The past few weeks have been one of considerable upheaval for how Epitaph is managed architecturally. We were previously hosted on Linode, and I (for a short time) ran a development server on one of my home computers so that people could do risky coding without it impacting on the game proper. If someone needed to dramatically restructure key objects, they could make sure it worked on a separate MUD before bringing it to the live game. When I moved down to Canterbury, the computer that hosted the dev system was damaged in transit, and we’ve been without a development system ever since.
Now, I posted a while ago (http://drakkos.co.uk/blog/blog.c?action=filter&blog=mud%20commentary&id=722) about our hosting situation. We moved over to Linode, who are an excellent company with good VPS solutions. They’re not *cheap*, but they were cheap enough for what I wanted at the time. The VPS I purchased from them came with 768 MB of memory, which seemed ample but before too long we started to bash our heads against it. A month or so after moving to Linode, I upgraded the package to a full gigabyte of memory. Even so, long uptimes would leave us feeling the pinch after a while. Still, it was fine and I was happy enough with it until I heard about BHost (https://www.bhost.net/home?afid=119631).
I have no problems at all with Linode – good people, good support, and an excellent architecture for managing my VPS. But it cost me 25.50 pounds a month for the Xen hosting I had (the Linode 1024 package). For a mere 12 pounds in comparison, BHost offered twice as much memory, a faster host, more storage space and three times as much of a data transfer limit. It seemed, and *seems* too good to be true. I decided to give it a go, buying their platinum package to host a proper dev server. So impressed was I in the end that I bought a second platinum package, and now for 24 pounds a month (less than what we were paying at Linode) we have two rock solid platforms, each with two gigabytes of memory. Seriously, I am SO HAPPY with BHost it’s a little bit creepy. I did a web search on ‘how do bhost stay in business', and the google responses suggested it may be a consequence of restless spirits. Still, if all I have to give up to get this kind of hosting is my eternal soul, it’s a small sacrifice.
Anyway, this post isn’t supposed to be an advert for BHost (although it totally is, if you need hosting try them out and use the affiliate link I posted above!). It’s just an overview of the change in our hosting packages so that you can appreciate what’s going on under the hood. The biggest impact for people coming to Epitaph is that there are now two MUDs and two URLs for those MUDs.
Epitaph Black Ops sits at drakkos.co.uk – this is now our development address, and very soon we’ll be closing the doors on this MUD to players. It’ll be used for all of the development that goes on, and as a ‘test server’ for play testers. This means a more stable experience for players because they’ll never have creators messing around with stuff in the background on the MUD they are playing on. All permissions for creators on the live MUD will be removed. It’ll all be done on Epitaph Black Ops.
That leaves the proper, ‘live’ Epitaph at epitaphonline.co.uk (note the address). It’s a completely separate MUD on a completely separate VPS – we aren’t sharing memory, CPU time or anything. It’s also hosted on exactly the same *package* of VPS, which means we know from Epitaph Black Ops how code is going to perform, and we can easily sync the two platforms up without worrying about incompatibilities in the configurations.
The impact of this is that players aren’t going to see fixes to Epitaph on a day to day basis. Previously bug-fixing has been a process of going through the database, making the change on the live game and then marking the bug as fixed. Now, fixing a bug means that you’ll see the consequences of the fix in the next patch. It means slower responses, and that minor niggles may be niggles for longer than they otherwise would, but it also means we can be more disciplined about the game that we provide. We can properly numerate the mudlib versions (which we can’t sensibly do at the moment), and package up new game content into one push. We can provide formal patch notes for changes (like you might see on Warcraft or SWTOR) in a sensible and coherent way. My plan is to have minor patches on a monthly basis (bug fixes, efficiency improvements and so on), with major patches on a bi-monthly basis (new areas, quests, items and so forth). Patching the live game is assimple as my entering a particular shell command and then rebooting. It’s all very slick, and it’ll mean a smoother experience for anyone playing.
The downside to separating out the MUDs is that we duplicate some things that shouldn’t be duplicated (the boards and the bug database for example) and make it more difficult to ensure social cohesion. We can’t currently chat across MUDs without going through one of the intermud channels, but that’s a temporary problem. Eek@Epitaph is putting together some truly fantastic infrastructure to let us do everything from single sign ons to Epitaph-specific cross MUD chatting. His architecture is going to let us hook ourselves into a number of external applications too – we’ll probably end up ditching our current bulletin boards in favour of a fully external forum that authenticates directly against the MUD. We’ll be able to link Mediawiki accounts to players on Epitaph. All of that will hopefully be coming together in the near future. It’s all very exciting.
As for how Epitaph itself is progressing, I’m very close to calling an end to our open testing period so that we can focus on adding in some more areas, fixing the current show-stopping issues, and polishing everything up in line with the feedback we’ve received thus far. I think the open period has been very useful – we’ve gone through a lot of iterations of systems and concepts. We’ve gotten rid of mechanics that just didn’t work, and evolved systems that needed it. I certainly have gotten a lot out of the process.
The MUD running at epitaphonline.co.uk is most likely going to remain open to players in its current state, with only minor maintenance patches being pushed across to fix runtimes. Behind the scenes, on Epitaph Black Ops, more substantial development will be going on. We’re currently working on a range of exciting things that have been pencilled in at the last minute for the release of Epitaph 1.0. Many of these new areas and concepts are also laying the foundations for the features and content that will make up Epitaph 2.0.
The past few months we have seen pretty slow progress because of the demands of my new job, but we’ve been joined in recent weeks by several new creators in whom I am investing very high hopes. The new content will start to provide the branching play experiences that I want to see, and also move the focus of the game away from the single city of Dunglen and on to a wider and more varied environment.
The future’s so bright I gotta wear shades.
 Epitaph, as a consequence of the lib we run on, is a GUTSY BASTARD.
 A system from which we are almost certainly going to be banned sooner or later.