These forums have been archived and are now read-only.

The new forums are live and can be found at https://forums.eveonline.com/

EVE General Discussion

 
  • Topic is locked indefinitely.
 

Tears aside was HeD the biggest?

First post
Author
Lucas Kell
Solitude Trading
#81 - 2014-01-26 19:52:59 UTC
silens vesica wrote:
Lucas Kell wrote:
Well that's not quite right. It's not automatically new bugs,

Yeah, it is. ANY time you do non-trivial re-code, there will be new bugs. Which isn't to say that they can't be over-come; they can be.
It's not automatic. It's not a guarantee that bugs will simply happen. With proper testing, bugs can be minimised. All code is going to have bugs, sure, but that doesn't mean that you get begs by rewriting code by default. The main point being that it shouldn't be used as a reason to not bother.

silens vesica wrote:
Quote:
it does remove their processor cap that they are currently sitting at.
Maybe. Maybe not. Probably, but no guarantee. Meanwhile, the *funding* for such a massive undertaking must be found. It's not a cheap OR simple exercise you're asking for. Are you willing to sacrifice new content and patches for a year or so? Will CCP survive the natural user-attrition over time whilst re-doing the code without new content offerings to draw new subs?

It's not a simple thing.
Of course it does, they 100% for definite will have more power to play with if they can use a server's full range of cores over a single overclocked core that they are already hitting 100% usage at 75% player capacity.
And they already develop plenty of rubbish features. Would I rather they improved performance than chucked in a handful of mobile structures and utterly pointless utilities like ISIS? You're damn right I would.

Now if they don't want to, and want to instead focus of small features and additions, that's fine, but market that. At the moment though they like to market big fleet battles, and what is delivered is pretty much waiting around for nothing to happen. It's not fun in any way, so it's really not good for a game.

The Indecisive Noob - EVE fan blog.

Wholesale Trading - The new bulk trading mailing list.

Prince Kobol
#82 - 2014-01-26 20:11:39 UTC  |  Edited by: Prince Kobol
Lucas Kell wrote:
Nevyn Auscent wrote:
A code rewrite is not an automatic 'EVE runs faster'.
A code rewrite is however an automatic 'New bugs'.

It's a huge gamble for potentially negative results to port EVE to a new programming language with no predictable results beyond the introduction of new bugs in the rewrite. Some of EVE might even break due to interdependencies that have strange effects under a new coding system.

As well as Devs no longer being familiar with the language and how to get the most leverage out of it, instead having to learn all the tiny tips & tricks of a new language.

So.... Yea. Not seeing them ever doing a code base rewrite unless python stops being developed/supported.
Well that's not quite right. It's not automatically new bugs, and while it's not an automatic "eve runs faster" it does remove their processor cap that they are currently sitting at. They are at the limit of how fast a CPU core can go. There's nothing in the industry to grant them more power from a core. While a rewrite wouldn't make it automatically faster, it certainly would give them a lot more options to spread out the load.

By the way, most senior programmers can write a range of languages, and adapt to new ones in a pretty short time.

And sure, they may not rewrite the codebase, but then the amount of power they have no is about the extent of it. They can do tweaks and workarounds like biab, and work with drone AI to make it a bit more streamlined, but the rate of the increase in fights size will continue to outrun the rate of performance increase. And TIDI will have to stay, which means fights will remain incredibly boring even when they do run right. I don't think I know anyone that can honestly say they find a TIDI fight a fun experience, and they consider TIDI to be a solution, madness.


Here is thing, my wife is a senior software engineer with 15+ years experience.

One of her jobs was working for Bae Systems coding for the Tornado GR4.

Now at one time they had to do a full recode of the software from one version of the language they were using to a more updated version for a customer who was purchasing GR4's at the time to allow for new technology and for better efficiency.

Now this rewrite took approx 18 months and had approx 70 people working full time on the project with a hell of a budget for overtime and extra bodies when needed.

You also have to consider that project to update the code base for the Tornado GR4 is actually a lot easier then it would be to rewrite the code base for Eve.

Most of the code for the GR4 could be ported over where as I suspect for Eve this could not be the case. For a rewrite of Eve you have to take into consideration what actions uses which cores, something which the existing code base does not do as it doesn't have to.

I really can not stress how much time and money it takes to rewrite something like Eve, in effect you are basically creating a brand new game.

So whilst yes it should improve performance (who knows by how much)but at what cost?

To be truthful I doubt CCP could even afford to do a complete re-write of Eve.
Lucas Kell
Solitude Trading
#83 - 2014-01-26 20:59:40 UTC
Prince Kobol wrote:
Here is thing, my wife is a senior software engineer with 15+ years experience.

One of her jobs was working for Bae Systems coding for the Tornado GR4.

Now at one time they had to do a full recode of the software from one version of the language they were using to a more updated version for a customer who was purchasing GR4's at the time to allow for new technology and for better efficiency.

Now this rewrite took approx 18 months and had approx 70 people working full time on the project with a hell of a budget for overtime and extra bodies when needed.

You also have to consider that project to update the code base for the Tornado GR4 is actually a lot easier then it would be to rewrite the code base for Eve.

Most of the code for the GR4 could be ported over where as I suspect for Eve this could not be the case. For a rewrite of Eve you have to take into consideration what actions uses which cores, something which the existing code base does not do as it doesn't have to.

I really can not stress how much time and money it takes to rewrite something like Eve, in effect you are basically creating a brand new game.

So whilst yes it should improve performance (who knows by how much)but at what cost?

To be truthful I doubt CCP could even afford to do a complete re-write of Eve.
Honestly I can't be bothered to go on arguing this, as we're going to just go round in circles. The fact is they have a development team who are spending time working on changes which are utterly pointless to the game, and they have reached the limit of the hardware that exists to support their game. If they are not going to change this, it will continue to degrade, and they should simply give up on the idea of marketing the game as having big fleet battles. On top of that, things like POS and corp mechanics are stuck in their pan in the ass state, because of their "we don't touch the old code" policy.

I'm fairly sure you are overstating the amount of time a rewrite would take (I've been involved in a few) and the amount of funding CCP have. Comparisons to aircraft programming which is under considerably more rigorous testing is not really fair. But honestly, it doesn't matter. The amount of effort it takes is pretty much beside the point. The game does not support the features they want it to, and no amount of tweaks is going to suddenly make that happen.

One day they will have to change it or stop supporting large fleet battles. There's simply no way around that. Performance has been an issue now for years, and yet they would rather spend time on totally useless features and rubbish like DUST (which basically plays like a cheap FPS as you would expect).

The Indecisive Noob - EVE fan blog.

Wholesale Trading - The new bulk trading mailing list.

Nevyn Auscent
Broke Sauce
#84 - 2014-01-27 02:45:49 UTC
200 v 200 is a big fleet battle.
Most MMO's cap you at about 40 v 40 and have terrible lag issues while doing that much.
2000v2000 is truly epic, and 8VT showed it can be supported.
Hed they have shown stats on server performance, discussed why it had more issues, and contemplated some solutions to those issues as well as updating us on progress of the BIAB status.

So.... Yea, I'm happy that CCP are being honest in their advertising, and are working in things appropriately.
So sorry you are a glass half empty never satisfied person though.
silens vesica
Corsair Cartel
#85 - 2014-01-27 03:56:14 UTC
Lucas Kell wrote:
silens vesica wrote:

Yeah, it is. ANY time you do non-trivial re-code, there will be new bugs. Which isn't to say that they can't be over-come; they can be.
It's not automatic. It's not a guarantee that bugs will simply happen. With proper testing, bugs can be minimised. All code is going to have bugs, sure, but that doesn't mean that you get begs by rewriting code by default. The main point being that it shouldn't be used as a reason to not bother.
Name me one major code re-vamp you've been associated with that hasn't had new bugs introduced.

I've been involved in many, complete with government-compliant validation testing via audited SOPs (and no, nothing as amateurish as "Healthcare - dot -Gov). Bugs ALWAYS creep in.

But if you get to the end, and read for comprehension, you'll see that I agree; they can be handled.

Quote:
silens vesica wrote:
Maybe. Maybe not. Probably, but no guarantee. Meanwhile, the *funding* for such a massive undertaking must be found. It's not a cheap OR simple exercise you're asking for. Are you willing to sacrifice new content and patches for a year or so? Will CCP survive the natural user-attrition over time whilst re-doing the code without new content offerings to draw new subs?

It's not a simple thing.
Of course it does, they 100% for definite will have more power to play with if they can use a server's full range of cores over a single overclocked core that they are already hitting 100% usage at 75% player capacity.
Interesting that you can make such a guarantee. Where do you source your crystal balls? I'd like to acquire a few - I'll be fabulously wealthy in no time.

*Probably* so. No guarantees.

Quote:
And they already develop plenty of rubbish features. Would I rather they improved performance than chucked in a handful of mobile structures and utterly pointless utilities like ISIS? You're damn right I would.
Fair enough. Now, how about all the *rest* of the user base? Can you speak for them? I suppose you can, if your crystal ball is as good as you think it is.

Tell someone you love them today, because life is short. But scream it at them in Esperanto, because life is also terrifying and confusing.

Didn't vote? Then you voted for NulBloc

Lucas Kell
Solitude Trading
#86 - 2014-01-27 08:20:23 UTC
silens vesica wrote:
Name me one major code re-vamp you've been associated with that hasn't had new bugs introduced.

I've been involved in many, complete with government-compliant validation testing via audited SOPs (and no, nothing as amateurish as "Healthcare - dot -Gov). Bugs ALWAYS creep in.

But if you get to the end, and read for comprehension, you'll see that I agree; they can be handled.
Yes, bugs creep in, but not at a rate any higher than through a regular code change, and they are just as avoidable, thus not "automatic". It should never just be used as an excuse not to bother. The way you stated it is essentially that they shouldn't do a change for fear of bugs, which is ridiculous.

silens vesica wrote:
Interesting that you can make such a guarantee. Where do you source your crystal balls? I'd like to acquire a few - I'll be fabulously wealthy in no time.

*Probably* so. No guarantees.
Well not only is it plain old common sense, but CCP have stated it themselves. When they were explaining that they weren't going to go down that line and started looking at BIAB. The only way there would be no performance gained is it they were truly awful programmers, which obviously they are not. And you know this, you are just being difficult for the sake of it.

silens vesica wrote:
Fair enough. Now, how about all the *rest* of the user base? Can you speak for them? I suppose you can, if your crystal ball is as good as you think it is.
I don;t need to speak for them, they've been pretty vocal about it. Not to mention that there was a breakdown of rubicon's launch spike (or lack thereof) not that long ago.

The Indecisive Noob - EVE fan blog.

Wholesale Trading - The new bulk trading mailing list.

Prince Kobol
#87 - 2014-01-27 10:14:24 UTC
silens vesica wrote:

I've been involved in many, complete with government-compliant validation testing via audited SOPs (and no, nothing as amateurish as "Healthcare - dot -Gov). Bugs ALWAYS creep in.

But if you get to the end, and read for comprehension, you'll see that I agree; they can be handled.


Yes some bug will creep in but you will get bugs with any new code you write, that is the nature of the beast, you do not write new code because you "might get bugs"

That is what QA Testing is for.

I was involved in the initial planning as well as the initial roll-out of the New IT system for the NHS that failed in an epic fashion and that failed not because of bugs but because of idiots in management who were clueless and thought they new better then the professions.

I have also been involved in number of military deployments for new software and hardware with budgets running into the hundreds of millions pounds and most of the problems which we encountered were not due to bugs in new software but poor planning and people in management who knew nothing about IT but were only interested in numbers.

So yes bugs will come but that is no excuse for not writing new code.

silens vesica wrote:
Interesting that you can make such a guarantee. Where do you source your crystal balls? I'd like to acquire a few - I'll be fabulously wealthy in no time.

*Probably* so. No guarantees.


Of course a rewrite will bring a performance increase, to claim it wouldn't is just silly. The issue isn't whether it would bring a performance increase but is what ever increase in performance is achieved is it worth the cost incurred?

silens vesica wrote:
Fair enough. Now, how about all the *rest* of the user base? Can you speak for them? I suppose you can, if your crystal ball is as good as you think it is.


I think the response to the last couple of expansions speaks for itself.

You just have to look at the average PCU count over the last year to see that there is a problem.