Why is SM3 dead?

[HeadHunter]

So what happened to SM3? Was it bad or what? Why it is dead?

[HeadHunter]

Anyone ever dares to comment on why SM3 is dead?

[Beherith]

I split this topic off map hardness, to dispel the myths surrounding sm3:

There are many issues with sm3, not limited to:

Fixed level of detail per texture phase
No blend+overlay filter
No proper overlay
Broken shadows
Broken reflections
Broken water appearence
Bad performance - especially above 3 texture phases and bump mapping.
No ability to paint on texture- thus only procedural texturing, difficult to draw metal patches, geo spots, roads etc, flow lines for erosion.
No free toolset (though the basic edition of world machine would still be good for it - theres nothing it cant do, but requires a procedural mindset)

Kloot's SSMF is superior in almost all aspects, Argh never released screenshots about how awesome he can get sm3 to look.

[PicassoCT]

I killed it. I admit it. Also, Behe is right. It sucked, and fell into the rift between high expectations and poor results.

[SeanHeron]

Hmm - according to your arguments Beherith, it was born dead, no? (since those issues seem to me to go way back). I would argue those are symptoms of the format being dead.
I think the ultimate reason is: It wasn't adopted by map makers (which I can hardly take issue with - given all the weaknesses Beherith pointed out), and therefore in turn, there was little incentives for engine devs to put more work into it. So a chicken and egg problem, I think.

Plenty of caveats to this (eg there were some maps made - but these all were more demo style, was my impression), and of course my reasoning altogether may lay wrong.
----
On an aside - which free tools for making SSMF maps are you talking about Beherith ? (ie SSMF maps of the quality you produce).

@ Picasso - Yes, you killed it. You as our resident avantgardist should have been the early adopter. Plus you had me waste half a day trying to compile the Busan map... (when it really is an easy one to put together as SM3).

[Beherith]

It sucked, and fell into the rift between high expectations and poor results.

100% true.

World machine Basic edition (free with the single limitation of allowing only 513*513 max exports) is more than powerful enough to do SSMF splatting.

Many of us did try to use sm3, like Forb, smoth and I, and our results were less than satisfactory. I spent a few weeks trying to take it to the max, which is when I determined its inviability because of the reasons shown below.

The only nice things sm3 had over SSMF was relatively smaller file sizes, no need for compiling and normal map support. But normal map support came at a performance price that could not be justified.

I challenge anyone to make a map in sm3 (dont mind the rendering bugs) that looks even half as nice as something released by Artturi.

In conclusion, sm3 is not only dead, it has been surpassed by SSMF.

[FLOZi]

IMO it should be removed

[Beherith]

I dont agree with removal, it doesnt occupy resources, but someone may see possibilities in it, and wish to rebuild it.

[PicassoCT]

NOOO, thats amputation, we never give up features. U never know when they might be of use one day. Just imagine, its the year 3000 and suddenly, they want to celebrate by throwing the old, well working mapping format out of the window and start from scratch (that always works so well in OS)..
No we got to carry it around, mumified maybee, shrunken to a skull, but it will stick around, till the weight of dead code brings the engine down.

[AF]

I dont agree with removal, it doesnt occupy resources, but someone may see possibilities in it, and wish to rebuild it.

Split to a branch and remove in master?

[Beherith]

Ah true, git has proper tools to manage this, then there should be no reason to keep it in master, except for the following:
http://springfiles.com/spring/spring-maps/sevenislands
This map has recieved 3000 minutes of play time in the past year, and is a nice 1.5 megabytes in size while the map is a whopping 32*32.

[oksnoop2]

I would like to make an SM3 map for ct. It sounds like it's procedural system could work well for what i have to do. My vote would be for it to remain.

[SeanHeron]

Same here. My vote (not that one is being held I don't think..) is for keep.

[FLOZi]

But it's still broken and is unlikely to be fixed.

[hunterw]

update some tutorials, don't let any noob mappers get tricked in to using sm3

[PicassoCT]

I always thaught that was the purpose of a torturial - scaring away the newbournes?