Ok here is a list of just some general things I want to have in the rewrite of onlink. From ideas people have throw out in a topic but not made a topic for them.

1. Rewritten gateway system ie. (see gateway types list)
2. More types of missions (see mission type list)
3. More types of government servers and more types of servers (see server type list)
4. More upgrades (see upgrade list)
5. Misc (see Misc list)
6. Better Messages

Gateway list

1. Standard gateways
2. Satellites
3. Load balancing gateways
4. Gateway brands (like the apple brand or dell etc.)

Mission list

1. Hack satellites
2. Crash satellites
3. Disrupt Space missions
4. Disrupt Stock Exchange
5. Firestorm (look it up it is in movie, a hacker movie)
6. Nuclear Missions (launch a missile at this person, destroy a nuke on a pad etc.)
7. Misc Missions (hack a rover on mars, hack a government official's cell phone and copy all the data, etc.)
8. Keylogger Missions
9. Trojan Missions

Server types

1. Super computer
2. Watches
3. Cell phones
4. Load-balancing servers
5. FBI (lan)
6. CIA (lan)
7. British intelligence agency (lan)
8. Satellites
9. Space Stations

Upgrades

1. More CPU upgrades
2. More Hard drive upgrades
3. Operating System Upgrades
4. More Types of security systems (with missions to suite each type)

Misc

1. Ability to make a super computer out of multiple gateways
2. Ability to use other computer's resources (like the FBI central server's cpu or the FBI central server's ram)

I think that is all the ideas I have in one neat thread.

I'll update this thread with more if I get any.

If you want me to be more specific about some things like security systems (cause I know one of you is going to ask), just ask and I will tell you when I get some.

I want to see no Uplink, frankly.

Plus:

4 generations of gateways, delimited by board type:

1. PCB
2. Uplink's old blue board
3. Tempered glass/optical
4. Cloud System

A world that moves and breathes. This was the main issue I had with Uplink. Everything is generated for the sake of the end user - people, companies, news, missions, etc. There's nothing wrong with an expansive, "static" world between new games (ask any other RPG out there) - provided there's good replay value.

burningmace's list of special awards is good for that. As is discovering ever-increasing levels of security, crackers, quests, etc (Freelancer-style).

An online league. Posting scores, achievements and awards from Onlink to a scoreboard isn't impossible, and should encourage post-storyline playing.

Along with that: ingame hacking challenges. Modlink had this, but it faded.

Shit that goes wrong. Connections that drop for no reason. Machines that go down for restarts. Fires in buildings, earthquakes, disasters that render parts of the net inaccessible (remember the snapped cable in Egypt?).

Special hardware gifted from more powerful corporations as opposed to money rewards.

Several, different, slanted news/job board sources.

The ability to open multiple connections to the same server - and the requirement to have multiple connections in order to bypass certain security systems.

Workspaces and widgets.

Context menus.

A connection map that can go per-country.

A much more fragmented online world, brought to your attention through news, stock and infighting. The old uplink was basically, as far as the user was concerned, a static utopia.

More viruses, more hackers. People trying to break into your properties occasionally.

Cloud computing. Not just for research, but to use as your main method of hacking. Distributed processing slows you down, but a backtrace only takes out one node, and heightens public awareness (Singularity-style).

More vs. corporation missions where the target offers you a better deal for a counterhack - for instance, Arunmor and ARC, and where you choose which side you're on.

For people's PCs to be at the same place as their telephone. (VoIP)

Smaller companies should have their entire infrastructure running off their LAN.

Personal Computers should be the easiest to break into.

Email Managers/Directors trojans to steal passwords, backdoors into their personal machines, financial info.

Multiple concurrent plotlines.

OTHER HACKERS.

Freelance hacking job boards (you'll probably have to drop Uplink Corp as the sole provider anyway).

The inability to buy 90% of your software from any one source. Different companies specializing in different things. Some software stolen from companies, compiled, awarded for missions.

And so on.

ok I think you got my other 500 ideas in that one post.

Heheh, wouldn't surprise me. I've been thinking about them literally ever since I played the first edition of Uplink. Now that we're finally able to completely break away from that mold, I'd say it's about time

MULTIPLAYER

*cough*

Yes and no.

As awesome as multiplayer would be - is it really suitable for cerberus?

Ultimately, you'd want cerberus to be a slow, ponderous game, one that you could literally play for a month+. Multiplayer games (that I've seen) tend to be the fast-and-furious type.

Or let me put it like this: Would you invite friends over for a night of counterstrike, or a night of cerberus?

I trust I've made my point.

However - online play, now there's something else. Uplink wanted to go a little crazy with this. Have actual machines in the real world that you could hack into and do stuff. In their articles, they went as far as questioning the *legalities* (WTF?!) of an online play system.

But think of it in Freelancer terms. Freelancer built a single, massive, static universe. This was the same universe that you played in singleplayer, and in multiplayer, and it worked. It worked because it was such an intricately detailed ecosystem.

There's no reason a server can't be hosted online, with a static universe, that players can connect to. Or, go one better - peer-to-peer it, DEFCON-style.

This is all pie-in-the-sky, though, until (probably) the third edition of cerberus gets pushed out. In the meantime, though, an online ranking system is much easier to build, and will go a half-step towards true online play.

Some more ideas:

Have a real-world location, identity, and the ability to get arrested (a safety system for this comes later in the post)

Make it a lot harder for a beginner to get going. For instance, the brute-force password breaker. That really shouldn't be available to you until later, and you shouldn't be hacking into large companies right off the bat.

(In fact, it shouldn't exist as a program at all. There need to be alternatives to breaking down the password authentication barrier. And later, acquire a distributed brute force password breaker that runs off a network of compromised servers.)

Gateway - optional. In fact, there should be several gateway rental agencies, with fees that are too high for you to initially afford. As an alternative, you can rent temporary servers, proxies and shells to route through. Slow but safe.

For smaller jobs, have the ability to buy credentials off warez forums as opposed to banging the door down. Make the money back by creating a backdoor account and selling it.

Jobs: Remove the job board as much as possible. Several possible replacements:

* First-come first-serve forums with low-paying jobs.
* Direct referrals
* Watch the news. If a company loses data, offer it back to them.
* Proactively steal data, software and research and sell/auction it.

As you have companies and corporate alliances, you should have individual hackers and clans, along with reputation as well as a neuromancer rating.

If you're a strong member of an "immoral" clan, arrest isn't an issue, since they'll spring you out again. If you're a strong "moral" member, arrest isn't an issue either, since it'll probably turn into an official job offer, or a favor for freedom.

It's also through your clan that direct referrals for jobs become possible.

The concept of a home nation. Even by 2020, there will still be an intellectual cold war. Prove your skills, get recruited by your government, and get access to the highest levels.

In order for the home nation / multiple gateway agency idea to work, you'll most likely need a static universe. Score one for savefile size, too.

Multiple career paths? Via hacking you can do anything from law enforcement to organized crime to large-scale research.

Another small-detail thing: news. If it's a static universe with a static timeline, you can create a few thousand small articles, strung together in such a way that makes it look as if there's a real world out there. How many world events reach you today through a series of headlines and clippings?

In order for that to work, you'll need to remove the fastforward function, but given the amount of engagement required to play efficiently, that might not be that much of an issue. The only thing I ever used the FF for was waiting for new jobs/storyline advancements anyway. (or to see if I would get caught).

Ok, bed time. I'ts 4am, damn.

Hmm, by static universe do you mean Dwarf Fortress style, like each game is a continuation of events in the same world?

Hm... I like your static universe idea. When you install the game is when it generates your universe. Then, each agent enters the same world with the same servers in the same locations, and all the actions your previous (whether they have been disavowed or not) agents have done are still there. This would mean removing the fastforward function, since you could log off one agent and play as another in the same world, but I'm sure there would be some way around that. Hell, make hardware installed by robotics so that it's instant, or within like 30 seconds or whatever. This may not be possible, but it would definitely cut down on save time/size/complexity, while simultaneously allowing the world to become that much more complex and bigger, since it's generated many less times.

Of course, it's just a though.

"Static Universe" is when the design of the world is fixed. You'll always have the same companies with the same servers, IP addresses in the same locations, with the same security systems (upgraded over time, obviously).

Uplink used to generate a random universe from scratch every time you created a new profile. As hip and 19th century as that is, it does waste a lot of space, and I can't see where it adds any major game value. At least not now, when it's possible to add this much depth to the game.

For a game that employs the static universe effectively, try Digital Anvil's Freelancer. Or just about any RPG in existence today.

Additional advantage: Multiplayer. Since everyone's running the same world, it's not impossible to network the two.

@Soldier of Light: Not even then, I'd think. There's nothing wrong with having one universe for the entire game, for all players, like any other RPG. Think about it realistically - the specific IP addresses and physical locations of the servers you hack into and through aren't important, and neither are the phone numbers of specific people, to a very large extent.

yeah it would be a good idea.

Multiplayer with that would be nice but here is some things we would have to add in order to get multiplayer to work.

• Player has security systems and has to buy them himself.
• Player can hack into other players
• NPCs should be added to single player to attack the player as well
• Uplink should not be the only hacking corporation
• Players can create companies like uplink
• Players can issue missions
• NPCs are also in multiplayer
• Cerberus should be more complex (have more servers, more tpyes of systems, more missions, etc.
• In multiplayer you should not be able to use the speed changers
• The Player should be able to switch to multiplayer in a click of a button (not launch a whole different application like cod does)
• The Player can steal other people's software
• The player can make software for people and then complie it and sell it to add different things to the game (like they can make security systems, new gateways, new bypassers, etc. in game)
• There should be a black market or well many of them
• The player can work for the government
• The player can be an employee of companies like the government and make a salary
• The player can now trace people who are hacking into their system
• The player can trace people using a software tracer.
• The player can "call the police" on people who've hacked into their system.
• When the player goes to jail he has to wait a certain amount of time before he can go back and play again. (like one hour or 30 minutes)

I think those are the things we would need also.

didn't notice that before but your post says 10:13pm to me

But here is one down side of multiplayer I just realized it, in c++ is is much harder than you would think. It is doable but it is also hard for people who haven't done networking before.

I myself have only done getting some text off a webpage and that was getting hard already. And then for multiplayer you have to do that constantly without wasting too much bandwidth, cpu, etc. and it all adds up. In order to do it efficiently you have to be thoughtful about what you transfer and send which is a why we need a static universe for it. But then we also need to do other things with it. Here are the minimum things we need to transfer.

some of these things can be rpc calls but still it is alot

• time
• player locations
• news
• stocks
• Player's traces
• Player's money
• Player's gateway
• Slow traces for each player
• logs
• companies servers and everything on them (what software, data, etc.)

this is just a small list but everything transferred costs something in bandwidth and somewhat cpu.

now do this list (transfer every item in the list at once and looping it) and do it in-efficiently and you have a game that uses 2mb of bandwidth per second which is alot (you may not think it is alot but it really is). so you need to think about what is constantly changing and send that to all the players, but then the things that don't change constantly like bank money, firewalls being up or down, etc. make those rpg calls. Now do this all from the main central server to get a anti-cheat proof game. You need to send the players commands (keyboard mouse clicks) to a server and then the server validates them and if they are valid sends the result back to the player and then also to all the to the other players. But now here is another problem there is a thing called connection speed if your connection is slow for every click you might have to wait 5 milliseconds or more to get a result back. So, you need to add one more thing client side prediction, where the player sends the key commands to the server and then the client will also do what it thinks the server will do so that there is no lag, but then if the client is wrong the player has to be moved back to the state the server thinks that player should be in.

Now all of this has to be done and this is why multiplayer is hard in c++ (and programing langs in general) because you have to do all this, be mindful of your bandwidth and cpu usage, make a good client side prediction that is 99.9% of the time never wrong, make a server to receive all the data, and then send it back to the player in a reasonable time (the server must be a good server not a crap one (one with a pentium 3 1.0ghz cpu and 256mb of ram) ).

This is why multiplayer should be something we do but after we implement single player to where everyone likes it. Multiplayer shouldn't be something we add right away.

Edit:
opps, I did it again
I made a 3 post
in a row.
(sing that to the hits song from 2000 (I think) )

Edit: one more thing we need to do encryption

Encrypt all the data the player sends so that it is

1. not readable by the player to hack
2. not readable by anyone but the server

now you have to do this on the fly but make sure that data is secure.

this is what adds another thing to networking which is hard. You have to choose what to encrypt because encrypt takes up more bandwidth generally, encryption takes more cpu usage, encryption is only necessary for things the player should not be able to change outside of the game (like the money of the player but not the time, etc.). Now you really need client side prediction because the server has to also un-encrypt the data which takes more time and with more players it becomes even harder. Now if you have 5 players on one server it is fine but (here is a scenario for you)

15 players can join a server which doesn't use encryption

how many can join with encryption and keep the same time (amount of lag etc.)

Well lets find out

First we need to decide on an encryption method.

md5? pretty quick but not hard to crack.

md5 with some salt? pretty quick but now the server can't crack it and neither can the player

sh-a2 512? not so quick to encrypt and pretty hard to decrypt

custom method? (the best solution) You need to make a custom method which changes all the character's around etc.

Now with are encryption being a custom method (a random name, cmpt) we can go on to the server side and deal with time,

so how long does it take to encrypt some data in our custom encryption method?

lets say for practical purposes 1 millisecond.

so how long to decrypt? let's take the same amount of time 1 millisecond. (now here is where it gets interesting)

15 players on one server unencrypted

they now all take the same time to connect and receive data 15 milliseconds

now add the encryption and inorder to get the same time to receive data you can only support 13 players (add 2 milliseconds to each player, client side encryption and server side encryption)

now the more players you want you need to do one of 2 things (I think)

A. Get a faster server with a faster connection
Works to a point but then lets say 1500 people that needs to be one powerful server (a super computer most likely). Now that is really expensive

B.have multiple load balancing servers. (one in each continent, North America, Europe, Asia, etc.)
The best solution (I think).

The advantages

• Less time for a person in Asia to connect to the server in North America (because they only connect to the one in asia)
• Can support more people without a really expensive single computer

Those are the main two advantages

Now the disadvantages.

• Expensive (less than the super computer idea but still expensive)
• The servers now have to communicate with each other
• who is the master one which decides who is right? (see Section B)
• You now need a really fast connection for each server to each other

Section B:
Who is right?

now here is one solution to that, the worst solution.

Add all the data of the other servers that this server doesn't have and keep everything else I already have.

Advantages:
None

Disadvantages:
The servers will get out of sync pretty quickly and in the end the servers might be giving the player a negative time value or might just give the player his data he already has back.

now let's find a good solution

How about change all you have to what the other servers have? nope, one server will then become the master server and you lose your data about players in your area.

How about compare the servers and see which one is the best? best solution.

Now how do we go about and do this?

We take all the data from all the other servers and compare them and then on each point take the majority.

Time for example.

server 1 has the time being 10:13pm

server 2, 3, 4, 5, 6, 7 have it being 10:12pm.

Server 1 is wrong and so corrects itself to what the other servers say.

Now you might wonder what happens if the servers all have different things?

this should never happen if you write the code correctly but then again if you don't you can always have a web interface where a server operator can correct the values on all the servers.

Now don't you think this is kinda slow? well yes it actually is but if you ad one more thing into the game? server-side prediction. The server sends out what it has since it last contacted the other servers and goes with what is says but when it gets the other data from the servers if it finds it isn't in the majority on something it issues a correction to all of it's players.

problem solved.

Now any other problems with this way of networking? none I can think of.

But now you see how complex networking can get. You have a bunch of things to factor in, you have to be wise about what you make synced constantly and what is an rpc call, and you have to worry about hacking. Now to make this all work good you have to be efficient which is hard but doable. Now you guys hopefully know what is involved with networking and now understand why it shouldn't be done right away and made a sure feature.

some of this I bet is wrong in some views but hey it is my opinion and some facts.

correct me on any of this.

Also to the mods and admins you can if you feel it is necessary move this into another topic (but I don't think it is).

Edit:

One more thing which I will not talk about much.

Bandwidth.

Everything costs something to transfer now do this will 1500 people and you have a huge amount of bandwidth, this is where you need to be efficient because bandwidth is really what can kill games, a person with a dsl connection like most have is 768kb which is slow and so the less you transfer the more people will play. This is where you can use compression but then again it takes more strain on the computers un-compressing and compressing files, but if the files are small you can do it quickly enough to make it work.

My lesson for the day. Multiplayer and Networking.

I should just note that we are in fact planning a static world generation that functions much like DF.

<3

I was wondering though - is there any way to make the game P2P? Each user that logs in becomes a node that's hackable, and all the nodes that are in the game world are players.

Instead of having a central server, the game state is maintained between all clients. (Not sure if that made sense, but yeah.)