Differenciated Radar Blips and The General Approach to Intel

huh, then we pretty much agree, I didn't suggest t1 and t2 differentiation, I don't agree so much with that. I do think however that naval should be told apart from land.

this is pretty much still up for debate.

Not really.

Everyone agrees.

It's simply not feature complete.

It's tatsujb's thread necro time!

...

Time to necro - once again.

Not only should radar blips remain identified, but should also leave behind visible blips when loosing track + merging these leftovers again when either the unit is identified again somwhere else or when the "ghost's" position is scouted again.

I mean, that would even be possible using just the client already, but it's not nice since it wouldn't integrate with the Chronocam at all. Ghosts and ID flag are something which must be handled serverside.

........ : D !!!!!

I DIDN'T DO IT THIS TIME ! SEE ?!

thanks @exterminans !

I fully support this, except for naval needing to be different to land, i mean it's on the sea isn't it, wheres the confusion.

there : https://forums.uberent.com/threads/...al-approach-to-intel.49424/page-2#post-764593

Am I the only one who can pretty much identify every unit just by the movingspeed of the blibs?

No, you are not.
As long as it is moving.

But it's utter nonsense, that strat icons for identified units drop back to anonymous blips when leaving vision, the lack of radar ghosts when loosing track, and as you noticed the lack of visualizations of obvious characteristics.

These are all information the client already has (at least partly) and which are omitted ... for what reason are they omitted again?

Terrain type + speed does give a good indication imo.

I wonder if it's deliberate to *increase the importance of orbital* as the orbital radar stuff gives you vision as well as radar...?

Yeah I guess you're right. I've never played SC1/2 and never played a RTS with that mechanic in game. I am not used to it so I don't need it. (As I think the radar is allready Super Mega powerfull). But I can see your desires for it.

That would be a jackass move. Crippling (and not even in a consequent way!) the radar output just to promote vision?

But if that was the intention, why would they still allow you to see individual units by radar?

yeah but why the hell is it your job to memorise that + on a spherical map?!!!

I much rather be strategising. And scouting much more! I would if the data I obtained I actually kept. I'm sorry but I have **** memory and having to handle a bagillion things doesn't help trying to remember that one structure's location and nature.

I was only speculating.... in all likelihood it's simply a case that Uber haven't implemented the features required for it to do this. It sounds simple, but when you think about it there are quite a few things that need to be tracked for it to work properly:

Unit in radar range (no been seen visually) = standard 'dot' as now.
Unit visible = visible / strategic icon (as now)
Unit that *was visible to player* but now out in radar range = unit type specific radar dot or strategic icon (?)
Unit that *was visible to player* then tracked on radar but left the radar range should revert to 'dot'.

That last bit may actually be quite difficult. Also you need to bear in mind the visibility rules need to be controlled separately for each team individually. That's actually quite a lot of additional checking that is required so probably isn't that trivial.

Nothing difficult about that. It's a simple state machine with 4 states which can be encoded at the cost of 1 byte per player. One of the most basic concepts in computer science.

The tests are being performed already, so all it would require was to store that state machine per player.
3 possible inputs: ["invisible", "radar", "vision"]
4 states: ["invisible", "radar, but never visible", "visible", "radar and was visible once"]
And a 3x4 table which maps the new state for each possible input.

It's really that simple.

Ok... except that is 1 byte, per unit per player.

So if you have a game with 10,000 units and 40 players your actually creating 3mb of data. That would have to be done *per sim tick*.... It could start to add overhead quickly unless I'm missing something?

I agree the supreme commander way is better. Propably the best there is for now.
Units should have different symbols based on movement type they show using (a bit tricky for amphibic vehicles).
Units that you once got visual contact with, should continue to be shown with thier unit icon until they leave the radar coverage (even just briefly because you had a energy shortage).

At the very least different dot's for different types (building, vehicle, air, bot, naval) would help. I recently tried to aim a Holkins on factory using only radar blips and previous scouting information. And I had no idea wich ot the two very close points was the factory and wich the unit currently in production.

Of course I also see why that is difficulty to implement. Especialyl as the server has to keep track of it all.
I asume it comes around at it's due time.

You do
State transition can only happen when input occurs - that means when vision / radar is sampled at all, which wouldn't happen any more often than it does now.

And even then, the state has to change, or the Chronocam server won't even record it.

Which means that in your example the working set is increased by 3MB, but not all of that data is touched every frame, nor does it go over network.