f13.net

Latest live devblog has a crapload of exciting changes, some to do with balance, some new stuff.  Enjoy it, complete with unedited, amareutish, comedy-cutz beginning: listen to it here (http://myeve.eve-online.com/devblog.asp?a=blog&bid=547) or just download it directly from the site (http://ccp.vo.llnwd.net/o2/devblog/livedevblog11-pp.mp3).

Highlights

- Exploration to be pretty much totally rewritten, to be harder but less about dreary sitting around and waiting, and more about player (not character) skill.  Also, the ability to probe down most cloaked ships (except forcoverts, SBs, black ops, recons).  Unlucky, for the Chinese cloaking ravens.
- Some changes to salvaging.  I'll listen to this again but they want to make it less annoying, though I suspect that this can't mean much faster, since income-per-hour on salvaging is already pretty good.  More uses for salvage parts, too.
- Some improvements across the industrial side, mining in highsec etc but fuuuck knows what this Chronotus guy is saying.
- Changes to insurance (I hope they remove insurance payouts for self-destructing.  Sounds like it may be aimed at suicide ganks, though)
- More UI changes, supporting higher resolutions, less overlapping windows.  An end to "Windows 3.1 in space".  More abilities with key-bindings.
- Enhanced NPC AI, more variation in missions, enhanced smuggling and customs systems.  More Bounty hunting stuff, too, like tradable kill rights.
- More lowsec incentives, such as factional warfare, player sov of some sort etc.  Sounds fun for old UO-style pirates vs police type interactions.
- Assault frigates were going to get more speed and more dps.  Sensibly, they are not doing this now.  Looks like some resistance to webs, which will make for useful tackling capabilities up close.  Depends if they pair that with some nos/neut resistance, really.
- More useful ships that are easier to get into.  I like this one so much it gets bolded.  I like their idea of command frigates: The Groundskeeper now has Wing Command 4 and is moving towards Skirmish Warfare 5 and mindlinks for the end of next week.  It would be nice to be able to bring a small ship on Frigate Club ops rather than being the guy in the battlecruiser at the back.
- Capitals to be readdressed.  Titans and motherships to be redesigned as staging posts for the frontlines, with enhanced jump-cloning capabilities.  Tick tock.  Not only sounding like reduced offensive capabilities but also increased zerging possibilities.
- More nerfs to speed module stacking, especially poycarbs, which should be in just in time for my vagabond character :(

Edit:

- Local may be nerfed in some way to give less intel, either through delaying, making it constellation-wide or something.  Easier hunting?

Starting next month (perhaps), some of the big design ideas will be being published in dev blogs.

Good thing I can suicide gank in a cruiser with my new char, no insurance payout just means less competion, which means the targets will get complacent..and then BAM, ISKIES FOR ME!

I approve wholeheartedly of these changes.

Can the people you typically suicide gank actually GET more complacent?

Had a dude get away with 100+ BPOs in a punisher (I was in a torp raven, d'oh!), he could've been in a noob ship i suppose :P

Me too.  CCP are definitely on an upwards curve at the moment, with their design team.  I now really look forward to reading the devblogs.

About 15-20mins into it so far, so far everything sounds really great, I can't believe they havn't already converted all icons to vectorgraphs for scaling already.

The only problems are going to be:
a) how long it takes them to implement it (forever?)
b) How many tries it'll take to get it right (probably alot)
c) Amarr are still going to be pissed about lasers.

Edit: Changing local to constellation would be fucking awesome.

  :awesome_for_real:

Maybe they've been listening in to rookie chat while this Steam promotion has been running. The crappy UI probably loses them a ton of ppl who might actually sign up for a month or two if they didnt quit in total frustration 2 days in.

Yes, yes and a hundred times yes.  I was just coming here to post this.

I mean, now we have access to 0.0 space BAT people might think "fuck ratting becomes more dangerous".  That's true.  But we're not, primarily, PvE types: it makes stalking, ambushing and ganking way easier if this happens, and that means less time roaming pointlessly looking for ganks.

Plus, where you have a high quality intel channel, like TheCitadel, you have a bit of an edge.

Looks like they mention a lot of stuff they have been schmoozing about for a long time now. It has a lot of promise, let's see what happens in the next patches.

If they nerf polys without touching snakes they're tarded.  All the other stuff sounds great though.  I expect suicide ganks to increase as more dumb people haul billions around in T1 industrials thinking they are safe now.

Good stuff all around.

Im not sure I agree with this. In less populated constellations, its almost an early warning system when you see my -10 red ass jump in from 3 jumps away.  Fuck, for me I'd rather it just have local. This way I guess I'll have to try to catch people on constellation borders. I guess to me it sounds like ratting and mining in populated 0.0 becomes more dangerous, but in the drone regions you are gravy baby!

My guess is it will be better, but I like playing devil's advocate at times.

I think it makes lowsec and busy npc 0.0 space better for hunting in.  It'll really hurt the Chinese if they have to cloak up (or log out, with the probing change) whenever us honkeys roll into the same constellation.

I think that's a good thing for the gankee, though.  Currently, they only have a couple seconds TOPS to react to local spiking with reds, and I think that the constant concentration required for that is somewhat frustrating.  The constellation warning system will separate the smart ones from the dead ones, rather than the way it is now, where only the smart AND extremely nimble ones survive.

Plus, they may build in delays to the Constellation channel.  Or, even better, remove the pilots list.  Dunno.  We'll see.

Sounds like vaporware, for now, though.  All these promises, with only CCP's track record to bespeak for them.

I don't understand how getting rid of system local is a good thing.  If you don't have scanning skills, you have no way of knowing when there are enemies near you right?  You know they're in the same constellation, but they mght be on the complete other side of it.  Isn't that what local defense channels are for?

New changes / game mechanics that aren't mentioned above

system 'intel' revamp: possibly replacing/integrating the directional scanner, I suspect something along the lines of an 'overview' window in the solarsystem view that contains all that data.

Replacing / changing local: they don't like the idea of having the name of every person in the system automatically. Changing local to constellation would partially fix this (you wouldn't know WHICH system in the constellation the guy would be in), this ties into the cloaking mechanics below and the intel revamp above.

A baby rorqual for hisec mining to "give all the industrialists / logistics guys something else to do". Not the best way of putting it... people already know this one's coming so it's not news.

Salvage system: adding more things you can build with salvage parts outside rigs (Not mentioned, but drone parts perhaps?), changing salvaging to possibly require only one salvager per ship. They don't like the idea of a dedicated destroyer and they want to fix the mechanic to not require it. Some of the other ideas on this topic sound absolutely terrible, timeconsuming targeting minigames, this guy Chronotus or whatever shouldn't open his mouth until he's vetted his ideas with the other devs.

Doing something to make missions more interesting/challenging without being harder. Some mindless talk about intelligent NPCs, I suspect nothing of interest will be added to the game here for a long time. It strongly sounds like they are talking about 'randomizing' missions -- random spawns, random objectives, not just 30-40 or so specifically crafted missions. I suspect hilarity is on the way for this one.

lowsec revamp: they realize low-sec is generally higher risk than 0.0, and they want this reversed, they want more risk mitigation options for players who are willing to put their toe in the lowsec waters but aren't willing to go to 0.0.

Increased payouts for L5 missions to bring them more in line with what they should be.

Investigating making 'smuggling' profitable and other ways to lure people into lowsec

Bounty hunting revamp, giving a framework or system for players to trade/sell kill rights, other ways to make bounty hunting a viable profession (and lure people in to lowsec)

UI: They know it sucks. Changing the ship hud, replacing the huge 'locked targets' at the top, combining weapons into a single group / Fkey so you can have an alpha-strike. Moving to scaling vector graphics for customizable text sizes

They don't want to add heavy UI customization / scripting because they are afraid that people would be able to automate the game and would widen the divide between experienced players who would quickly add the latest addons and the newbies who would be stuck with the 'normal' one. Fuck CCP right in the ear on this one.

Lots of shortcuts - adding hotkeys for heat, different ammo types, among others. One of the devs would "Like to play only using the keyboard". Woot.

Changes to titans, possibly adding a special clone jump 'back' for titans that ignores the 1 per day jump limit (or possibly just add a 'back' for everyone? This would rock)

Forum changes -- splitting GTC / character forums into two forums shortly

More dev blogs; specifically community feedback on CCP game changes / directions and player feedback on them before they go live.

Recon/Blackops changes: Possibly adding a 'fuel' cargo bay for jumpfuel (as the cargo bay is too small to hold enough fuel for behind the lines maneuvers) and some slight buffs as they were released deliberately weak.

Assault Frigates changes: They had some changes ready, including AB/speed buffs, and they pulled it for future changes - changing roles to make them more useful, playing around with ideas such as webifier resistances and things to make them more useful in gang fleets and more than just a 'frigate plus'

Cloaking balance: non-covops cloaks are too powerful, they need to be balanced - either by removing local or some of those changes or allowing them to be probed out.

Removing system local has been loudly called for by experienced pvper's, and those who have the ability to form decent sized blobs, for a long time.  These people hold a lot of sway with CCP, and it looks like they are going to finally get their wish.  After it's implemented, those players in any place other than high sec will now have to:

* be in a hunting pack
* hide when they see something in constellation
* be willing to put up with nonstop scanning.

Expect to see even less carebears, new players, solo players, and even smallish gangs in low sec/0.0 after this is implemented.

It forces anyone in low sec to use scouts, which I guess CCP wants for two reasons:  incentive to not solo, and something that the newbie can do thus keeping them busy. 

I'd like to see the list of pilots in the system appear in an overview-like style (small text, multiple columns of info, no pics), rather than the clumsy way it is presented now, with those fat pictures and unsortable and in the local channel.  But, if they don't add that, we'll just have to scout for the info we don't have.

Basically, the change to the local channel from a system channel to a constellation channel moves 0.0 from being in favor of the defender (easy situational awareness, you know exactly what system the aggressors are not in, and that the total number of defenders available is unknown until there is a meeting engagement) to in favor of the aggressor (you know exactly what you are up against and if moderately fast moving are nearly impossible to find).

I doubt this idea makes it past the open testing phase as it's described here.  Additionally, adding a constellation channel along with the current local system would probably be far too helpful to defenders as it would make engagements entirely at the defenders discretion.

There is already a constellation channel, btw.

news to me.

New characters get it turned on by default, but I've never seen it on my older char.

Does every person in the Constellation get added to it automatically?  Because if not, it's not the same thing.

Well, you can close the constellation channel and thus leave it.  But if you have it on, as soon as you enter a constellation, everyone else in that constellation can see you in the channel (if they have it on).  Most people close it though.

Was this in the text or your between the lines reading? I listened to the devblog and all I got was changes being considered, not the direction of those changes. I was actually hoping they'd do something to rectify the insurance situation on T2 ships. I think both you and I are projecting our desires onto the devs' comments.

The only thing they said are "We are going to be examining insurance payouts. It's a system that hasn't been looked at since the start of the game and has been around for 5 years. We believe there may be some room for improvement".

When they think a system that has been around for 5 years with no complaints suddenly needs "improvement", all you need to do is look at jihadswarm to see the mpetus of that thought.

There are plenty of people who have complained over the years that the entire concept of the insurance system was idiotic, and predicted its abuse.  They were shouted down by some, but usually simply ignored.

Always some noobs gotta ruin it for us 'honest pirates'.

But unless they come up with something drastic, I'll just gank without insurance and target juicier stuff.

CCP has always been pretty reactive to 'divergent gameplay'. Whether that's a good thing or not largely depends on whether it was your particular quasi-legal niche that just got nerfed. It's also a major contributor to the tinfoil hattery regarding dev bias as a lot of the divergent gameplay comes from a small part of the community.

I wonder if the issue is more econimic than social.  Insurance payouts must be a massive isk fountain.  Insurance must also skew the economy because the price of ships and goods are determined by the markets but insurance payouts are static.

If Jihadswarm gets big enough CCP won't have any choice but to end suicide ganking. And that will spoil everyone's fun.

Some systems are going to become an utter crap for mining/ratting (namely every constellation with highway solarsystems). Lot of false warnings will make players ignore them and then die. Now you usually have at least 30 seconds to react - I have been mining on 2 hulk accounts for 6 hours straight and I still had good enough reaction time to the whole op intime. So it's not that of a burden.

It's also a boost for afk cloaking griefrs, now they can grief entire constellation, not just one solarsystem:D 

GoonFleet, destroying EVE one game system at time. :ye_gods:

But yeah, stuff like that is the only way to get CCP to actually do something. They're incredibly lazy fucks. After all insurance system have been screaming for t2 insurance fix, 0.0 insurance fix, BS insurance fix, pvp insurance fix, etc. for fucking five years.

You do realize that there are both players, and dev's, that not only are resisting those things but would like to see the entire insurance program removed from the game?  It's not a matter of dragging their feet, or being "lazy fucks", they don't know where they want to go yet.

It's unlikely that you'll ever find a Project/Product Manager on any project, anywhere, who will prioritize changes to an existing "working" system over adding new features or fixing "really bad" bugs.

Good manager would see "really bad bugs" as what they really are and anticipate their massive exploitation.

Sometimes fixing a 'working' system causes more problems than you'd ever be able to imagine. See: Star Wars Galaxies.

But I beg to differ, I do try to improve 'working' systems as much as possible - but coming up with an ROI that the execs understand can be very difficult.

Uh, motherfucking awesome comes to mind?  :drill:

Yeah, I would love to see them add in a way, necessarily complicated to do justice to what it means, that would make it harder for players with a bounty on their head to permanently reside inside a station. I also don't see the insurance system as a helpful thing. Yeah, it's nice to get some money back and all, but at the very least it should be a sliding scale paying out virtually nothing for the top ships, and paying the most to the bottom tier ships. This would help the new players out a lot, where the players capable of piloting BS and titans likely are not in need of financial propping up.

I like the idea of local going away, and also of pilot names not showing up regardless. It's too cheesy to be able to see that info automatically, AND it adds in some interesting new roles for Scouts... who would now be a more useful conduit of data for others since they can, through scanning and so on, know who's around.

But if they only do one thing, PLEASE God, let it be scalable text and a UI overhaul. I am so tired of sitting up and squinting at the screen to make out the text there.

Ah, but...  these new roles for scouts, they'll be neccessary, yes, but fun?  I think it'll amount to gatecamping.  You're not camping anything as a scout, you're just scanning.  Your miners are mining, boring but at least they can see some numbers go up every minute, and you just repeatedly hit the scan button trying to find reds.  And if you have a false alarm, they blame you for lost productivity.

I'd rather see them tank up haulers and mining vessels.  What transports have right now in terms of defense, the T1 indies should have.  And then go up from there.  But eh, a lot of you will disagree, and I won't argue anymore; CCP does whatever they want.

Trinity brought in quite a few new people, and the Steam thing brought some more, and it sounds like now they've all seen some of the flaws and are complaining/cancelling enough that the Devs see the need to do something.  Or maybe I'm just imagining things.

If they beef up t1 indies to what transport ships have now, should they just make transports and freighters invulnerable, then?  Haulers are supposed to be slow and weak, especially cheap ones that don't take any skills to fly.

As to the insurance thing, removing it from suicide ganking is good, removing it altogether, or in 0.0/pvp/etc?  People would riot, as alot of people couldn't afford to replace jack shit.  Battleships don't mean big money, hell it'd even hurt me if I didn't get insurance payouts on battleships, everyone would be roaming around in cruisers and BC's until they could fly HACs, which are better than battleships in every way, and without insurance, cheaper.

If they want to stop suiciding can't they just void insurance when Concord is involved and throw in a fine? It would probably be easier to make it economically less viable (or downright painful) instead of rebalancing ships.

They don't want to stop suiciding, though.  They may want to make it less profitable for any but the really well-planned gank - shift the profit-levels to make it less ubiquitous - but the nature of Eve includes risk.  I suspect that there is a big argument going on about Jihadswarm amongst the Eve devs right now: some will think it's gone too far, while others will be keen to see the pressure it places upon high-end mining ship owners to move to 0.0 or lowsec as part of a larger revamp.  I suspect that all will be delighted that it has virtually doubled the cost of ISK from the Chinese.

Regarding suiciding there is NO risk whatsoever being the aggressor, you lose absolutely nothing (except standing). No payout when being CONCORDED should have been put in game a long time ago.

I don't say this as I think high sec space should be more safe, just that only being a really juicy target should be risky, now you can get blown away for cargo worth very little.

I personally started running tech II fittings on my suiciders....just cause im a DPS whore and want to be able to take down transports...so there is some financial risk there. :)

and to counter the suiciders people start with CONCORD bombs: Sending in a newbie alt attacking to get CONCORD on the gate before hauling their goods...

Concord bombing is only good if you want to protect the belt in which you mine in high sec, since that is static. Its slightly less work to simply not go afk while hauling than getting an alt on every gate you want to use and have them shoot you as you afk by.

Concord bombing 101: Get your alt to shoot you at the belt you want to mine, concord will show up and hang around (for ten minutes?) to instapop every jihadi that opens fire on you. Rince and repeat.

There is a way to clear belts, too, though  :evil:

Concord sticks around until someone else gets concorded, so jihads could just concord bomb a station and then run in to rape your face if they were so inclined.

Don't spoil the sekrits!!
Next thing you know the RMT prices will fall back to their old levels and once again it will be easier for the people with money to catch up with the people who have time. And that's immoral!!

Welp, that's the end of BoB.  :grin:

One would have hoped. It would have been nice from a 'move on' point of view if the train would have obliterated BoB but it didn't and now the train has lost its momentum.
Sure, another multi-alliance orchestrated push might win the coalition another system and more formidable efforts could have yielded even more systems but reality shows us that the steam is gone, the train has stopped, the warriors went home to fight their own battles.
The big nap fest, the global peace, is weakening alliances because peace is boring and eve is war.
All the steam that is escaping from the kettles of this once mighty train is being used elsewhere and the pressure is building up on global peace. Bored corps are leaving their alliance, alliances are eyeing their friends, naps become nuisances and plans are born.
Peace in eve is always preparing for war.
There will be another age after this great war and it will be an age of shattering.

Wait and see.

My char was created 4-5 weeks ago and doesnt have it so it must be real new.

Let me elaborate: The only thing keeping BoB out of NPC stations at the moment is a huge cap/supercap swarm* defending their own systems under a cynojammer. Supercaps get nerfed (again) to a purely logistical tool, BoB loses those systems. Goonswarm? Well, we tend to use our titan mostly as a mobile jumpbridge anyway, so turning them into mobile stations is an upgrade for us.

* Seven (eight?) titans and counting, dozens of motherships, scores of carriers all on grid simultaneously with all possibly fighters out. Nothing in EVE can attack that and survive, short of an equal zerg...and if that happened, one side would desynch or the node would crash.

Sorry, but that's just cheesy. If they can't fight, lag out the system?

From the people that brought us supercapital pos bowling, fighter lag-bombing is pretty low on their list of sins.

Yeah, that's goonswarm you're talking about.

We don't lag the system on purpose, I mean, how the hell few dozen super-ships is supposed to lag the system anyway, as compared to infamous 600 noobship swarm?
Droes don't add to server lag as much as they do for client side and besides, our defense strategy does not rely on fighter swarming. As in, how the hell fighter swarm to work with multiple doomsdays?! In practice fighters are nto only as lagged as your ships, but also have tendency to get bugged - it's impossible to realiably call them off every few minutes for a doomsday. But again, why let facts get in the way of Simond's pasting.

The solution to the problem isn't nerfing already nerfed sup caps(especially lol motherships - 150% dmg, 3,000% the cost), but limiting JB to BS and smaller ships only, so you have to take Cyno Jammer down to sneak those titans in.

PS. Sneak in some dreads.
PS2. Don't pretend that cowardice to use your titan in combat is a virtue.   

For those who are in any doubt, Joe is trolling.

So what IS your defensive strategy if it isn't lag swarming?  We'd like to know  :oh_i_see:

PS3.  A superior system to the PS2 in nearly every way.

I guess it's just another case of the dev's not being prepared for what the players will do to "their" game. You'd *think* that the Devs would have adequate hardware, and efficient programming to allow relatively smooth combat between 2 massive armies. Barring that, they should just cap the number of people per alliance in any one system. Or something like that.

Too bad though. I was hoping one day to maybe do a fly-by of a titan while running the Space Balls scene in my head where the president is running to the bridge and says (something like) "The ship is too big, if I walked, the movie owuld be over".

Have you seen their hardware? They're not skimping, I can promise you that. There doesn't exist adequate hardware, but they're using the best you can buy.

As for efficient coding -- obviously there's no way to tell, but judging from Devblogs in the past they're not shy about bringing in outside help to check their code. It's just what people are trying to do is pretty insane -- when CCP manages go get nodes stable and playable for 100v100 with drones and fighters out the ass, people bring in 200v200. Since they handle 40k concurrent across their hardware and do manage to support pretty damn large battles, they're not doing too badly -- I'd imagine their server and netcode is probably the best of the MMORPGs on the market.

Although, just for giggles, I remember a thread on EVE's forums where some idiot suggested to the EVE devs that they use "threads" for their servers, to make it run smoother. He was serious. An EVE Dev -- rather patiently, and without use of the phrase "you fucking moron", noted that yes, they DO thread their stuff -- microthread it, in fact, using stacklass python.

I can't recall what the genius in question responded with....

What they *don't* have is efficient netcode. They are working on it, I think, completely replacing the stack and when that's done performance / loading grid should be quite a bit better.

Really? I mean, how are you judging "efficient" here? Like as in "they're switching to a custom, application-driven netcode instead of using the standard TCP/IP stack internally" (which is how I read it) efficient, or "They're really doing stupid shit with TCP/IP" efficiency?

'cause I couldn't see most of the other MMORPGs laboring under that sort of load with their current setup, but I've ALSO seen systems (non-game) running better with higher loads (although from what their DEV blog indicates, their DB setup is pretty optimal).

Offhand, though, the ask some insane things of their systems, just from what they've released. There's places that ask more (you should have seen the specs for Kennedy's next-gen launch control setup in the late 90s -- doable now, but at the time it was a serious hardware and LAN killer), but I haven't seen another MMORPG that makes those demands.

There's not much they can do.  While they've sharded their server differently than everyone else (shards/nodes being dynamically assigned), all MMO's still have the problem that if too many people bunch up in one zone or location, there will be huge lag.

Maybe they can make the client pre-load, gradually, a busy system, from several systems away.   Won't really work (there will still be rubber-banding), and then someone can set up a sniffer to "detect" your gate camp from 3-4 systems away.  Another way may be to designate priorities based on ship type and fleet role, and then give packet priority to the fleet commanders etc, and extend the cloak to however long it takes for you to load up the environment + 30s.  Shrug, I'm not a dev, I don't know.

Their client communication code is terrible. Last I read, during large fleet battles, they send an 8 megabyte data dump when you load grid. There is a ton of extraneous data in there and it's all legacy code from the start of the game.

Needless to say, they're working on pairing that down.

Good for them. I like their approach to lag and slowodnws -- throw hardware, change gameplay, optimize code. Too many people just do one of the three. :)

Of course, their players reward them by shoveling more shit into a battle until it lags again.....I fear one day they'll take away my precious drones.

microthreads =/ threading (as in using multiple cores), which is what the whole discussion was about

Link:

http://myeve.eve-online.com/devblog.asp?a=blog&bid=286
 (http://myeve.eve-online.com/devblog.asp?a=blog&bid=286)

The above post and quote are more than 2 years old, but it shows that the SERVERS are threaded and multithreaded.  While the clients are not, but isn't the discussion about the servers?

(Edit to add:  I might be stupid, but I can't really find a big difference between microthread and multithreaded.)

So their servers are running python  :ye_gods:

Well, I guess CPU isn't the bottleneck of any MMO server (might be on an MMOFPS or something which actually needs to complex collision detection). Networks and databases are still slow.

Bottleneck is grid-loading, which is DB+network+usual lag.

DB, from their devblog on it, is insane. While I'm sure their SQL could use optimization (both in actual calls and how it's used in code), since that's true of everyone, the last bottleneck on SQL they were facing was on disk reads -- they were queuing up so many DB calls that while the DB engine was fine, the disks couldn't keep up. They swapped out for solid-state RAM (and did a number of other changes) and things were really smooth until people started tripling the number of pilots for fleet ops.

I think Bhodi was saying that they're working on drastically reducing the amount of data they need to shove down the pipe for grid-loading.

I found an old presentation (http://www.stackless.com/Members/rmtew/News%20Archive/newsPyCon2006Pres) by CCP on stackless Python. One of the slides contains a diagram of the server as it was in Oct 2004. Anyway, even the client seems to be written in stackless Python (except for the parts that need to be fast like graphics etc).


You can compare microthreads and threads. You can say using multiple threads is multi-threaded. You can also say that using multiple microthreads is multi-threaded. The traditional multi-threading is that you have one process that uses multiple threads. With 'real' threads, the OS can allocate each of those threads to run on a separate core. In the case of stackless Python, all of those microthreads run inside one thread (and thus one core).

Why would you want to use microthreads then if they don't take advantage of the new cool multi-core hardware?

Microthreads in stackless Python are called tasklets. The reason you want to use them is because they are fast. Typically they are at least 10x faster than normal threads. Of course it would be cool if you could get those 10x faster tasklets to run on multiple cores as well. From what I have read, CCP has hired experts in the past to try to figure out ways around this, but the problem is that to get stackless Python run on multiple cores, you will lose all the benefits it brings that were the reason you wanted to use it in the first place.

There are some ways around this limitation of course. In order to use multiple cores with Python you have to use processes. You can launch multiple Python interpreters and each of them can run on a different core. But then you need inter-process communication which is a pain compared to just using multiple threads inside a single process (or multiple microthreads inside a single thread).

AFAIK: The EVE server has many CPUs for simulating the solar systems. Each CPU can simulate one solar system or multiple solar systems. But it can't simulate one solar system with multiple CPUs. Also, the server does not support dynamic load-balancing, i.e. the CPUs are allocated at downtime. So each day at downtime, they check which systems are empty or have very little people in them and put many of those systems to run on a single CPU. Then they see that Jita has crapload of people, so they allocate one CPU just to run Jita and nothing else etc.

This is why (I think) you can lag an entire region sometimes if you jump large amounts of people into a system that didn't have very many people in it during downtime. As the CPU suddenly has to do a lot more work (or wait for network/databases) every solar system simulated on that CPU suddenly feels it.

So, basically, there's not much that they can do to improve lag.  Biggest thing would be to somehow convince players to:

a.  not blob so much
b.  spread out instead of concentrating in Jita/Empire

Not sure if it's possible.

tax overcrowded regions.

Revert the HP boosts on capitals, supercaps and POSes.
Knock 10% off all of their resistances, while they're at it.

The savings is gained by not having to do a full system call for context switches.  For example, on an ARM9 cpu (near and dear to my heart), you can switch contexts (save/restore registers, swap stacks) to another thread in 180 cycles.  Doing this at the kernel/syscall level (on linux) costs 10-20x as much.

The downside is that you get no kernel support so you have to jump through hoops on syscalls, do *all* IO asynch because any blocking operation blocks all threads, and often you tend to do cooperative threading because doing preemptive micro/userspace threads with timers is 1. gross 2. hard and 3. brings you more overhead.

Python is not inherently multi-core/multi-thread friendly.  It has one big "interpreter lock" that must be held while executing the bytecode which pretty much kills performance right there.  Moving from big global locking around the interpreter to lightweight synchronization is a really large and difficult change.  One of the things I will say about Java (which has plenty of horrible horrible issues) is that they built support for multithreaded runtimes into the VM design, so it Just Works -- I wish some more of the nice little interpretive languages did that.

- Q