I just saw this one on steam today, don't know if it's any good, but it looks cool: http://autotilegen.com/

In my opinion, this is a terrible idea. having a whole site generated in a canvas (regardless of it being made with construct or not) is pretty much the same as having a site made with flash: no search engine indexing, pre-load times, no browsing history, hard to maintain and update. And all that mostly for some animations that the user would rather skip to get directly to the content. It's 1998 all over again, when everyone was going crazy with flash websites.

When you create a new instance of an object in an event,this new instance become the default pick for that object within the same event, so you can just access the UID of the new instance like this:

On mouse click >> Create object of type Cube

>> Browser alert(Cube.UID) // will alert the UID of the cube you juste created

You can disable diagonal movements in the pathfinding properties. You you set the cell size properties to the same size as your graphics, you should get the result you want

When an object is pinned, you can get the UID of the object it is pinned to using the PinnedUID expression.

So when you want to destroy a monster, before destroying it, pick textbox by UID where the UID is equal to monster.Pin.PinnedUID. Destroy the textbox, then you can destroy the moster

Well, yes it is possible, but like any game that is more complex than a simple puzzle or arcade game, it will take some work. But technically, you already have everything you need: the most complexe part would be pathfinding and line of sight and those are already implemented in Construct (granted, line of sight in the perspective view of commando might be a bit trickier than a simple top view line of sight)

But, yeah, you won't do it in 5 minutes, but it can be done with some work

— Yes I know XDK an Ejecta are getting better everyday, but when I started working on this project, around october, cocoon was the only option. Since that's how it's implemented now, "if it ain't broke, don't fix fit"

Thank you guys. Like — said, I don't feel limited by C2 for PC. For mobile it's a different story. For instance, this adventure project would simply not work with cocoon or ejecta as it relies heavily on the XML plugin to load the game data. If fact it's kinda lame, the reason I went with a complex dynamic scene loading rather than having each scene stored in a different layout, was to get around the memory limitations of mobile devices. Turned out I introduced a new problem with XML.

Overall, I have two construct projects: a private one that I use to build the scenes and generate the xml file, and a public one, the engine, which uses the xml data to rebuild the scene.

I also created my own scripting system, also relying on xml. I just create my own script script in a text editor, then each action trigger a script. Scripts can be as simple as one line of dialogue, or more complex sequence of event, like when you try to look at the sky in the main room (character walk to window, says something, pause, turn around and finish his sentence)

The inventory works, I tested adding, removing objects and filling several "pages", the "up" and "down" arrows work. I just need to implement the actual "pick up" action, this is the next step

You can also take a look at my first prototype which was not an actual clone of the first monkey island but had an interface more like the one of Full Throttle and monkey island 3 (right click to bring up menu and inventory) : http://soletme.free.fr/PointNClick/

The particularity of this one is that it supported mouse/keyboard/controller/touch seamlessly, meaning you could start playing with the mouse as a regular point and click, then pick up your controller and move the player with the stick or use your keyboard arrows/awsd at any time without switching any option

I should get back to it sometimes next month, right now I'm working on my portfolio

I'm still working on my own adventure game engine in construct 2. I have not made any progress in a while because at the moment all my free time goes to updating my 3d portfolio, but so far I have been able to recreate the interface of the old lucasart games: http://soletme.free.fr/SCUMMClone/

So I believe, with some work, like everything, it can be done in construct 2. Sure, adventure game studio would probably be a better choice for what I'm trying to achieve, but I like construct better.

Blinx123

I did not use any toJSON or toXML built in function, I generated custom XML from scratch, by just running a big loop in which I concatenate text into a variable. For each element, I export its type (wall, object, door...), position, size, angle and any additional property. Then in the engine part of the game, I do the opposite on start of layout, I load the XML with AJAX and loop through the it and rebuild the elements.

For the path, I just used the built in pathfinding plugin. I just create walls for the areas you can't walk to. The walls are just invisible sprites with the solid behavior.

Blinx123 : No plugin or external program, I build the scenes in a separate construct2 project. When I run this project in a browser, it builds a big XML string using the scene data, which I then output to a textbox. I can then copy and paste it to an xml file. The actual game imports this XML file and parses it to rebuild the scene.

In all opinion, I don't know if it's worth the trouble. The main reason I did this was for memory management on mobile devices and to no have to edit a bunch of layouts whenever I decide to add or move a layer. But in the end I'm developing everything with only a couple of scenary, so when I finally move onto creating the actual content, the game engine itself is not likely to change much.

Psynaptik

I'm not quiet sure yet, this is the next big step for me, NPCs and dialogs. But I'll probably keep the XML approach, so I can tie it to my scripting system (since some dialog options won't just trigger text but also some actions). The harder part will be to keep track of stuff: Some dialog options are only available after certain actions, other can only happen once... Same goes with objects actually. It will require some kind of database

You can put the two objects in a Pair family and add an instance variable to this family, like "pairUID", in wich you store the paired object UID: when you pair object A to object B, set A.pairUID to B.UID and B.pairUID to A.UID.

Then, when you need to pick the two objects, just do use two conditions in a single event:

• A condition that pick object A
• B.pairUID = A.UID