Yarpen's Gamedev Ramblings

New address

2006-06-04T18:57:00.000+02:00

Moved here.

Shameless bragging

2006-05-20T23:03:00.000+02:00

Seems like The Witcher has won Best E3 2006 RPG Award at IGN and just-rpg.com. At least all the crunching was worth something :). On the other hand it's funny how reviewers have problems with remembering so simple name as Geralt. I've already seen Gerard, Gerhard and Gestalt(!).

Pet project

2006-05-16T19:21:00.000+02:00

As 90% of professional programmers, when I return home after 8+ hours of coding, I code some more. After rather long break I returned to my home engine. It must be the most overengineered piece of software I wrote, but I like it. My latest discoveries are:
- Source Monitor
- Typegen library
Typegen is cool, because it helps in solving one my pet peeves - setters/getters methods. Blasphemy, I know, you're not supposed to expose variables. I normally dont, but sometimes it's tempting, when I see a pair of Set/Get methods, and they do NOTHING else than setting/getting a value (and cluttering the code). Sure, solution with accessors is more future-proof, you can always add additional code the if need arises. Now, honestly, how often does it happen? And if it does, how hard would it be to search&replace? One reason to have very simple setter method is to test if the value is in certain limits. With Typegen, this test is moved to the value type class. I run it with asserts in debug build and then switch to simple typedef in release. Works very nice.

Game idea

2005-04-06T14:47:00.000+02:00

Pretty simple, actually. Massive (well, kinda) Multiplayer Online Football. 11 on 11, every footballer controlled by single player... When someone implements it just remember I thought about it first :-).

Expression templates

2005-03-03T09:55:00.000+01:00

OK, as WoW takes all my sparetime, today I'll go the easy way and simply redirect you to one of my old articles. Things have changed since I wrote it and now I think that ET's are simply not worth it, they complicate code too much and make it hard to maintain. Anyway, it's fun to try 'em out and it surely helps in understanding certain issues. So, if you're interested - follow this link.

Test Driven Development

2005-03-01T09:38:00.000+01:00

Noel Llopis wrote a very interesting article (two, to be exact) about TDD. I've to admit that while I use unit tests extensively, I've some problems with employing TDD (talking about my "home" project). I'm not giving up hoping it's one of these things that have to "click" and then it goes easier. One of the best introductory articles I know is this. Noel lists advantages of TDD in his text, I'll add one more, the most appealing to me - programs written using TDD tend to have really nice, simple interface. Sometimes, when one designs things traditionally he may overcomplicate things, not care about the user so much. Result is that you have to write 15 lines to create mesh for example (create renderer, register factories and whatnot). With TDD you just put yourself in the client's shoes, you work on interface first.
Side note: next week I'm going to my first GDC, can't wait! Hope to meet you guys there!

WoW impression

2005-02-21T12:19:00.000+01:00

I've been playing WoW almost constantly in my spare time during the last week and it's only getting better and better. This is my first MMORPG, so I may be surprised by features that are considered standard, but here's my list of things that make this title shine in my opinion:
- Graphics (please remember, it's MMORPG). We've come a long way since text MUDs. It may not be the most detailed 3D world, but I really like the style and colors, it reminds me Fable a little bit. Mage Quarter district of Stormwind is my favourite so far. The only thing that looks plain ugly IMHO is water.
- "coolness" - there are so many ideas here that can only be described as "cool". Pets are the first example that springs to mind. I've chosen my hero without really reading the manual (I always play dwarf warriors anyway :-), but when I've seen folks running around with tigers I was seriously considering switching to hunter (my next character will be either Tauren hunter or Night Elf Rogue). Some other nice ideas: shapechanging for druids, gryphon flights, mounts...
- World. Well, it looks like your classic, high-fantasy, a little "warhammerish" realm... But it's huge and I mean, HUGE. We've still not managed to gather all in one place (some of us started as night elves). I started near Ironforge and so far I've visited Elwynn Forest, Westfalls and parts of Loch Modan. This is like 5% of the whole world (and I can't claim I've been everywhere in these areas). Every land has it's own, distinct "feel", this can be best observed when flying from Stormwind to Ironforge for example. Some of the NPCs are "classic" (kobolds, wild beasts etc) and some are more interesting like mechanical golems. Plus, there's playable race that I really like - Taurens.
- Combat system. At the beginning I thought it's too static and a little boring. It started to change when I gained my first combat abilities and now at level 16 it already becomes interesting... And I'm still talking about PvE, combat movies from PvP (made by lvl 60 folks) are breathtaking. There's much complaining about certain classes (warriors are generally considered to be much too weak for example) on Blizzard forum, but I can't comment on that yet. I'm going to take a "crit-hit" route, which is supposed to work the best for warriors and we'll see. Anyway, the nice thing is that there are so many combinations of skills/professions/weapons that almost everyone can find something that suits him. I'm playing with fellow CDPRS programmers and each of us plays different class and in different "style".
On the other hand, I'm not so excited about the quests, big part of them is "bring X from A to B" or "kill 'n' Zs", but I guess this will get better as my level increases. Another complaint is that walking takes much time and it's just boring (try to get from Ironforge to Loch Modan for example). It's not so bad as in Morrowind, no, but still a little bit annoying (luckily, as soon as you discover the route you can use gryphons). I've ready somewhere that R. Garriott is going to address this issue in his incoming MMORPG - Tabula Rasa. Oh, and there's one thing I'd definitelly add - arenas where players would fight in gladiator tournaments (winner takes money/item prize) or maybe even some "All Stars"-like event where you can show off your agility, strength or magic powers.

The adventure begins..

2005-02-15T12:52:00.000+01:00

OK, folks, World of Warcraft is here. This means, I probably won't be updating very often :-). I only started playing yesterday, I'm Lvl7 Warrior (Dwarf), so basically I'm restricted to easy areas/quests only for the present, but I really enjoy the game, one of the best I've played since a long time.
PS. Oh, yeah, just in case anyone wants to say 'Hi' - I'm Zoltan at Emerald Something (forgot the exact name), "Normal" mode.

Visual Unit Tests

2005-02-11T11:06:00.000+01:00

Recently I've become a big fan of "visual unit tests". I use unit tests quite intensively in my home pet project. Basic framework is modified CppUnitLite. But I found out that testing algebra/gfx related code only with CHECK_EQUAL etc is not very helpful. If, for example, I test AABB/triangle intersection I don't want to calculate correct results by hand (and what if you make mistake in calculations?) and see if I received them. I want to visualize it. It turned out to be pretty simple, all we need is an image writer (for TGA it's like 5 minutes of work). Now you can draw all your triangles, boxes, curves, lines and whatever. Of course, the disadvantage is that it isn't automatic anymore, so it's better to leave basic algorithm tests in code as well.

Sidenote: In order to get my VUT running I needed some line drawing routine. Of course, speed is not important, so I could simply go with naive algo, but I thought: "heck, let's see if I still remember Bresenham". Well, as it turned out, I didn't. After 15 minutes of trying to write it without thinking I had to go ahead and practically reinvent the algorithm from scratch. It ain't even funny, because only 5-6 years ago all kinds software rasterizers were my bread & butter (oh, the demoscene days). I got to the point where I could probably write textured triangle filler in 30 minutes in hand-optimized assembler. I tell ya, all those 3D accelerators are pure evil :-).

Addendum

2005-02-03T17:18:00.000+01:00

Seems like we're not the only one fighting with the problem of prototyping combat system without all the animations: http://www.gamedevblog.com/2005/01/notes_on_ema_th.html.

Importance of early prototyping

2005-01-17T11:11:00.000+01:00

With every title I work on I'm more convinced that one of the key ingredients of a successful game is early prototyping. At the same time, I've to admit that it wasn't really applied in my previous project, but we're trying this (to some extent) now with "The Witcher". It doesn't go as smooth as one could hope, but there are reasons for that. This is our first title (as a team), we have to develop/extend the engine in the same time (it should be much easier when starting with a ready one). Also, it turns out that the sentence: "if it's supposed to be fun it has to be fun even without the graphics" is not always true. In our case main problem is combat. In order to predict how it'll all work together we need a complete (or almost complete) set of combat animations, code supporting that, at least basic implementation of combat system and so on (trying to tweak combat using two boxes is not really helpful here, feedback would be significantly reduced). We're finally getting to the point where focus shifts to gameplay. Of course the cost of potential design/code changes increases together with the work progress, so the earlier you can find out your crucial weaknesses and strong points - the better. I've read a PM to "Dawn of War" lately, those guys started gameplay tweaking and prototyping very early and it shows, I really hope to improve this process in our next project. Let's see at "Jade Empire". It seems like the game itself was almost finished one year ago, there's not much difference in gfx from the latest screenshots and those presented at E3. That means, they have ~1 year to polish the gameplay... Knowing Bioware, I expect JE to be their best game and one of the best RPGs ever. Another example - Halo. As Charles Bloom wrote - Halo2 is a kind of insult to other developers, bacause it doesn't really include any groundbreaking stuff. Sure, graphics & music is top notch, but the strength of Halo is that crucial gameplay elements are identified, took out and polished. It focuses on things that really matters and the result is that's just pure, condensated fun (another last year example that comes to mind would be latest Burnout, it's almost perfect).

Wanda and the Colossus

2005-01-14T10:23:00.000+01:00

This must be one of the prettiest games I've ever seen (juding by the screenshots). Can't wait to get it, especially that it's developed by the Ico guys. Breathtaking stuff.

JC starts blogging!

2005-01-04T10:58:00.000+01:00

You can find John Carmack's blog here.

Heroes in FPP games

2004-12-29T15:39:00.000+01:00

I first started to think about this after playing Painkiller. It's a solid, old school FPS with great graphics and very good physics. But when it's finished there's not much to remember and return to. Now, don't get me wrong, I like the game, kudos to PCF guys (they're located in the same city, so I had an occasion to meet them in person), but it will be hard to build a strong IP from that. The key difference in such case is that people may wait for your next release, just because you've proven you've potential, but they don't really care if it's a sequel or not, the brand itself is not very strong. This, of course, is not only PK problem, it's problem of first person shooters in general. There are several memorable and world-wide known characters from TPP games, maybe it's just because you see them on the screen, so some design work has to be done. How many true FPS heroes can you rememember? For me, there's virtually one - Duke Nukem (and the game is from, what, 1996?). Doom, Quake are all strong titles as well, but they base on fame and charisma of their creators, main characters in id's games are practically anonymous. Heck, we knew more about the main protagonist from Wolfenstein (name, face at least) then we do about the one from Doom3. id's game defend themselves also because they have some other elements that can cover lack of believable hero = monsters (original Doom was great in this aspect, Quake quite OK, too (Shambler)) and weapons (legendary chainsaw, BFG). The perfect example from the movie world is "Alien". We have the main character here, but true symbol of the serie are "bad guys" (see most DVD covers, no Ripley in sight). In such case the enemies make the IP. This is another problem with PK, IMO. The weapons are interesting, but monster could have been designed much better, they're all too similar to each other (especially AI-wise). It's still possible to create a good game even without all those elements, but it's very hard to achieve a cult status this way. It's kinda unnatural to identify just with a title, especially if consecutive parts are not really connected. It's much easier to get excited about something more specific (hero, enemy, weapon, world). I'll mention DN again. It's an old game, but it's still remembered, Duke has many fans and people still hope DNF will be released one day (and I hope Scott will write a note about designing and creating Duke character). I'm sure that if only there were more DN games he would be one of the industry icons now together with Lara Croft and Max Payne (3DRealms does it right again, coincidence? I don't think so).
One may wonder if FPS even needs a strong main character or story. In most cases it doesn't even affect the gameplay (you just have to kill everyone in sight, who cares whether they're from Mars or Pluto). But it seems to stimulate people's imagination (at lease some of them). Let's see how does the hero look like (remember bathroom mirrors?), let us hear his voice, give him personality, it's hard to sell a lifeless robot. Add some background characters, this opens a great possibility for spin-offs if done right (see MGS, it's not a FPP shooter, but good example nonetheless, their characters are just real), think about the world, it's geography and history. Most of this is pure design work, can be almost 100% orthogonal to the gameplay, but works best only if blended carefully. Only this way the game universe and hero can start living its own life in fan fictions, unofficial sites, comic books (which are, BTW, great tutorial of how to build believable characters/worlds) and so on.
Most of this stuff is implemented in Half-Life2, which may not be your typical kill-everything-that-moves-and-their-families shooter, but it's FPP game after all. The only think I don't like here is main protagonist, who seems a little "flat". But anyway, great work by Valve, hat's off. I hope this example will show other developers that this additional work pays off and we'll see some more interesting stories in upcoming titles.

About Me

2004-12-28T15:57:00.000+01:00

Let's start with a few words about me and this blog. My name is Maciej Sinilo, I work as a game programmer. My current project is The Witcher - action RPG based on famous Polish fantasy books by Andrzej (Andrew) Sapkowski. For more details about my professional life, you can try my CV, which should be somewhere here. The main subject of this blog will be game development of course, mainly design. I'm not a professional designer, but after some years in trenches you start to think about it more and more. Plus, we're in this because we love games and play them a lot, we have contact with brilliant and not-so-brilliant ideas every day. Anyway, expect rather some loose ramblings about games I play(ed)/would like to play/make, than professional design essays.
Finally, please excuse my English. I know it's far from perfect, I've learnt it mainly from technical documents, Eurosport and Cartoon Network :-). I didn't decide to write in Polish, because it would drastically reduce my potential audience (gamedev industry in Poland is still in diapers).

Hello

2004-12-28T15:11:00.000+01:00

Hello everyone. This is my first post, just to see if everything works. I'll put some more serious content today or tomorrow.