Scmdraft 2 - 0.9.0 Preview - Page 4

Post a Reply

Prev 1 2 3 4 5 6 Next All

neobowman

Canada3324 Posts

July 07 2017 02:10 GMT

#61

Could you explain ramp vortexes real quick? And do the Korean mapmakers know?

JungleTerrain

Chile799 Posts

July 07 2017 02:43 GMT

#62

On July 07 2017 11:10 neobowman wrote:
Could you explain ramp vortexes real quick? And do the Korean mapmakers know?

As far as we know, we just discovered this, almost 20 years after the game's release! (Unless someone else has looked into this). I doubt (most) Korean mapmakers know about any of this but I don't know Korean so I have no way of finding out if this is true.

However, based on the posts on the ASL submission page about resource gathering "research", Korean mapmakers don't know more than we knew as of recently.

And correct me if I'm wrong freak, but vortices seem to be where pathfinding region nodes meet (the approximate "center" of the regions, which are almost always irregular shapes in most maps) BUT if the nodes are on top of impassable terrain, then a vortex is created, where units tend to get stuck and stack. This usually happens on ramps because of the small and irregular shapes they force pathfinding regions to take.

When it happens on a ramp, we can call it a ramp vortex. If it happens anywhere else, well it's just a vortex lol.

Again, correct me I'm from but that is my understanding as of now.

Also I am a bit confused still on the difference between the short and long pathfinding algorithm and their relation to the pathfinding region. As in, the pathfinding regions are used almost entirely for long pathfinding algorithm solutions, or do they also play a part in short pathfinding? It seems to be that they do have something to do with short pathfinding since units get stuck on these nodes.

And do units passing through a region "tend" towards the node as they pass through? As in, the node pulls on the unit to pass through it to get through the region or what?

Freakling

Germany1533 Posts

July 07 2017 07:07 GMT

#63

If you want to know all the details, I suggest you study the Open BW code closely...

Freakling

Germany1533 Posts

July 08 2017 00:04 GMT

#64

So here's a bug-fixed version of Circuit Breakers.

+ Show Spoiler [pictures] +

List of updates

Completely redid all the reversed ramps (using my own space ramps) to fix lots of terrain level bugs and adjust their width and angle to that of normal High Platform ramps.
Fixed the bugged geyser in the top right main (turned out a single cliff tile swap sufficed)
Cleaned up the rather messy terrain editing around the 3/9 o'clock expansions
Filled in a lots of potential drop holes around the map
cleaned up some other messy terrain editing

Things I haven't changed:

I kept all the low ground ramps as they are. I'd consider the kind of old-schoolish lack of reverse ramps a nice touch and the map accounts for it with a layout that does not really require reverse ramps to maintain good balance.
The top right natural geyser still mines very slowly. Not much that can be done about that, as it's just a bad geyser position.
Looks as dull as ever. I am not going to start decorating other people's maps...
I kept main ramp positions between top and bottom mains asymmetrical because changing it would have meant to make the gap between resource depot and ramp tighter for the bottom positions, which I think would be more of an issue. The original ramps were certainly put in their respective positions for exactly this reason.
Aside from the top right main geyser, no resources where changed, mostly because it was not necessary: Luckily all the relevant path-finding regions turned out either the same or even cleaner than before, so if there are any bugs with worker pathing they ought to have been there before.
That being said, some of the formations are really asymmetrical and there is certainly room for improvement (particularly with Comsat issues). So this might be something to look into for future updates.

neobowman

Canada3324 Posts

July 08 2017 04:58 GMT

#65

This sounds like the perfect time for you guys to make an article about this and see if someone can translate it to korean to post on one of the big Korean boards. This is actually pretty huge.

GTR

51599 Posts

July 08 2017 05:03 GMT

#66

For those still confused with the concept of a ramp vortex, here's a real-time example.
[image loading]

S.I.

58 Posts

July 08 2017 05:22 GMT

#67

On July 08 2017 14:03 GTR wrote:
For those still confused with the concept of a ramp vortex, here's a real-time example.
[image loading]

Thanks for finding an example =)

JungleTerrain

Chile799 Posts

July 08 2017 05:42 GMT

#68

On July 08 2017 14:22 S.I. wrote:

Show nested quote +

Thanks for finding an example =)

I've actually seen that before lol poor Bisu

Freakling

Germany1533 Posts

July 08 2017 09:21 GMT

#69

On July 08 2017 14:03 GTR wrote:
For those still confused with the concept of a ramp vortex, here's a real-time example.
[image loading]

I remember that gif. That's what I was looking for. Thanks.

Freakling

Germany1533 Posts

July 08 2017 09:23 GMT

#70

On July 08 2017 13:58 neobowman wrote:
This sounds like the perfect time for you guys to make an article about this and see if someone can translate it to korean to post on one of the big Korean boards. This is actually pretty huge.

I think I'll just post both updates on the Afreeca page.

Falling

Canada11516 Posts

July 13 2017 08:38 GMT

#71

Oh, I just saw this. I'm so glad someone's been updating ScmDraft. I just got back into making maps and the old ScmDraft had a weird bug on my computer for the last couple years that would crash whenever I zoomed out. This one doesn't so we're back in business

S.I.

58 Posts

July 13 2017 17:00 GMT

#72

On July 13 2017 17:38 Falling wrote:
Oh, I just saw this. I'm so glad someone's been updating ScmDraft. I just got back into making maps and the old ScmDraft had a weird bug on my computer for the last couple years that would crash whenever I zoomed out. This one doesn't so we're back in business

There were one or two buffer overflows in the old overlay rendering code which you may have run into. Most of the time I spent on the update was to fix / update / modernize / improve the code and not just add new features, and its much better now.

iopq

United States1084 Posts

July 16 2017 20:41 GMT

#73

On July 08 2017 09:04 Freakling wrote:
So here's a bug-fixed version of Circuit Breakers.

+ Show Spoiler [pictures] +

List of updates

Completely redid all the reversed ramps (using my own space ramps) to fix lots of terrain level bugs and adjust their width and angle to that of normal High Platform ramps.
Fixed the bugged geyser in the top right main (turned out a single cliff tile swap sufficed)
Cleaned up the rather messy terrain editing around the 3/9 o'clock expansions
Filled in a lots of potential drop holes around the map
cleaned up some other messy terrain editing

Things I haven't changed:

I kept all the low ground ramps as they are. I'd consider the kind of old-schoolish lack of reverse ramps a nice touch and the map accounts for it with a layout that does not really require reverse ramps to maintain good balance.
The top right natural geyser still mines very slowly. Not much that can be done about that, as it's just a bad geyser position.
Looks as dull as ever. I am not going to start decorating other people's maps...
I kept main ramp positions between top and bottom mains asymmetrical because changing it would have meant to make the gap between resource depot and ramp tighter for the bottom positions, which I think would be more of an issue. The original ramps were certainly put in their respective positions for exactly this reason.
Aside from the top right main geyser, no resources where changed, mostly because it was not necessary: Luckily all the relevant path-finding regions turned out either the same or even cleaner than before, so if there are any bugs with worker pathing they ought to have been there before.
That being said, some of the formations are really asymmetrical and there is certainly room for improvement (particularly with Comsat issues). So this might be something to look into for future updates.

can we get this version into ASL 4?

Freakling

Germany1533 Posts

July 16 2017 21:42 GMT

#74

Maybe.

The point is: I am not quite sure I should call this an official update yet. There are some things that could still be improved upon from the old version, like some worker pathfinding issues and very unequal mineral formations between mains.
Why stop half-way when now we have all the tools to efficiently solve all of these old issues?

Guess I'll be working on it a bit more...

In fact, I think I'll start a thread on getting all the old Kespa maps up-to-date soon. I am also have a quick-fixed version of Demon's Forest without stuck stack bugs (not that the map hadn't still tons of problems, overdoing a very restrictive mechanic being not the least of them), an update for Dante's Peak SE that fixes the ramp vortices and a ton of terrain level bugs and an update of Polaris Rhapsody that clears up the absolute mess that it's been terrain-level-wise.

If I knew which maps ASL4 is planning to use it would probably a good idea to get those a complete overhaul as well...

iopq

United States1084 Posts

July 17 2017 07:57 GMT

#75

I would probably fix the "unmined minerals" where workers don't automatically transfer over to the next patch because it's off to the side too far. I would also just completely move the natural gasses so that they mine at a better rate.

Freakling

Germany1533 Posts

July 17 2017 11:15 GMT

#76

On July 17 2017 16:57 iopq wrote:
I would probably fix the "unmined minerals" where workers don't automatically transfer over to the next patch because it's off to the side too far.

What exactly are you referring to here?
There is a bug where minerals are placed centre on unwalkable ground, which makes them effectively unmineable (except through click spam – if you are lucky). This is so serious that I doubt any map actually has a bug like this.

So I am not quite sure what you mean. The mineral formations at the 3/9 and 12/6 o'clock expansions? The lack of automatic migration between mineral clusters spread like that is pretty much a feature, not a bug. It's on the players to optimize their mining rate by splitting their workers into two equal groups. The benefit of this kind of arrangement is that less worker migration also means more efficient mining at higher saturations.

I would also just completely move the natural gasses so that they mine at a better rate.

This would change other core aspects: how Vulnerable the gas is to being sniped, how much it gets in the way of units and buildings, how the mineral formations are placed (to make room for the geysers), and hence how vulnerable they are to Muta harassment ... and so forth. In short: This is a change that could impact map balance negatively in lots of other ways, so I'd rather not do it.

iopq

United States1084 Posts

July 17 2017 11:23 GMT

#77

http://screenshot.sh/mKS5Oiv5CHSzE

eight minutes later

http://screenshot.sh/odCiTBWtokQbE

is this a feature? I think you shouldn't have to manually put workers on each patch when you take one off to make a pylon or whatever. That's not a feature in FS, for example, where workers will transfer to the next mineral smoothly over the course of the game.

Freakling

Germany1533 Posts

July 17 2017 12:27 GMT

#78

Is that consistently happening in every game?

I don't see any particular reason why that one patch would mine so much slower.

iopq

United States1084 Posts

July 17 2017 13:19 GMT

#79

Good question, I couldn't replicate it perfectly myself. But workers do tend to mine the patches on the left instead of that one if you send them to the neighboring patch. I'm guessing Bonyth took the only probe mining that patch to make a pylon so no other probes mined it.

Freakling

Germany1533 Posts

July 17 2017 13:50 GMT

#80

Well, how many probes were mining there to begin with? If it were just 8, or maybe 9/10 you cannot really expect them to migrate to any one particular patch automatically, as there are always other free patches available closer by. Often you will see something like three Probes only alternating between two patches in a mostly stable pattern which will take a long time to be broken.

Prev 1 2 3 4 5 6 Next All

Please or register to reply.

Scmdraft 2 - 0.9.0 Preview - Page 4

Completed

Ongoing

Upcoming