Pages: 1 [2] 3 4 5 :: one page |
Author |
Thread Statistics | Show CCP posts - 20 post(s) |

Matthew
|
Posted - 2005.10.06 20:00:00 -
[31]
Edited by: Matthew on 06/10/2005 20:00:00
Originally by: Wanoah I still wish Eve could run on one of IBM's mainframes. I honestly don't know whether the architecture would work at all for games (I suspect not) but the reliability and stability is phenominal. They make Intel-based servers look like they were built by retarded monkeys.
Would scalability be the factor here - i.e. it's easier to upgrade the bladecentre when you get more users?
You can do anything. But you can't do everything. |

Helison
|
Posted - 2005.10.06 20:04:00 -
[32]
Oveur: Would it help for lag (and would it be economical justifiable) to reduce the number (availability) of the 2-weeks test-accounts? Or reduce the 2 weeks to 1 week?
|

The Enslaver
|
Posted - 2005.10.06 20:12:00 -
[33]
Oveur! I wub j00 
Anyway, one main thing that is causing serious aggrovation for a lot of people (especially those with multiple accounts) is the char sheet. I agree that the forums are far more important; however would it be possible to set it up so instead of automatically updating with TQ constantly, it only updated every say, 6 or 12 hours? Would seriously aid those of us that rely on the EVE Char manager.
Cheers! :) --------
FireFoxx80: If you think you can do a better job, go find yourself a datacentre to host a box, get a copy of Visual Studio, and STFU. |

Tal Oman
|
Posted - 2005.10.06 20:14:00 -
[34]
I wouldn't be playing right now if I only had a week if thats any help. My first week was so dull and boaring that I was considering not continuing. Second week was a blast as I finally started interacting with good folks and alot of things started making sence.
So I guess it would help the lag if it were only a 1 week trial 
|

Hal2
|
Posted - 2005.10.06 20:21:00 -
[35]
Originally by: Oveur
Taking in contract programmers into our environment does two things:
- Takes time from current programmers to get the new programmers familiar with our environment.
- Time it takes for the new programmers to get familiar is wasted because he cant start doing anything useful until he is familiar with everything
Even worse side-effects of taking on contract programmers is that you are wasting both of those factors in knowledge which then leaves the company.
This is not taking into the calculation of the fact that programming teams scale badly in terms of coordination effort and productivity.
However, we recently hired 3 new programmers, estimated time to full productivity is 2-3 months.
Thanks for your reply and I must say the hiring of 3 new programmers is a good thing - my thoughts which perhaps I didn't expand on enough was to get maybe a few more of the 15 moved over into the optimization project as I assume that maybe they are doing more maintenance or other mundane tasks which of course I could be wrong. I know the issues of bringing in contract persons as I often have to do it myself and the problems that go with it. But if there is any type of task that could be done by them allowing the freeing up of those that could assist I would still consider it.
Originally by: Garath Kane
As a coder, I would like to comment that its not that easy to just bring in 'contract programmers' to bolster the team. Bringing someone new into an existing, ongoing project is throwing that person to the wolves. The amount of time it would take to get familiar with the systems to be able to OPTIMIZE pre-existing code is almost always more than the amount of time existing team programmers can devote to optimization. In other words, it could take a new coder months before he/she is comfortable enough to make significant optimizations and that would defeat the purpose of bringing in a contact coder to get things done quicker. Better to keep existing staff working on the optimizations (as they are doing) and look towards long term return on hiring new full-time programmers.
As I clarified above I guess I didn't make my point clear enough, but if the use of contract help may help to alleviate some of the more mundane tasks that 8 of the programmers are still doing then it allows more of the "experts" to work on the problems.
Originally by: markol
As a software engineer myself, I would like to point out to you where you are mistaken. More programmers does not mean the work will get done faster. I can only imagine the codebase size of this game. It's probably on the order of BILLIONS of lines of code. Not that it means anything to an actual programmer, but that should give you (I am guessing a non-programmer) a point of perspective. The ramp-up time (the time it takes to get familiar) with a code-base this size is probably around months.
Even if you hire 20 new people to work on the code changes, you are not going to see any actual output from them for at least 2 months. This is not counting all the testing time needed for whatever they produce. Furthermore, the more people you have, the more they step on eachothers toes, further reducing productivity.
I'm not a programmer per se, although it was my first carreer goal and I have written some code that is in use by my company even after 6 years I am actually a Telecom/Network Engineer which involves a lot of programming, although no where near something like Eve obviously, but I can still appreciate what it means to maintain something of this complexity. But in any IT function there is always somethings that any competent individual that has the proper knowledge can do. I.E. I hold mutiple Cisco Certifications should I be the one actually putting a LAN switch into a rack and plugging in all the cables or would my skills be better utilized elsewhere? I assume that the same could be true here.
|

Skarsnik
|
Posted - 2005.10.06 20:30:00 -
[36]
Have to say, thank you to both the OP and to Oveur for such an informative post. It has certainly allowed for a further undertsnading of the problems at hand, and the resolutions currently ongoing to deal with them as efficeintly as possible.
I've worked with large scale clusters in the past and I also deal with load testing of programs of all types (thin/fat client and web based) I was wondering and would apprecite and answer on wether you have some kind of Stress testing solution in place to determine the volumes achievable by the current hardware?
Just a geeky thought  ------------------------------------------------- Statement of the Obvious
Contrary to Popular belief - it is NOT possible to walk on water.
|

Xune
|
Posted - 2005.10.06 20:42:00 -
[37]
Some maybe know..... i often told that i find Ovour¦s posts extremly anoying and somtimes even worse then the post from some forum-troll¦s.
But this one was realy good.. i mean... why cant you ever replay honest and profesinal like that ?
Xune
|

Whiskette
|
Posted - 2005.10.06 20:46:00 -
[38]
And this is one of the reasons I love this game. I've never gotten details about whats going on on the server side of things like this. Keep up the good work!
|

Mangus Thermopyle
|
Posted - 2005.10.06 20:48:00 -
[39]
Originally by: markol ... It's probably on the order of BILLIONS of lines of code. ...
Haha, I can assure you, its not. Its probably a few millions though. As software projects go, EVE is a medium sized project. I have worked with bigger projects myself (200 person project at Ericsson with over a million manhours in the software).
But you are right, adding developers scale really bad. Adding more developers costs more than you gain for the first weeks, unless you can isolate very specific and context free problems for them.
|

toto007
|
Posted - 2005.10.06 21:28:00 -
[40]
Another proof that CCP is not afraid to admit its hardware/software weaknesses! Which makes this game all the better!
Thanks for this very interesting and sincere information, Oveur.
|

infused
|
Posted - 2005.10.06 21:35:00 -
[41]
Awesome post.
Have the devs concidered using personal blogs as a way of letting us know stuff? It would be cool to see that you are testing a hotfix this week and what it will be doing.
Just my .1 isk.
|

Rawiana
|
Posted - 2005.10.06 21:56:00 -
[42]
Originally by: Helison Oveur: Would it help for lag (and would it be economical justifiable) to reduce the number (availability) of the 2-weeks test-accounts? Or reduce the 2 weeks to 1 week?
I've only seen 1600 trial accounts active at once, not sure if that would help lag or not by reducing it to 1000 active instead of 3000.
But for me I avoided Jita after the first time I went through...took me 10 minutes to get across the system, I couldn't imagine running missions there.
|
|

Redundancy

|
Posted - 2005.10.06 22:43:00 -
[43]
Polite posts, constructive criticism, well thought out ideas (positives and negatives both considered) and the like are always more likely to get answered. You grow a pretty thick skin as a dev reading the forums, but perhaps one of the most biting comments people can make is that we don't care, because nothing could be further from the truth.
|
|

Wanoah
|
Posted - 2005.10.06 22:48:00 -
[44]
Originally by: Matthew
Originally by: Wanoah I still wish Eve could run on one of IBM's mainframes. I honestly don't know whether the architecture would work at all for games (I suspect not) but the reliability and stability is phenominal. They make Intel-based servers look like they were built by retarded monkeys.
Would scalability be the factor here - i.e. it's easier to upgrade the bladecentre when you get more users?
Oh yeah, it's definitely easier to plug in a couple more servers than add extra mainframes to your cluster. I do vaguely remember that you can cluster upto 256 of the Z series servers. That would be awesome power!
We are all in the gutter, but some of us are looking at stars. (Sig best viewed with Firefox)
|

Dale Cussler
|
Posted - 2005.10.06 22:49:00 -
[45]
Originally by: Mangus Thermopyle
Originally by: markol ... It's probably on the order of BILLIONS of lines of code. ...
Haha, I can assure you, its not. Its probably a few millions though. As software projects go, EVE is a medium sized project. I have worked with bigger projects myself (200 person project at Ericsson with over a million manhours in the software).
But you are right, adding developers scale really bad. Adding more developers costs more than you gain for the first weeks, unless you can isolate very specific and context free problems for them.
Dunno, judging by the looks of the files I kept around since Beta, I'd say 80% of the game is actually written in Python, with just enough C/C++ thrown in to be "glue" to DirectX and some other libraries. Python tends to be .. sickeningly small as far as line count goes, so it might even be less than a million lines of code.
Then again considering the intricacies of the entire game world, it will take a long time for a new person to get familiar with the code, and even longer for them to be able to make changes that will not affect anything adversely.
As far as the hardware goes, the only thing I could possibly fault CCP for is that they didn't order it sooner, but 20/20 hindsight always has all the answers. And as far as hardware delivery goes, if you want something custom built, or bleeding edge, expect to pay through the nose for it, and wait a few eons for it to actually show up. Delivery time on 46 regular cheap-ass build-a-cluster-out-of-std-1U-rackmount boxes is up to 6 weeks now and that's just ordering 46 standard boxes without anything bleeding edge.
I'm just glad a good OP was made, and a good answer was given :D /Dale |

Wanoah
|
Posted - 2005.10.06 22:51:00 -
[46]
Originally by: Redundancy You grow a pretty thick skin as a dev reading the forums, but perhaps one of the most biting comments people can make is that we don't care, because nothing could be further from the truth.
The evidence of that is that you're posting at 22:43 - not exactly normal office hours!
We are all in the gutter, but some of us are looking at stars. (Sig best viewed with Firefox)
|

Flee Now
|
Posted - 2005.10.06 22:52:00 -
[47]
Edited by: Flee Now on 06/10/2005 22:52:19 :)
|

Tredegar
|
Posted - 2005.10.07 01:10:00 -
[48]
It sounds like CCP is starting to be a victim of it's own success. 
Keep up the good work.
Remember technology is a great thing.... until it doesn't work!!! 
"I may be a craven little coward, but I'm a greedy craven little coward." Daffy Duck |

Ticondrius
|
Posted - 2005.10.07 01:36:00 -
[49]
I'm hoping that hardware upgrades <= 2 months to go in..or Christmas lag is going to be a bi***!
Thanks for the VERY informative post Oveur! It's good to see CCP going in the direction of upgrades, bugfixing and optimizations rather than spamming us with semi-broken content every month or two...not intentionally I'm sure. 
IMO, if you have to push back the first of the big Kali patches until after the holidays and the hardware upgrades, I won't complain a bit.
|

M3ta7h3ad
|
Posted - 2005.10.07 01:39:00 -
[50]
Solid State?
Would this suffer from similar "limited write cycle" issue that currently plagues consumer devices built along the same lines (i.e. flash memory devices).
I.E. it will drastically improve performance in the short to medium term, but 2 years down the line could we end up in the same position?
However, kudos to both the OP and Oveur for their posts. Both well constructed and informative :) Good to hear something is being done. My only worry is that as I mentioned above, maintenance issues will once again plague EVE in the not so near future.
|

Sorja
|
Posted - 2005.10.07 02:03:00 -
[51]
Good stuff.
Thx for the informative reply Oveur.
Kill mails |

Kermit Frogly
|
Posted - 2005.10.07 02:57:00 -
[52]
Thank you Kismet for taking the time to write a letter and represent all of us, and thank you Oveur for taking the time to respond in a very clear and concise manner.
These are the reasons I enjoy Eve.
|

MiloMorai
|
Posted - 2005.10.07 02:58:00 -
[53]
Best. Post. And. Response. EVER.
|

Nyxus
|
Posted - 2005.10.07 04:38:00 -
[54]
THIS is why EVE has grown and will continue to grow.
Most of the playerbase isn't 12 years old and drooling. We have expectations and questions, and when you come out and tell us clearly what the problems are what you are doing to resolve them we can understand to some degree the horrendous amount of work that you have to do. We respect that, and we respect you for your candor.
Lag (or any other problems) where the developer fails to communicate all the hard work they are doing to alleviate the issue makes the playerbase start to look for another game.
A problem that is directly dealt with and the upcoming plans (and hard work they entail) that is commmunicated directly to the users makes them say "Oh. Glad they are working so hard on it. If I know it will get better in a few weeks/months/quarters then I will stick with it". Player retention 4TW.
Communication: One of CCP's greatest assets. And no one in the industry does it better. Hat's off to you CCP.
Nyxus
|

Aelius
|
Posted - 2005.10.07 06:44:00 -
[55]
Although i don't have enough time to play EVE like i used to a year ago, lag can come from a wide range of causes. Even with the tech advances on that area you are still playing online you know?!
Personaly i don't notice an outrageous amount of lag, true when there are 300+ people on a system that takes some time to load, but that isn't the end of the world.
I know i pay to play (2 acounts) but i'm still happy i pay for lag in EVE than for lag in some grinding level MMORPG with poor gaming experience.
But don't let all this "fanboy" talk fool you, i'm concerned. Not about coding, hardware, optimizations or number of programers, i'm concerned about CCP lack of planning in growing slow but steady.
We see buddy programs, adds in every MMORPG site and others, free 14 day trials and all the advertising, when the servers are hiting the roof??? What kind of planing is this CCP?! Can't you see that beside "hurting" your faithfull playerbase your are still "burning" possible future costumers?!
Don't want to grow over your possibilities, grow slow but steady. Megalomania is BAD and tell your marketing team thinking on buying a Ferrari to "hold down their horses".
Just a piece of advice Oveur, don't kill the "golden egg goose" plan ahead. 
|

Ophenbach
|
Posted - 2005.10.07 07:40:00 -
[56]
I too would like to say thanks for the info on what is happening.
Thanks 
A few people have come close to "Brook's Law", but none have named it, feel free to Google it if you're unfamiliar.
Keep up the good work.
----- Gone today, however,
Ophenbach
|

Admiral Keyes
|
Posted - 2005.10.07 07:48:00 -
[57]
Originally by: MiloMorai Best. Post. And. Response. EVER.
Word! 
|

Bad Harlequin
|
Posted - 2005.10.07 07:50:00 -
[58]
Originally by: Original Downer when someone with high intelligence and maturity posts in a reasonable tone with clear and concise questions. 
These words you speak, they seem alien to me  Seriously, where'd this come from? Thank you both, for the post and the info, questions and answers in a format that don't make my eyes bleed (avatar pic notwithstanding).
|

Xelios
|
Posted - 2005.10.07 07:52:00 -
[59]
I keep saying EVE's customer relations are among the best in the business. I still remember way back when a member of the community (I think it was Winterblink) was involved in a car accident, and out of the blue.dll ( ) CCP sent him some autographed EVE posters and other goodies. Never have I seen a dev team go out of their way, and take money from their pockets, to brighten up one of their customer's hospital rooms.
They've put years of effort into this game, and continue to find ways to improve it every day. To say they don't care about what they've made, or that they're not proud of it, is ridiculous.
So thanks again for the info Oveur and Redundancy  _____________________________________
|

Buddrow
|
Posted - 2005.10.07 08:01:00 -
[60]
i can't really do anything but clap at this post
when eve started we talked of a few hundred in system, 2-3 max... well when we finally reach that nodes crash... well hell you hit the roof your bound to fall, now we just need to make that 200 playable ---------------------------------- "Give me but one firm spot on which to stand, and I will move earth." Archimedes c.287 - 212 BC
|
|
|
Pages: 1 [2] 3 4 5 :: one page |
First page | Previous page | Next page | Last page |