Happy New Year all; sorry for the extended absence. Last year was a tough one for me. If you read this blog as some kind of “Statements from CRS” outlet and view me as a spokesperson for CRS; don’t read this post. Seriously, fuck off. I’m gonna write me an honest-to-god blog entry. It just happens to involve WWII Online by way of the fact I work on it. This one is for the folks who know me. If you don’t know why “Gyroscope” is funny, try reading Rafter’s blog instead.
2006 and I got off on the wrong foot: we had problems over Xmas 05-06, and Doc and I were on duty covering them. It literally took me months to recover from that.
Then we finally got onto TOEs – tying equipment to brigades rather than towns. This is something we both want and dread. For instance, shilling logged in over xmas and /reported this:
I once had a hand in this games future but to log in and see the pitiful state it has come to. Nothing to Spawn, No more Formations of units be it ground or air. Lemming style attacks its a shame to see this. There once was a dream of WW2OL and this is not it. Youve lost your way and … Sad to see guys Im sorry but More and More im glad I dont play anymore
I’ve noticed a number of returning players air similar grievances. Those of us who’ve been around have forgotten how massively brigades changed the face of the game, and those of us who’ve survived thus far probably don’t realize the whole learning curve to figuring out where to spawn. When you know how the system works, its really obvious now. If you’re looking at the game with 2003 or 2001 eyes, you might well be even more confused than you were back then.
In honesty, we weren’t ready for TOEs. If we’d added TOEs as priority #1 for 1.25, the map would have collapsed worse than our christmas “frame rate” patch did. You do realize that it’s not a bug, Toto pointed out that its a cool new feature that whenever a plane is about to crash he enters bullet time; course we have to do it to everyone else in the area for it to work… Ok. You’re not buying that. Bah.
In the next few [working] days I’ll be decoupling the supply system and connecting the Brigade Supply system in its place, and we’ll be taking it into beta. I’m backtracking a little on the system we were going to try out and instead of hard-coding everything, I’m switching all the critical code paths to Lua script calls. This will massively improve the development time – allowing us to modify a script and reload it rather than requiring lots of server ups/downs/restarts/whatnots. It’ll allow me to combine development and tuning and give us a much greater flexibility when we deliver it.
If all that was chinese, the translation is “pressing on to get to open beta stat”. Two Weeks(TM). (Those who don’t have a 6+ year history with the CRS staff might need it explained that Two Weeks(TM) is not a literal measure of Real Time)
Our pre-xmas patch was atrocious and it needn’t have been. The last version we beta tested did not have the abysmal afflictions we subjected you to. Nor did the first 1.25 release. But we allowed ourselves to tie up some loose ends: spit & polish. We fixed some bugs and rolled out a hotfix.
Because we use Microsoft SourceSafe it can be real easy to get caught with your pants down. As I mentioned in my other post, branching requires creating a full-on copy of the project you are branching, including all the history, etc, etc. On a project of our size and longevity, that has come to be a real time consuming task, and frankly there are times where we’ve screwed the pooch just because some has decided there won’t be any harm in a checkin to something they’ve been working on when, with another source system, they might have checked their changes into a quick branch; or because they’ve checked in a change to the wrong repository.
Incidentally: I am always confused by the terminology we use here; when 1.23 went out, $/pn/ww2/ww2ol/ was ‘branched’ to become $/pn/ww2/ww2ol-188.8.131.52/. That’s our working snapshot of the source as of 184.108.40.206 release. It took me a long time to get my head around that just because I was used to the stable, release “head” of the development trunk being the baseline and the in-development not-yet-accepted work being the branch. IMCMUW the baseline is what you know works.
Yesterday was my first day back after the misery that was this christmas. I’ve actively taken time out of the game over the vacation period because working on the game andplaying it would have left me dead in the water at the start of the year. But a little absence is good. When I came in yesterday I mentioned a couple of thoughts I’d had over Xmas about some “add on” content we could look at adding, features that require development work we want and need to do, but which take it a step further without imbalancing the game.
I really don’t want to go into any detail, we’re going to be refocusing on “Stable, Fast, Fun” again at the start of the year, I’m going to be squirreling away on finishing TOEs this next couple of weeks. I’ll be working out the details of the upcoming, and hopefully final, major host sweep which will include one or more of:
. merging map and chat so that all lobby functionality occurs through a single, persistent, chat connection;
. finalizing Netcode2 usage on the cellhosts – netcode1 will still be used for the map/chat stuff, but all the world data will be netcode2;
. removing the ‘redirector’ which transfers players between map<->cell hosts and is the blackhole of “waiting for mission results”;
. replacing the server-side implementation of the spawning process to make vehicle creation and player attachment fully independent
That right there does very little for you the player. Its somewhere between 2-6 weeks of work. You’ll see some connection stability improvements, maybe some reduction in warpiness, possibly an increase in visible infantry for players with higher bandwidth, possibly some improvements to multicrew. You might not notice things running slightly more smoothly on the host side, perhaps a few less sudden despawns.
But this is groundwork stuff. It opens the door for creating/changing missions from in-game, improved multicrew, it takes a serious chunk out of polycrew and vehicle entry/bailing development time, it will enable us to finally tackle scoring with a 21st century approach (and without the current flakiness).
But this isn’t even the stuff we were talking about; these are things we were planning last year. Rather it seems several of us, over Xmas, have been thinking about some of the same “down the line” elements of the project, and how we might structure our development schedule to make them happen sooner, and to introduce some value-added (edit) content based on that work to help cost the development.
For instance, if we released our next patch with an in-game card game, many of our players would resent the fact we’d diverted development time into that game-irrelevant project. But consider, instead, if the next patch contained bug fixes, fps improvements, in-game mission posting and creation AND for a one-time $19.99 add-on fee you can get an add-on pack which includes in-game ports of the 5 card games that Playnet owns source code for, in-game email, and a bunch of other features which don’t modify gameplay but they embelish the game playing experience, which are all spin-offs of the neccessary work done to the “standard” version of the game that were implemented by hired help (which is why it would have to cost a fee).
Yes, I know you’d much rather we hired help to fix bugs and improve FPS. But do you want to pay for WWII Online: The Bug Fix Pack? No, you don’t, so quit being a fucking retard. In the fictitious example outlined, people who want to play cards in WWIIOL would pay money for an add-on that includes it, and in doing so would have contributed to bringing in extra resources that went towards some of the stuff that you didn’t pay extra for. edit: This is a ficticious example which assumes that somehow introducing card games would double-up with some existing, essential fix/repair work in the game that we are otherwise finding it difficult to factor in to our development roadmap.
Consider our last patch: we added bouncing grenades. Lets say that cost us $10k to implement. Now lets say we had thought about it and decided to add a trivial “golf course” to 1.25 in such a way that we could just plug grenades in to it and have bouncing grenades “free” as a function of that work. And lets say we then sold that “golf” add-on for $25 a pop and 1000 people bought it. This is a ficticious example again, the content isn’t worth $25, but if you don’t get the idea by now, the close button is right up top :)
Anyway, the stuff we discussed, I have to say, I’m enthused about. Its stuff that we talked about on the beta forums 6-7 years ago and even on the WarBirds forums 8-9 years ago; other games have realized some of those ideas already so we won’t be breaking new ground, but I think it will actually surprize people by how much it adds to the game, and it has some really exciting possibilities.
Fear not those of you who read on despite me telling you not to: It is not in development now. It’s not even on the whiteboard yet. Time is not being diverted from TOEs to work on a card game or in-game mail or porno mags in trucks (although, IMHO, the latter probably warrants it). Look, didn’t I tell you to fuck off?
We’ve also cleared some of the last Chapter 11 (bankruptcy protection) stuff so cashflow is improved, as long as we didn’t destroy ourselves with the xmas patch. Gophur, Martini and Rick are working feverishly on a patch that fixes not just the 2fps disaster but it may fix some of the “randomly getting really low FPS some nights” issues too.
Netcode2 stats are looking pretty healthy. Very low connection loss rate, which I think is largely accounted for by people crashing. That’s a bad thing unless you’re wearing the host-coder-hates-CTHL hat.
Speaking of which, maybe I should open a blog-store with some CTHL hats and t-shirts to help covering the costs of my driving lessons :)
So 2007 starts out looking fairly positive.