Game News Javelin - 3.5 OGL based Rogue-Like cRPG Released

Oh shit, a party-based RL? I am all over that shit. Giving this a go.

EDIT: Man, this thing is barely functioning. Pass.

 

Well, that was disheartening.

 

doesnt crash for me
but thats the only positive aspect of it
i think id rather play pillars of mediocrity than this

 

No, this is disheartening.

 

hmm not "finished" enough for me.

 

The graphics glitches like crazy for me and the AI takes a decade to make its move causing me to fall asleep on enemy phase

 

Are you sure your graphics are actually glitching? This game makes an "unusual" choice about how player vision works: When it's a given character's turn, the player can only see what that one character can see. Even more oddly, when it's an enemy's turn the player can only see what that enemy can see. If you don't know what's going on, when it flips between characters the result could easily be mistaken for glitchy graphics.

Get all of your characters the dark-vision upgrade and things improve somewhat. Unfortunately you can't buy dark-vision for your opponents, and some of them can't see very well. (Darkmantles seem to have only one or two squares of vision.) Things get especially wonky when your characters are spread so far apart that they can't even see each other, which for some mysterious reason is sometimes the case right at the beginning of combat. (There's a minimap but it's not very clear, making it not very useful.)

You can edit the "ai.maxsecondsthinking" value in the "preferences.properties" file to make AI turns go faster. E.g., set it to something like 0.1 seconds. I can't imagine it will perform well though. Even at the default 3 seconds (and running on 4 cores) the AI is very weak. However, if you were to actually set it to 315576000.0 seconds (a decade), it might be pretty smart.

***

I just got some message about an invasion, and now it says "Invasion!" in the bottom-right part of the overworld display data. I'm sure that means something, but I'm still walking around and murdering random ponies and dogs and what not, and am not seeing anything that actually looks like an invasion.

 

I was curious because i also did prototype d20 implementation once, so i was interested how others would approach it (J1M). So very nice to see some code.

Kudos to this guy for having the determination to push through to "1.0", but the code does not look very good. Many things are way too mixed up.

Implementing d20 is a real bitch, because if you don't design properly you end up with thousand of lines of unreadable code. There are just so many nasty exceptions to the rules, that one tends to quickly patch around those with shitty hacks. At least that is what stopped me - i wanted too much too fast and realized at one point that it become too much effort refactoring the code when encountering new requirements.

 

Hello guys, I'm the creator of Javelin! Sorry it took me a while to find this thread...

There was a bug that would act up on slower computers. Version 1.1 should fix it!

Again 1.1 brings major changes in this aspect. Even though the gameplay remains unchanged the screen now tries to show the entire battle and makes "unseen spots" half-grey so you (as a player) can both see everything that's happening and also understand that your character can't see that far. The screen also doesn't jump around every time a new character is acting - it tries to remain as still as possible making the game now feel more as if you were playing a table-top game with miniatures. It's not perfect but hopefully much much better!

Indeed you can and the game is designed around that: it's supposed to be a game with a though AI opponent, not an action game but you can always bring the thinking time down if that is your style. In 1.1 the AI has been tweaked to act smarter and also improved as far as thinking speed is concerned. If you don't change the thinking time it will still think for the same time as before but hopefully make better decisions.

I intentionally left some things for the player to find out for themselves! The invasion is certainly around there somewhere for you to find it - and you should sooner rather than later, if you're strong enough to move around the map without having to fear for your life!

Some of it is tangled indeed because I used another game (Tyrant) as the code base for Javelin and when you create something new from inside a big game things can be difficult but otherwise being a programmer for all of my professional life I can be quite proud of how well the code is structured and working together. Of course, if you plan to go in and find a d20 Java library you can plug-and-play into your own project I'm sorry, it''s just not what the game is for - and since the AI is extra sensitive to performance issues it just can't be programmed that way. Believe me, I tried to use something like that at first http://sourceforge.net/projects/numbat/ My plan is to get rid of all old code by 2.0.

This became a long post but if you're still with me then here are the highlights for 1.1, coming soon (plus lots of other minor improvements):

* Dungeons!
* 12 or so new randomly-generated battle maps
* Vehicles!
* World and battle screens are largely improved, creating a much better game experience
* Each town has a unique selection of items and upgrades
* Towns now take a certain amount of time to craft items or upgrade heroes

I also want to put up a crystal system like in Final Fantasy Tactics, even though it's not very D&D-ish. It would happen only on special battles though, so die-hard D&D fans need not worry.

 

top kek

 

Why Java?



Otherwise the premise sounds good.

 

ASCII would make it look better

 

> MFW my 20-year old nickname becomes a meme

Obviously I really like Java but to get up-close and personal: Java, being statically-typed is great for big bulky projects. A complex game with real AI, hundreds of monsters, several screens, network code (someday perhaps), etc takes time to maintain. Java lets you make big changes and have compile errors everywhere immediately so you can start working on those. If you have the same on Python, Javascript, LisP, etc you're either gonna have to have a lot of time in your hands or to write a lot of tests, which also takes time. There are some projects I'd never use Java for but big bulky games can benefit from heavy object-oriented programming, great refactoring/debugging/profiling tools and so on. The fact that it works on any operating system, can be downloaded via Java Web Start and is one of the languages with most open-source code available is just the icing on the cake.

But trust me, you weren't the first person to immediately ask me that :/

I really doubt it. A collaborator spent several days working on the monster images - you just don't get to that level with ASCII. Perhaps it's more your thing but I like 32-bitish stuff, it's a lot richer while still leaving most to your imagination. The fact that there has been a lot of work updating classic roguelikes to graphical interfaces should say that it's the general trend - take NotEye (link) for example.

 

Meh. Java may run on nearly every operating system, but it runs like shit on all of them too.

 

I have nothing against graphical roguelikes, it's just that Javelin looks not very good, and yes, most ASCII roguelikes look better (more coherent art direction, dungeon and environment design is much more varied) due to the simplicity of their art style. And if you compare it directly to something like Brogue that makes every dungeon look like a work of art, Javelin doesn't even come close.

As for graphical RLs, take a look at stuff like CastlevaniaRL, that looks good.

That said, I like the overall concept, I would like too see more party-based roguelikes.

 

I "stole" all graphics from Tyrant, the open-source roguelike that I used as Javelin's code base. I myself spent very little time working on graphics so I agree it could be largely improved. Hopefully by 2.0 all images will be in individual external files so it will be super easy for the community to make changes and enhancements to the game's visuals. All monster images are already like that on the "avatars" folder - these were mostly compiled form various sources by a collaborator.

I'll take a look at CastlevaniaRL when I have the time. I saw that before but to be honest when I think of Castlevania I'd much rather fire up my Aria of Sorrow and beat it from start to finish than to play a RL.

That was the sole reason I started working on Javelin. I mean, I had other ideas too but I could not let myself see hundreds of great roguelikes online and none with decent, focused party-based gameplay. Hopefully it can become a much better game as new versions come out!

 

kek Is there a timeline for release of 1.1? How about 2.0?

Also how big of a party can you have? Don't really want to try this until its updated.

 

redacted

 

oldmanpaco : 1.1 is coming out before the year ends, hopefully. It's done so it's more about playtesting than anything else now. Besides that I might add a final feature to the version still.

There is no limit to how many heroes you can have in your party or how many parties you can have at once. The screen is not optimized for anything beyond 5, 6 or 7 per party though. The current maximum number of enemies per fight is 9 but you can change it up to 16 in the configuration file (pretty easy to do). Remember this game is inspired in D&D so the baseline is 3 to 5 characters per party.

Severian Silk: thanks but I don't think making the game isometric is an option, not at least until 2.0. The problem isn't even the lack of usable tiles and graphics, it's the fact that right now I'm working alone and I have plenty to do. If someone is seriously interested in helping with that it would be a great addition to the team.