tjohnman / knoh Goto Github PK

Hey cool, great idea! Thanks a lot for that.

Unfortunately I cannot run the binary:

not a valid Win32 app

I'm on Win XP Pro SP3, 32bit.

I promised to get VS up and look at it myself, and I will. Planned for the weekend.
But being able to just run a binary would be really great. It would also help to keep the hurdle low for others interested in this.

First of all (moving over from the zachtronics forum): again Hello @tjohnman, and Thank You!

I really appreciate your initiative, as well as your spirit ("fans can make). Therefore I'm pretty sure we'll be able to get up something interesting together.

As mentioned, I'm really not so much into C++. And of SFML I heard ... well, today. I have, however, heard of (and used) VS, although mostly C#.
My vision - and first step - was to re-implement it in JS, so it'd still be a browser game but not in Flash. Yes.

• So here then is my 1st q: What are you thinking right now? I mean what's your genuine, immediate reaction to this "C++ vs JS"?
  You know, not so much about where / whether we go. Rather that I'd like to get to know you. And I do know stupid prejudices amongst this or that camp, believe me. So please, tell me right away. Should there really be such an issue then well, let's put it up and solve it. Either way, I can already imagine several ways of doing so, including my getting into C++ & SFML, good reasons presumed.
• Next, let me ask what you're thinking of my proposal of applying temporal logic, particularly model checking. Have you done something re that or is it rather like me & C++?
• Another thing, although not so pressing atm, is how to relate to zachtronics. I mean the primary purpose of my posting in their forum was to actually get some statement from them, but none so far. I skimmed through the repo and noticed that you did make your own graphics. In my own (still private) attempt I just copied everything, although I think the user XP / user interface does deserve some major improvement. Nevertheless, staying compatible - even on the UI level - to some extent seems important to me. Anyways, having copied 1:1 (ie the graphics) is what kept me from making it public so far. I do, however, share your spirit of "totally open" (obviously?).
  So that's basically it: should this really be getting anywhere workable we have to have an official relationship / agreement with zachtronics about it.
• And finally - maybe anticipating the discussion a bit - have you thought of, or better: heard of, or even better: yet experienced the possibility of serious JS development? I mean unit testing, at the least. It does exist.
  Here's one more: in my view, JS relates more to Scheme or LISP than to Java or C(++) or what-have-you. How can you relate to that?
  So, my question really is: assuming you haven't already, is it conceivable for you to get into that business in order to build a truly amazing Kohctpryktop?

Ok, so this one's for getting to know each other, or set-up. It's meant to open discussion, and I do mean open. I can even imagine to have implementations in both, C++ and JS, and still benefit from each other, really.
Beyond that I'm looking forward to delving into details. Must refrain myself from dropping some :)

...are a good thing (TM).

Can I use them too, to label issues opened by me?

Being non-Russian as I am I'm missing any pun that might be there.
Transliteration yields "C?on" which ... as I'm looking at it ... does give me some idea.

But what did you have in mind?
Maybe drop a word about it in README.md?

What the heck, I just can't resist:

So far I've found out that it's base64 encoded deflated (RFC1950) binary data that more-or-less represents the design area cell by cell. That's the rough idea.

BUT: plz note that in fact things aren't (can't be) as easy as "cell by cell". That's because of the connections that can be in the metal layer or the silicon (P or N, which of them might be implicit), or both.
So one (stupid) way to encode all this into a one-cell representative datum would be to enumerate all possible combinations.
However, that
a) would mean a lot of redundancy (think: connected-to-right implies right neighbour to be connected-to-left and so on). Although, one might just say (ie have said): "let deflate do its job"...
and
b) does not match the _in_flated size of the data compared to what I calculated (actual is way less than what I think it needed to be)

Hence I guess there's rather two chunks of information:

• cell contents: metal or not, silicon or not and if so which, via or not, junction or not; like so (of course not all of the combinations)
• connections between cells: a grid very much like the former, only displaced "half a cell" and one less datum in each direction (hope that's clear...?)

Those two chunks might be either one-after-the-other or interleaved. To me it looks rather like the latter. At least at the moment...

Branching off of #1, particulary

I have no idea what temporal logic is. Sounds interesting, could you tell me more about it?

and

I'll have to think a bit more about how to best explain, you know. But I'll just try a start and hope that you'll tell me where I should go faster and where to add more detail.

For completeness: reposting some links:

• CTL (computation tree logic), implemented using
• BDDs (binary decision diagrams).

Remark: there's also linear time logic (as opposed to branching time as in CTL). And also logics with some sort of "watches", see below. However, we'll be focussing on CTL and BDDs.

Basically, it's about the verification part of the game, and the scoring. And it will allow for much improved feedback to the player.

That was kind of the foreword and the tagline; intro part below.

Let's think about and disseminate it.
Then do better.

See kongregate high-scores (or rather "low-scores"?).