How about paid DLC? On Steam these are clearly marked and there is a lot of these. I have not noticed any on Desura... Is DLC an option on Desura? Is it feasible on Desura?

I went through Feedback, DevLogs, Announcements... and I can't really find any strategy games. Why is it? Is this genre not appealing to indie devs? To hard to make? To hard to sell? Some other reason maybe?

Make a lot of smaller games. Not because it's easier, but because you WANT to fail often. If you make a game and it is great you learn nothing (you simply got lucky once and you don't know how to recreate this), only if you make a game and fail you learn what is not working.

Deadline. Deadlines are fun and help you finish games during your life time. Notice that any moron can make the best RPG ever made if given unlimited time. It is the ratio of quality to time used that counts.

The graphics (blocks) are very nice, this is what made me try the game

As for gameplay it wasn't too fun for me, it felt... heavy and stressful. Especially the level where you had to keep the box in mid air without it exploding. You make one error and you die (in comparison Tetris feels much more relaxed). I think I would enjoy it more if it was more into direction of puzzle than arcade.
Note that I'm not a fan of this kind of games (tried it because of graphics) so you probably could ignore what I said.

There are two approaches to what a player is:
a) a player is an idiot who plays primitive games only and his buying habits are purely based on impulse and graphics are the only thing that counts
b) a player is smarter than you, he wants to play good games and is capable of judging is a game is good or bad before buying

If you believe in a) you should make Bejeweled (match 3) and at all cost get to top 3 charts on the casual portals. If you believe in b) you should make Minecraft and ignore market research and marketing.

Of course these are extremes, usually devs believe in something between these two.

that may includes developing games seriously but then it will certanily not be fun

Definitely not true. I have been making games seriously for years and I still find it fun. I would even say that the additional worry of "not getting bankrupt" adds to the total enjoyment of making games (althrough, I might simply be totally insane, that's a posibility too :D), the constrants of "making the game sell" does not make the whole process unfun for me, on the contrary. It's a lot of fun that I have to worry if a game sells and this worry actually made my games better (because I first have to assure that the game is playable, which is not obvious if you are not doing a commercial game). It is the years of being the amateur who is doing games only for own enjoyment and artistic vision that weared me out. It definitelly is fun and thrilling.

As for crazy competition... You highly overestimate the number of people that take it seriously. It's not the 5% you mentioned, it's more like 0.005% :D Being half serious is probably enough.

It's fundamentally wrong to use word "should" in relation to a player. The developers should or should not do certain things, but players are to only have fun. Thinking in terms that they have to/should play in any certain way is crazy.

OK, players should not take the game from the shop and leave without paying, but that's the only exception related to the player and the game relationship :D

This topic can be seen from so many angles... I will just post random related very short sentences, without explanation why, these might help you make up your mind. Otherwise I would need to post like 10 pages long :D

- Always cut down the number of features, never cut down testing and polish.
- Finishing the game as fast as posible is a virtue (true x10 if this is your first game).
- If you spend less time making the game it will be better than if you spend a lot of time (contrary to logic).
- Before you choose a specific money related thing ask yourself what are the consequences of not genberating X money on Y time (in short, when you are going to go bankrupt). Without this it is merely a discussion about ethic.
- Is your goal to be rich or to make the kind of gmes you want? You rarely can choose both.
- If you do it, whatever it is you do, will the players who bought the game: a) sign up to your mailing list and wait eagerly for your next game; b) curse you and feel cheated
- Are you building your fanbase or releasing game after game to random consumers you will never meet again?
- If you feel that "And if you don't sell much, then what was the point of investing all this time in gameplay?" what is the point in making games in the first place? Make business software, it always sells better than games.
- I don't know where the "crappy games sell" comes from, from my personal experience only things I loved brough money (proportional to how much I love them). It might not necessarily be true for you through.
- Always cut down the number of features, never cut down testing and polish (I feel this one is worth posting twice )

In 3D they usually have it easier because they have predefined and static map. You can simply manually define choke points in editor and that's it. In the game I have in mind there would be a lot of destructible terrain, reshaping the dungeon by the players, etc. It might be extremelly dificult to define what a choke point is. Also, in 3D you most of the time have passages that accomodate more than 1 unit (so at least 2 way route can be created), in tile based there would be a lot 1 tile wide corridors which makes cylindrical collision useless.
But maybe instead of queue system make some priority system? Like a unit that waited too long gets a "priority passage" flag?

As for "some unit has choosen this tile as destination flag" I think better not. The distances will be sometimes high, so in some cases the room could be used and then freed by another unit before the one that wanted to use it first arrived. Also it would make rooms artifically cramped in case of traffic jams (the room would be fully occupied by flags of units that can't reach the room yet).
Or maybe set the flags only if the unit is really near the destination?

The crowd behaviour (hive) concept sounds interesting...


In game scenario:
There is a room in the north and a room in the south, connected by a very long 1 tile wide corridor. There are two units, the 1st is in the north room and want to reach south, the 2nd goes from south to north. The assumption is the engine does not understand the definition of "corridor". How to handle it?

I guess, with the 1 unit per tile rule it can't be done... They would either need to occupy the same tile for a while or swap.

But let's say there was a queue system implemented and it works. Now, wouldn't it look strange? I mean, it's not how it works in real life. You don't know that there is someone on the other end of a very long tunner, you just try to to enter the tunnel, unless you see someone almost leaving it (in such case you wait and the enter). But if the other person is somewhere far, far... you don't know it and enter, even if you expect someone is there, you hope to somehow miracolously avoid each other in the middle
So, I wonder id dirty cheap swapping position would not be better for gameplay reasons than perfect avoidance algorithm...

Hmmm, 80% is not bad, it means 20% payed above minimum (I run freemium MMOs and 95% never pay anything, not that this model is compatible, but still ).

Comparing Xienen's stats and http://joostdevblog.blogspot.com/2011/10/proun-is-big-success-pay-what-you-want.html it seems that $20 is the maximum you can expect from a player (he got 2 players paying $50 while total earnings were $19k, Xienen got also 2 who paid $50 while the earings seems much smaller, it seems like the amount of $50 is fixed and related to friends than normal fans, anyway, in both cases there is no point expecting anything above $20 in any reasonable quantity).

Total revenue breakthrough (Proun):
$2-2.99 price point - $4k
$5-6 price point - $9k (he said the $6 was from added tax, so we could count it as $5 I guess).
$10 price point - $6k
$20 price point - $1k

Surprisingly there were basicly no $1 price point sales... Seem as if for those who actually decided to pay $2 or $3 was the lowest. I wonder if this proves something...


Overall, I would speculate that $1 minimum is bad for 'pay what you want' model. PayPal has around $0.50 fixed fee, so $2 is around 3 times more than $1 for the developer (not relevant if it's sold via portals, but these don't allow this model so it is again irrelevant). Also it might seem that $2 is still seen as a very low price by the player.
On the other hand "just $1" sounds better.

I know this is not known topic and there are no experts on this kind of subject. I don't expect brilliant answers/ideas. But you could try posting how would you solve it. Your solution don't need to solve it perfectly, if it is better than the one posted previously it's more than enough

One thing, don't forget that the purpose is not to make the best AI but playable AI. The player is supposed to be the smart one, he is supposed to outsmart the AI and it is the AI who is supposed to have more units and still lose. If you upgrade the AI to an extend when the player need more units to win against AI it will result in a bad gaming experience (I'm the idiot and computer is the smart one).
Do not turn your game into chess, where no human on this planet can win over AI.

Prepare the base version on Windows, test it for a year. Then compile on Mac and that's it. I plan to use precompiled binaries of SDL, this libary has been here for almost 20 years already, I doubt it could cause problems and the game code has been already tested on Windows... Bugs could arise if there was a difference in behaviour of Windows and Mac, and since it is both C++ and SDL and even the same CPU I can't see problems other than upon compilation/linking stage here.

So, I have basicly 2 options, to acquire Mac or use emulator? There are no affordable companies providing this service, or some software that let you compile/convert sources to Mac version directly on Windows (or Linux since I could acquire Linux easily), or some website converter (now I'm making this up, but hey, who knows, humans are known to invent crazy stuff all the time ) or something else?

I want only to compile it, testing will be done by testers and by all those who complained "why haven't you released Mac version" It would be nice if I could also run what I compiled to check if it does not crash immediatelly (performance is irrelevant), but that's not absolutely needed.
Also, the OpenGL would be extremelly simple (maximum GL1.3, nothing that was not supported by all hardware 10 years ago), I will be doing mostly pure 2D in SDL. Bugs related to porting are unlikely in such scenario...

There is a dungeon built with 3x3 rooms connected by corridors 1-2 tiles width. It's inhabited by several dozens of underground creatures that move around (so it's rather crampy environment). Let's say that each grid/tile can be occupied by only one creature at a time.

- Cheating is acceptable (like if two creatures block each other ways for too long they swap positions or they are allowed to violate the one creature per tile rule temporarily). But if it can be done without cheating it would be slightly better.
- I prefer if it is not too complex if possible, also I don't need a perfect solution (the units can appear a bit dumb and not always use the absolute optimal path, that's OK).
- Speed of the algorithm is of relatively low importance (but within reason).

How to organize their movement?

I think pathfinding probably should take into account only dungeon's shape (ignoring units/creatures), because it will end up in too many cases with "target impossible to reach" plus all these units move around all the time. Then when they encounter another unit they use some avoidance algorithm or use short range pathfinding which take into account both dungeon and units?

I'm also not sure how to define the target they want to reach. Units will frequently move to "rooms" and a room is 3x3, so not an exact coordinate. The most obvious would be to use the center of the room as the target, but if it is occupied by another unit it will be marked as not reachable which is not true since there are still 8 tiles left in the room.

BTW, the original Archon was not using A*, you could get away in such type of game using much more primitive algorithms. Like, move in the direction of the enemy, if obstacle try to go left or right.

Normally, nowadays we don't need to resort to such cheap techniques, but since you mentioned it's for Android and you had some performance problems, you might keep this as alternative method in case you are really starved for CPU power.

I found only these two posts.
http://forums.tigsource.com/index.php?topic=22080.0 (Pay what you want model : Proun report)
http://forums.tigsource.com/index.php?topic=12494.0 (Pay what you want : time limit?)

Quick question, a game developed on Windows: C++ & SDL & OpenGL (so extremely friendly to porting, no system dependent at all), how to port it to Mac?

Or even quickier question, do I need to buy a Mac to port a game or can I avoid it somehow?

Anyone tried it? Does it work? Is it worth it? Any articles about it? Anything.

To avoid confusion with donations, microtransactions, etc, here is the definition of "pay what you want"
Pay what you want - the product has no set price (althrough it always has some minimum price, at least $0.01, so you can't get it for free), you can pay any amount you see fit and then can download it.

Are you serious?
There is really no reason to develop a 3D game to WinXP these days. Not financial reason, at least.
Why not develop games for windows 98?
Also, DirectX9c got PIX as well... so...

Sigh... WinXP is the second most popular OS on this planet.

http://www.w3schools.com/browsers/browsers_os.asp