So in 0.x, it would change the variable type to whatever I set it to at runtime? I had no idea. :/

Newt's suggestion of a fourth type is a good one, I think. Also, please leave hash and array the way they are - having to manually set variable types for those would be a lot of extra steps.

Hey, nice to have another admin about.

Been thinking about this some more. I think there's a perception problem with the subscription/2 years of free updates model, and perception is important for the customer. I think these might be too far from what people are used to. Perhaps a somewhat more traditional approach might work.

Traditional model - every so often a new version is released. New features for next version are worked on in secret, with perhaps some public glimpses of what's being worked on to hype the next version, when it's done everyone has to pay for the new version, new features are used to justify customers paying again.

Updated model - each year or two marks a new version of construct (again, perhaps with year numbers. If C3 is used for the name, it doesn't mean another rewrite, it is essentially C2.1, no different from the current development method, essentially C2 a year or so later. It is C3 in name only).

People can preorder the next version and get access to the 'beta', meaning access to the current versions. Finished versions are released, people can preorder the next to get the features as they're developed. Everyone owns their copy, no one is confused, people are hopefully comfortable.

Wolfire does something like this, with preorders getting access to the beta. It seems to work very well for them.

Tokinsom

Because C2's �no major versions, lots of little ones' development model makes that pricing model not work the same way, because people wouldn't need to pay for new versions.

I suppose every once in a while Scirra could say "HAY new version time, you all gotta pay to upgrade and get updates from this point on" which is basically what other companies do with a +1 to the program version number, but that's not really much different from the "buy it and get two years of updates" model.

I suppose they could have yearly version numbers... C2 2011?

I think some clarification is needed, because people seem to be confused.

I am more than happy to help support the program and the devs by paying, but a subscription model is a huge turn off for me.

But I am heartened that you are intent on finding a good solution. The idea of buying two years(or whatever time frame) of updates for whatever price you set is one that really interests me. Being able to actually own something I bought and be able to support you chaps too.

That's what the subscription is. A new word is definitely needed, because 'subscription' is apparently confusing and bothering people.

People are not going to try free version if features are being held back for along period of time.

Perhaps, perhaps not. Game maker permanently holds back features from the free version and people pay for it. People often pay/pay more to avoid having to wait - seeing movies in theaters instead of on tv, paying full price for a game, etc. Paying/paying more for new stuff is common.

As far as I know, the method I described hasn't been tried before. And while I respect your knowledge, there is no one who knows everything about business (since many business techniques haven't been tried) so something new might work better than you expect it to, especially with the sometimes seemingly illogical behavior of consumers that can be at times difficult to predict. That said, I might be wrong too, but it might be worth a shot. I mean, who knew people would pay hundreds of dollars for smurfberries?

#1: Usability: We need animations but we don't need the bells and whistles just yet. Just enough to create frames, animation speed, and something to set hot spots/image points. A full-bodied image editor can be created latter.

#2: Play Time: ONE flexible behavior such as custom movement, 8 way, or platform, as they tend to be the most popular and robust. One is just enough for people to start getting their hands dirty!

#3,4,5...: Functional tools: Global Variables, global objects, event groups, Intellisense/autocomplete for expressions...

This.

I think a lot of people are missing an important point here: C2 was written so that they won't ever have to make a C3. C2 should be well written enough that they can incrementally improve it without having to do any sort of major rewrite. As such, that makes the normal model of paying for a new version obsolete.

However, I think when used with something getter77 said, it might work. I like the idea of having paid members get access to all of the stuff that's in development for the next version.

Imagine this: each year, construct has a new version released for users of the free version that includes all of the newly developed features over the past year.

People who have paid get access to the features as they are developed. Users of the free version have to wait for the next free version.

The people using the free version will hear about all of the cool features in development and want to use them, so you get a constant cycle of enticing people to buy.

Also, I think the nag screen should display an ad. Not anything that you have to wait through, or at least no more than 5 seconds, but that way you can still get some revenue from users off the free version as well. The important thing is to keep the nag screen from being annoying enough to make people pirate C2.

I also like the ideas above about the second two years costing less than the initial purchase, and "choose your free version" since people can't agree on what they'd like from the free version - have a nag screen or splash screen, take your pick. The problem there is if what if they change their minds when they're done with a game? But it's an interesting idea.

I was hoping for the initial behaviors to be ones that had things that couldn't be easily done with events mainly. With custom movement behavior, it can do a lot of things other behaviors do. That one alone might be enough for a while, though if bullet and 8 directional movement are quick and easy to implement, they sound good too.

I hope you're not going to do all the behaviors at once before getting to some other stuff. I think many things like global variables, arrays and event groups are more important than most of the behaviors aside from custom movement.

My vote is animations first, custom movement behavior second, families third.

You can set the filtering on a per-layer basis via the layer's properties as well.

I think you underestimate how much work that would be. To make a venture like that successful would be a full time job for multiple people. They'd be so busy with it they wouldn't be able to spend much time developing construct. Besides, what Ashley said he wants to do is develop construct, not make a games portal. There are plenty of other came creation tools making enough money to support their developers, there should be some way for scirra to, too.