and playability as a whole suffered greatly.

The net result of combat healing in any game is that you end up having one player who essentially sacrifices himself to make other players vastly more powerful. That creates 2 major problems right there - from the healer's perspective, it's a boring and thankless job, and from the designer's perspective it's a difficult dynamic that must be worked around.

In practice this leads to an unfortunate situation - one in which the game is, by necessity, designed around the potential groups with healers that are vastly more powerful than groups without. This means that, to keep content from becoming trivial to the "right" group, you must "design out" groups that are lacking healers, since an objective that's a challenge for a healer-equipped group would be downright impossible for a group with no healer.

The reason this is doubly bad is that, as I previously pointed out, healer is a thankless and unglamourous job, and one that most people want no part of. You sacrifice your own offensive ability to play backup to the superstars, the fighters and mages who are doing the fun work of engaging the enemy and making things dead.

That leads - you guessed it - to a healer deficit. Not enough people have the desire to play them, but they're needed for a lot of content.

Of course, some people enjoy playing the sidekick, but there will never be enough willing healers to provide every group with the perfect composition. The selfless among us, those who are willing to take one for the team, may set aside their preferred playstyles and create healers simply because healers are so essential. These poor souls suffer through the game just to make their groups viable, all the while wishing they were playing something else.

WoW, a game designed around having people solo a good chunk of the content, exascerbates the situation. Since (unlike EQ) you're expected to solo for most of your career, picking a dedicated healer means that you're simply less effective than most for a good chunk of gameplay. In a game with forced grouping, where almost nobody can solo effectively and you'll always be in a group no matter what you do in combat, healers are less painful - but in WoW, their offensive disadvantage is even more pronounced when they try to go it alone.

This can be somewhat mitigated by spreading out healing amongst severally equal viable alternative classes who all have other primary abilities; however, in the case of WoW, there is a single class that is simply superior to all of the "secondary" healers. Combat healing as a "class defining" ability creates the biggest issues - since healing is the single most powerful ability in party play, having a single class that is well above and beyond the rest of the healers means that:

1) The best (class defined) healing class will provide proportionally more power to a group when compared to his non-class-defined counterparts than any other class-defining role (e.g., the class defined healer will have abilities that vastly enhance the combat abilities and survivability of the group, whereas a class defined nuker or meatshield does not scale to such an extent).
2) You must balance the "difficult" parts of the game against groups that will have the class defined healer, which makes secondary healers undesirable in groups facing the difficult content.
3) Given the class defined healer's disproportional benefit to groups, his other abilities (damage dealing, etc) must by necessity suffer, or else you create a superman out of an already overpowered class.

All of this indicates that combat healing itself is the biggest barrier to balance in MMORPGs, and should be one of the biggest thorns in the side of players and developers alike. It forces specific group composition for maximum effectiveness, it forces design decisions to be based around that group, and it imposes a "correct" playstyle that simply must be worked around.

However, there are some possible solutions:

1) Use the potion-like cooldown for targets of healing, wherein each person can only be healed once in the course of a normal-length battle.
2) Remove the class defined healer and have several equally viable healing classes that have non-healing primary abilities.
3) Have only long cooldown times on healing spells, to prevent "spamming" heals which makes melee characters and pets virtually invulnerable.
4) Remove combat healing entirely, leaving only specialized very-high-cooldown abilities such as Lay on Hands that have value in being used sparingly, only to recover from dire situations.

The benefits of eliminating combat healing entirely are immediately evident:

1) You don't have to balance around a high-powered healer, thus enabling a wider variety of viable group compositions
2) You prevent the creation of an offensive gimp just to satisfy the need for a dedicated healer.
3) Combat will be faster since encounters will be designed without having characters with virtually infinite hitpoints.

The downside... well, the downside is that you take away healing from people who enjoy having it. A small price to pay, I think, to fix a broken game dynamic.

I've disabled comments entirely because I'm simply getting too many bogus ones posted from spambots. If you'd like a comment posted on anything you see here, send me an email and I'll add it myself.

What happens when it's actually fun inside the Skinner Box?

EQ was, well, not very fun. You did repetitive tasks to get more power, over and over. The rewards became more infrequent, the tasks more monotonous, until you were left just scratching your head asking why you were bothering - but you'd keep playing, because you just had to get that next level.

After leaving EQ, I simply hadn't encountered another game that had that kind of hold on me. Very few games had even held my attention beyond a day or two, and I had begun to suspect I was done with gaming entirely.

Now along comes WoW, the game that I always wanted EQ to be, and I find myself simply loving it - combat is fun and fast, quests are interesting and varied, and talents and skills are actually powerful against the monsters you're fighting.

Of course, as the levels increase, I do start to feel the Skinner Box closing in around me - quests involve killing more monsters, levels take longer to complete, and you have to rest longer after tough battles. Yet, even though I know what's happening, I just don't care. At level 17 now, I am still gaining levels without even noticing it, and I'm just flat out enjoying almost everything about the game.

I find myself being pulled into WoW the way I was originally pulled into EQ, only this time around the stupid annoyances that always drove me away from EQ are absent. I'm wondering if it's even possible for me to play this game reasonably, without sitting my ass down in front of the computer and pounding through 8-hour marathon sessions as I've been doing so far.

Hopefully the honeymoon is almost over and I'll start playing less, but I'm already starting to feel the first symptoms of MMORPG addiction. And here I was, thinking I was done with games entirely...

So, I threw FreeBSD on my IBM Thinkpad T20 the other day for shits and giggles. I've messed with FreeBSD in the past, and have some degree of familiarity with it, so this time I wasn't going in blind.

Installation was fast this time. I did a minimal install and was up and running with a basic system in about 30 minutes.

So far as I could tell, all of my hardware was working as it should out of the box. Network card "just works", and I got connected without having to do a damn thing. I installed X and had Fluxbox, Thunderbird, and Firefox working in short order (all pulled from the FreeBSD package repository).

FreeBSD is the system that Gentoo wants to be. Have I mentioned that before? If you're using Gentoo, you should probably be using Free or OpenBSD instead (unlease, of course, you're a GPL zealot...)

Let me say this - FreeBSD is flipping fast on this machine. I had previously been running Debian Unstable (my favorite Linux distribution), and I now feel a noticable speed boost in X since I switched. I mean, it's ungodly snappy from a system this old - of course, I *am* running Fluxbox, so it's supposed to be snappy, but I was running BlackBox with Debian so I don't think I can blame the window manager. FreeBSD is just faster.

I'm still a little scared of ports. With Debian, when I get a package, I know people have used the exact same package on tons of other systems. It's been tested, and it will work (well, not always in Unstable...). Ports I just have less faith in, but I'm not sure my lack of faith is really justified. Time will tell.

FreeBSD is so... minimal. It has an incredible manual, far better than anything I've seen for any Linux flavor. But when it comes to crazy system config tools... there just aren't many. You have your pkg_ utilities to deal with packages, and your cvsup and portupgrade utilities to deal with ports... but other than that, it's almost entirely standard Unix tools. I rather like that.

FreeBSD feels very "clean" to me, and it's rather growing on me. I may do more with it in the future.