[Dev Blog] - Debugging the new AI enhancements


#1

http://stonehearth.net/2014/01/28/desktop-tuesday-debugging-the-new-ai-enhancements/


#2

While I realize you guys at Radiant are hard at work with the technical stuff (Stonehearth’s engine, Modding API, etc), could you give us some material that isn’t that technical, for us people who don’t understand software as well as you guys do? Perhaps some screenshots or a few concepts would be nice.


#3

in Radiant’s defense, many of us had been begging for something “meatier” for quite some time, as the vast majority of the initial DT’s were very visual in nature… still awesome, of course… just not much that really enabled us to look under the hood…

so, blame your resident n3rds… :wink:


#4

Aaah, good old “attempt to yield across C boundary”. I’ve had the same issue with RP until I’ve realised that coroutine itself already is protected, so wrapping it in a pcall is useless…


#5

I know you said we would have to wait until the bugs are fixed for the next release, but is there a chance that we might see a development build pushed out sooner via the opt-in beta (alpha? belpha?) feature on steam? It seem almost custom made for these kind of things.


#6

Actually this feature was already used in the past and we are “just” waiting for one certain fix at this moment. As far as I have understood the devs have already finished the coding-part. Now it is about to remove the bugs from this fix (not that it makes everything worse). So the actual build is not running flawlessly which means for me that there is no point in releasing the next patch without investing too much effort which can be spend differently. I think and hope that the debugging might not take that much longer anymore and we can expect the patch in the next few days (fingers crossed).


#7

yeah, as @voxel_pirate mentioned, this is a fairly significant update (the AI is obviously very pervasive) … once this is stable, we’ll see the major update pushed out, and then (most likely) see those smaller content updates roll out a bit more often…

well… then there’s saving… ok, so perhaps after that… :wink:


#8

I’m asking exactly because it was used already. It makes little sense to me to use a feature that is made to be used for fast iterations of updates, as compared to the slower main releases, only once. I for one wouldn’t mind getting new version “as-is” say, every Thursday, for the opt-in alpha, with the stable releases pushed out for everyone whenever the team feels it is ready.

If it crashes, so be it. You knew what you signed up for.


#9

The problem with that is the current state of the game.

[ KIND OF WORKING ]     [ KIND OF WORKING ]
                     ^
               YOU ARE HERE

If they released a version now, chances are it’s more broken than the current version. This would create lots of unnecessary bug reports about stuff they know is currently broken (or in development). It would do much more harm than good.


#10

Yea, it’d be kind of hard to test anything without a working AI, for example. :stuck_out_tongue:


#11

What I’ve seen so far is that you must break a game to fix it. Everything is connected so adding stuff messes up something else and then it needs to be fixed and is likely more buggy than the first release.
Also I don’t think many casuals are following the game this early so they have very little reason to dumb down their explanations.

I don’t understand any of it but its pretty and theres lots of stuff in pictures so I like it.