Jump to content

Map Contest


23 replies to this topic

#21 warner2

    Member

  • PipPipPipPipPipPipPipPip
  • 1,101 posts

Posted 10 April 2015 - 08:35 AM

Burktross wins the thread for having actual knowledge on the subject.

#22 Surn

    Member

  • PipPipPipPipPipPipPipPip
  • CS 2024 Top 25
  • CS 2024 Top 25
  • 1,079 posts
  • Twitter: Link
  • LocationSan Diego

Posted 10 April 2015 - 08:56 AM

The map editor already exists. Just allow user map submissions for review.

#23 Quelldrogo

    Rookie

  • Bad Company
  • 6 posts

Posted 13 September 2015 - 05:46 PM

View Postjoelmuzz, on 10 April 2015 - 06:28 AM, said:

PGI cant really host user generated maps without legal hassles and dramas.

The only way to really do it is to release server program so users can do whatever they like on their own servers outside of PGI's responsibility.
Cant see them doing that though.

-----

C'mon... just limit the map editor to existing MWO asset libraries. If you need a new item not in the library, just make a black box with white text saying what the intended object is. Then allow players to fly thru the map in spectator mode and vote on it. Maps with the most votes get polished to final live version in the game. Designer is given a shiny legend mech with "mapmaker" quirks. Done.

It's got to take just as much effort building and releasing all these new mech packs. Now they are at the point wasting time rebalancing all these mechs they made. That's fine, the mechs are beautiful. Let the fans build some maps and it is a win-win.

I'd prefer a $15 monthly sub like EVE Online, but free to play is fine as long as there is some incentive to create new content. If all the dollars are chasing new mechs, we will continue to be starved of new maps.

#24 Vellron2005

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • The Blood-Eye
  • The Blood-Eye
  • 5,445 posts
  • LocationIn the mechbay, telling the techs to put extra LRM ammo on.

Posted 14 September 2015 - 02:20 AM

View PostEgomane, on 06 April 2015 - 01:45 AM, said:

That's not even close to the truth. It's a romantic look on the issue that is ignoring several facts.

Fact 1
User generated maps for the games of Valve are run on servers in control of the players. The user controling the servers weed out bad maps and add new ones when they come up. They constantly have to keep watch to not implement a map that is bad for one reason or another (graphical or gameplay usually, I'll list others in Fact 3). No user will ever control a PGI server. Even private matches do not allow access to server settings (or files) for very good reasons.

Fact 2
PGI would need to check every map entry for art assets that are modified or brought into them from the outside. They would need to check if the foreign art asset is actually new content created by the user who submitted the new map, or if it was imported from somewhere else. If they implement a map with a foreign art asset, they run the risk of being sued by the original creator, for using his work without approval.
They only way to remove that risk, is to release a closed map creator, that can only work with already existing map elements. This map creator needs to be programmed and patched constantly. They can't risk to implement maps with new content on the promise of its creator that he created it by himself.

Fact 3
As PGI needs to work with a closed map editor, the tools they can provide are probably far from being professional level. So every map entry needs to be checked for technical and performance problems. A map with a field with thousend or tens of thousends of beautiful but tiny flowers could look very well and might have a well thought through gameplay plan, but will perform horribly on most computers. There might be problems with collision detection or missing textures. Or maybe the map creator forgot to add starting positions for one or more mechs.
These are technical issues PGI needs to check for every single map entry. Possibly resulting in a change or not implementing a map at all.

Each single fact already is a lot of work. Add all of them together and you'll soon get a logistical nightmare.

Just look at the amount of complaints we already have for map collision problems. Maps created in house by PGI with their open toolset and testservers to run them on. Yet those problems still arise and sometimes massivly. Now imagine what would happen with a closed up map creator with no real option to test a map for such things or maybe even to access the point of the problem. And now imagine the complaints on the forum if those maps get implemented but not updated, because the creator no longer has access to it and would need to send in an updated version based on hearsay. If he is able to make the update at all.

No, sorry, the idea might sound good, and that's the reason why it brought up every so often, but it is not nearly as easy as it sounds and it is most certainly not free for PGI. Not even close to it.


I agree with most that is said, and yes, it would not be free for PGI..

But it would be much cheeper, and it would be MUCH LESS TIME CONSUMING!

So what if they need to check the best maps? So what if they need to fix some technical details on them? They DON'T need to make them from the ground up! So for the price, and timeframe of 1 map, we get, 5, 6, or 10? That would be GREAT.

A MapMaker utility with a decent how-to instructions manual could do wonders for this game...





1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users