Page 2 of 3
Posted: Mon Jan 04, 2010 12:43 am
by Svengali
JeffBTX wrote:Anyway ... Torus... when I approach the Tori from certain angles the "trench" along the outer edge of the "wheel" is full of the artifact. It looks like many triangles that appear and disappear rapidly between frames (30 - 60 FPS or whatever...
Probably z-fighting, then.
Ahruman wrote:It happens when two polygons overlap without sufficient space between them. The workaround is to cut out the corresponding piece of the underlying model.
And exactly this happened with the Tori stations .-)
Posted: Mon Jan 04, 2010 3:31 am
by JazHaz
JeffBTX wrote:JazHaz wrote:
The new stations do display very well, however with some of the stations as they rotate, you can sometimes see through the stations to see the other side and see the edges rotating in the opposite direction (should be hidden).
JazHaz... NO you shouldn't be seeing that with PAGroove... I didn't see ANYTHING wrong with PAGroove, or GrittyCoriolis or Griffs Trade Outpost. It sounds like MAYBE its a driver problem, at a guess. Hopefully you can fixit, or live with it, or maybe your card ISNT fully shader ready.
OK here is a screenshot illustrating my problem. The station looks transparent!
Posted: Mon Jan 04, 2010 4:24 am
by JeffBTX
JazHaz; That's the "Solar..thingy..Variant" from PAGroove, and it is NOT supposed to look that way.
I assume some kind of shader problem. Maybe your card does not fully support shaders... maybe a DirectX or a driver problem. You can try updating DirectX (depending on your OS, though) and your VidCard driver. Try the VidCard driver first.
I suppose also that the OXP got corrupted somehow, or that there is a conflict with other OXP(s), but I doubt it.
Posted: Mon Jan 04, 2010 10:33 am
by JazHaz
JeffBTX wrote:I assume some kind of shader problem. Maybe your card does not fully support shaders... maybe a DirectX or a driver problem. You can try updating DirectX (depending on your OS, though) and your VidCard driver. Try the VidCard driver first.
First thing is that my system
doesn't at all support shaders!
DirectX surely has nothing to do with Oolite, as it uses OpenGL?
And my PC doesn't have a separate video card, its a laptop with integrated graphics!
Posted: Mon Jan 04, 2010 6:02 pm
by JeffBTX
Hmmm.....
I would suggest that if your LapTop does NOT support shaders, then this is a side effect of that, then...
... I don't know enough about them. I USED to code a lot, but I let my skills start slipping at about ~ DirectX 6 or so.
Posted: Fri Jan 08, 2010 8:27 am
by JeffBTX
For the Globe Stations OXP (globestation2.0.oxp), the first of two station types has this entry:
Code: Select all
globestation = {
model = "globestation_core.dat";
name = "Globe Station II";
port_radius = 500;
roles = "icosahedron(0.5) station(0.001) globestation";
What is the (0.001) for? The base probability is 0.5, but the SECOND value of 0.001?
I looked the the sections in the wiki on coding... the format just doesn't make sense, in THIS case?
Posted: Fri Jan 08, 2010 9:03 am
by Cmdr James
I tink it means that when oolite wants to create a ship with the role 'station' the globestation is weighted very low so it is possible, but unlikely that it will be chosen.
Posted: Fri Jan 08, 2010 11:19 am
by JeffBTX
CMDR James,
Thanks - it seems simpler just to use
roles = "icosahedron(0.5) station globestation";, or icosahedron(0.25) or icosahedron(0.001).
Posted: Fri Jan 08, 2010 11:39 am
by Svengali
JeffBTX wrote:What is the (0.001) for? The base probability is 0.5, but the SECOND value of 0.001?
I looked the the sections in the wiki on coding... the format just doesn't make sense, in THIS case?
It is there to help Oolite to see that this is a station - a safety mechanism .-)
Posted: Fri Jan 08, 2010 11:42 am
by JeffBTX
Okay, but why not simply
Code: Select all
roles = "icosahedron(0.001) station globestation";
instead of
Code: Select all
roles = "icosahedron(0.5) station(0.001) globestation";
Posted: Fri Jan 08, 2010 11:47 am
by Cmdr James
You have to understand how oolite choses the ships to create. It is asked to generate a ship based on a role, and it then selects from all the ships with that role acording to weighting.
So if you ask oolite for a station, it may chose globestation. In your code that is likely, in
Code: Select all
roles = "icosahedron(0.5) station(0.001) globestation";
it is relatively unlikely.
If you ask for a globestation then it is likely (in fact unless there are other ships with that role you will certainly get it).
To clarfiy, the globesetation has 3 roles and when requesting any of the 3, oolite may create a globestation. So each one has a probability set.
Posted: Fri Jan 08, 2010 11:52 am
by Svengali
Because 'station' without a probably defined, has a probability of 1. The Globestations are only meant to replace Icos and not other types and to avoid any confusion I gave it (0.001).
And Murgh (the author) has choosen icosahedron(0.5) to let room for other oxps AND the native stations. I think he has choosen it very careful - so I've taken this setting too.
The custom role globestation is another thingie and meant for scripters - and maybe I'll setup a planetinfo.plist too.
Posted: Fri Jan 08, 2010 12:04 pm
by JeffBTX
Okay... I'm still finding it hard to understand.
That *specific* format isn't in the wiki.
And the other "Station OXPs" that I've downloaded don't use that format; they simply specify (examples) "coriolas(0.25) station ..." or "dodocahedron(0.75) station...", there is no second number.
I've copied / pasted posts and snippets of code from the forum on this subject of mixing station OXPs and probabilities... I pretty much understand them... but so far the Globestation2 OXP is the *only* one I've seen that uses a second value like that.
Posted: Fri Jan 08, 2010 12:34 pm
by Cmdr James
The number is the probability (weighting) of the thing it immediately follows.
Code: Select all
roles = "icosahedron(1) station(2) globestation(3)";
So icosahedron has weighting 1, station weighting 2 and globestation 3.
Posted: Fri Jan 08, 2010 4:25 pm
by Svengali
Additionally to Cmdr James post.
JeffBTX wrote:That *specific* format isn't in the wiki.
When I've started a lot of stuff wasn't even mentioned or wrong on the WIKI, so it was often trial-and-error. Eric, Ahruman, the other devs and some other boardmembers have done a great job to update and complete/correct the docs. And even if there's still some work left, we have some pretty good docs right now.
But if you want to do oxps it is necessary to get a feeling for the relations between settings and this is (possibly) not a job for the WIKI. You can test these things fairly easy and if you're stuck with a problem the BB isn't far away .-) A lot of stuff has been explained here and using the search function (even in it's current form) and reading older topics (specially in the Expansion Pack forum) will bring up some 'pearls'.
JeffBTX wrote:And the other "Station OXPs" that I've downloaded don't use that format...
There're different ways to do the same, but it's necessary to know another piece of information. The probability is only one aspect, Oolite also needs a identifier that this entity is a station. This can be achieved by a few things: scan_class, roles or using 'station' in the entity-name (or combinations). For carriers there is additionally the is_carrier flag. In Globestations2 I've simply choosen the roles+entityname and in combination with the subent declaration Oolite knows that it's a dockable station.
And about the probability: Murgh has choosen the settings to make this type of station very rare. I know about pagrooves settings, but have decided to keep Murghs settings, because I think it's a good way to surprise the player. Finding a system with this station is a kind of 'Oh, I've explored something new' and has a big 'satisfaction factor' .-) I know that players with pagrooves stations installed will probably never see a Globestation, but I think that a probability of 10 is not the best approach, because other oxpers will give their creation then 100, 1000, 10000 to display their own stuff. That's why I've suggested to give station-oxps a planetinfo.plist (only a few systems) for specific stations.
Together with the probabilities it would be a way to avoid conflicts and to keep the 'pseudo-random' thingies too. And with Eric's suggestions of using conditions it would scatter things even more. So where should we go now (and sorry for too much words)?