pouët.net

Alone_Coder: this is what I meant in the other thread about seasoned coders putting their reputation on the line by making claims. The claim comes with a definition of what you're seeing, so the one making the claim must know what he's seeing. Here is another demo with texture mapping, there's also Voyager and a very bragging Brainstorm demo around this time (ah.).

You have linked two demos without texture mapping. That is to say, the claim saying that demo x did texture mapping first will not link to one of yours, but to a demo that solved the problem of doing texture mapping as we know it. Your two demos didn't solve that problem, but may come under a different claim with a different definition.

I've already shared my thoughts on framerate/size/quality.

By the same token there might be a first stencil vector claim and all the way up to perspective-correct(ed) texture mapping, and same but with distance filtering. One example of why claims should be moderated is the Real-time Raytracing claim on the C64 list of effects mentioned in the other thread. It is up to seasoned C64 coders to decide whether the demo in the claim solved the problem. There will be wailing and gnashing of teeth regarding the accuracy of the claims and definitions ;) so a system is needed, to avoid endless back and forth with loose definitions.

Yes, Cube-O-Matic seems to be earlier. But what do you mean by "not solving the problem of doing texture mapping"? I didn't link stencil vector effects, they are special case.

Having claims for each platform separately is crucial because, for example, texture mapping in the terms of "whole computer world" was made back in 1987: US4935879 patent. So, we need "first Amiga tmap", "first PC tmap", "first C64 tmap" and so on.

First ZX Spectrum tmap seems to be from august 1997: Spirius.

Not the same thing, one is fractal zoom and the other is feedback zoom effect.

Maybe "chaos zoomer" will be the correct name.

yes, separated between fractal zoom and chaos zoom. also added linear vs perspective texture map in the list. i just added some demos on some terms just to fill in. they're probably wrong anyway. here's an update

Alone_Coder: The list will include all platforms, so one can see where something was invented first in which platform and also on every other platform. That's the idea

Yes, chaos zoomer != fractal zoomer. And the first chaos zoomer still wasn't in Bananamen but the 40k at The Party 92 from Tizzy/Shining (although was slow and small, Dweezil was the first who did it properly in Bananamen)

ive accepted the Shining's. it looks like a chaos zoom to me.

Rax (released in June 1987) was probably the first demo to have "line vectors" (real-time 3d wireframe graphics) on PC. Probably also the first to have a 3d starfield effect on PC.

(that's around 3 years after Elite, but apparently the same year as the CGA/PC port of the game...)

chaos zoomer is known as Kumpa. In some demos (Pro.s.i.a.k.) you may see "Fractal zoomrotator"

This is weird. Just weird.

superplek: I'd imagine OCD listwriters to be quite common in our geeky environment. Gotta sort and categorize the world! ;)

@tomaes, this one came earlier: Microsoft Flight Simulator v1.0 (1982)

That's interesting: bibliography on texture mapping and image warping - the first scanline technique for texture mapping published in 1980, bump mapping in 1978, both in SIGGRAPH Proceedings.

Alone_Coder: Not a demo though. RT vector graphics as a concept is obviously much older (1950s).

Yes, there will be problems naming the effects, even for seasoned coders, if the author doesn't give one.

Also, some effects are obviously not possible on platform x, and so the author may be fibbing! ;) After all, he was just doing his best to fool us into thinking the illusion of effect x is effect x. That's what he's supposed to do! So a little "modified truth" or "not telling the whole story about the effect's limitations or weaknesses" is OK - in the demo itself that is. :)

At the same time, as mentioned, most effects are possible on any platform, and so the only difference to *demo* effects is that the demoscene has standards, such as decent framerate/size/quality. "Just writing the code" is sometimes very simple, let's say texture mapping in a floating-point BASIC, and while it "solves the problem", it's not much of an achievement, is it?

Since the best effects are these illusions, and making those illusions is instead the "problem to solve", I think the database or list should stick to only the effects invented within the demoscene. Lots and lots of rendering techniques were published in papers from the 1940s to the 1970s, and usually the hardware to run the algorithms were built to order, which makes it not the same arena or task.

Some of the illusions were possible from skipping steps ("correctness" quality), special-casing, precalc, prerender, baked tables, etc. This must be judged and separate the effect from the same effect using only generic tables, for example. "It looks like effect x" is not enough. When you've found out what it is you're seeing, you a connection to the general effect family can be made, and specifics used to give it a leaf on that effect tree branch. Some effects will be leaves of multiple branches.

Mapping out this tree might be fun for some, but it could be simplified to a list of *demoscene* effects, simply with two "required" fields to fill in by the claim-maker to motivate it being on the simplified list over others: Achievement and why it's an achievement (difference from similar effects or hardware/complexity/historical context).

If a list is chosen, it will be a "hall of fame" of few entries and is open to subjective ad-hoc motivations, so I would at least suggest an edit history. Suitable for a Wiki with seasoned coders as editors.

A tree would be much more complete and would allow browsing and understanding a fuller landscape of effects. It must be as strictly implemented as possible and not allow personal opinion. It's just for marking up "different enough" effects. So leaves must be pruned with a "strict but common sense" approach. It should not use categories but relations. Some fuzziness must be applied when creating "based on this effect" relations - for some demos it will be incorrect to make the relation but also incorrect to not make it. It might be better to have types of relations, such as "based on", "visually similar to" and "inspired by".

Should non-general-table-based or otherwise non-real-time effects be excluded, or at least from claims of firsts? It could be too exclusive for some 8-bit platforms if we excluded them, but if we do, adjective tags must be added, such as baked-table, 1-axis, no-perspective, no-sorting (and so on). Any demos with zero non-realtime tags might be considered the "real-time first" (for zero special-casing tags, "first complete implementation") and may appear before or after non-realtime entries, but I think that would be OK. This would also explain effects that are non-realtime but have gotten established effect names. The non-realtime elements of the effect that are generally tolerated on platform x would be dispassionately analyzed by the tags and get some background as bonus.

With the ambitious tree, it may be possible to map out branches so that it's easy to spot gaps within branches and make a new effect to fill that gap. Or even disregard all branches, imagine the excitement when someone release a demo to make a brand new branch on the family tree of effects! \o/

A list might be good enough. But since platforms are different I think there should be a list or tree for each platform. The lists or trees can then be compared against each other.

cheating is allowed. if anyone would like to make such a system, with branches then be my guest. i'll update this text-file because its just easy to do.wether its table based or not doesnt make any difference actually. effects are visual, they show on screen, and back in the early days, one didnt have enough ram or processor power to actually animate something anyway. let's just keep it simple.

Random nitpicking: the first texture mapping (not stencil!) on Amiga I know of is this. Released over a year before Cube-O-Matic. There might still be something out there I've forgotten or missed though.

Also, thumbs up to Photon from bringing reason and facts to the conversation :)

@break: Wow! O.o This dinosaur needs youtubing! :)

@Photon: maybe game implementations should be included, too, because they were made in the same or worse environment as demos and they were even supposed to solve the problem in USABLE WAY (so, no cheating, and little more than 1 FPS :)). The nature of demos "world firsts" is to show what was never done before, incl. in games.

im not including game-tech in this list. not even wolfenstein. if someone else wants to do that be my guest. mixing demos with games are bad idea just because it will be even more hard to maintain than what the list allready is.

the fact that it's such a hard to maintain and above all subjective list should give you a clue about what you're doing

or rather, should not be doing :-)

you should give me a clue, if you know what you're saying