Author Topic: New efforts!  (Read 7212746 times)

chuckv3

  • Elite Boss
  • *****
  • Posts: 362
Re: New efforts!
« Reply #13920 on: January 05, 2015, 08:30:38 PM »
...just a related thought to the current meandering thread topic:

How possible is it for the thoughtful folks who worked on Titan Icon to put in toggles for SS or SJ? Or to patch over the does-this-character-have-this-power check so that the Teleport bind can be made to work? Having fly is really nice, but some of us really didn't use fly very much in the game (yes, really!) That way we can re-familiarize ourselves with the other travel powers while patiently waiting for the game to return.

Arcana

  • Sultaness of Stats
  • Elite Boss
  • *****
  • Posts: 3,672
Re: New efforts!
« Reply #13921 on: January 05, 2015, 08:31:44 PM »
Edit: Since I mentioned the PFF problem, I'll explain it a bit better.  Basically, the stats were only affected by very specific slots of gear.  Recovery, strength and dexterity were exclusively improved in the offense slot.  Constitution and presence improved exclusively in the defense slot.  While Ego, endurance and intelligence were only improved by the utility slots.  Each peice of primary gear could only generally improve one stat effectively, with utility having a very, very tiny impact on one or two stats but in a very, very small amount.

.....You heard that right.  Now, because PFF's recharge was scaled by ego, and it's maximum capacity was scaledy by endurance, this meant you either improved it's recovery which was abysmally slow even with high ego or it's maximum capacity which scaled only with endurance.  And because it would always automatically take 100% damage, regardless of resists or blocking or dodging, it broke very very fast and generally never got back up to capacity.  So it was entirely useless.

"The" PFF problem was (as you mentioned) that PFF was outside block (I have no idea if it still is or not).  Given the mechanics of block, that one problem made all other issues practically irrelevant, because PFF was essentially worthless against anything that had super attacks that were designed to be blocked (even if you didn't use block for anything else).  The stats thing was just annoying by comparison.

Actually, "the" PFF problem was the devs explicitly stated that that was the mathematically balanced correct behavior for PFF, and to this day I have no idea how anyone could think that was possible, but it made it impossible to argue for a fix.  That was when I decided to give up analyzing CO, because it seemed pointless to me to calculate the stuff that was 10% out of balance when there were massive order of magnitude balance errors the devs thought were correct.  That would be like in CoH trying to calculate the proper recharge for Fire Blast powers when the devs thought all Fire Blast powers should do Scale 8 DoT.

Also, most players who were in Beta and rolled PFF and decided to do that for launch were instantly vaporized on launch without even knowing why.  It wasn't until testers like myself and others actually tested PFF carefully that the block behavior was discovered (that was one of the launch day nerfs, and I'm pretty sure it was a stealth one).

Arcana

  • Sultaness of Stats
  • Elite Boss
  • *****
  • Posts: 3,672
Re: New efforts!
« Reply #13922 on: January 05, 2015, 08:53:04 PM »
Except that's not cloud computing. Lots of servers sharing resources is clustering. Dynamically distributing load to clusters is part of what virtualization offers, along with fast provisioning (cloning) to scale to accommodate load. The industry has been doing all of those for much longer than the term 'Cloud' has been in vogue.

I find that wikipedia article amusing because it has a lot of self-referential terminology that describes the idea of 'cloud computing', but little to no details on how it's actually implemented. The concept is as nebulous as the name implies, and to those of us who actually work with the technologies every day it's nothing more than a pointless marketing catchphrase. Especially when people talk about 'private cloud', I scratch my head and say... "Uh, you mean datacenter? Yeah, we've got one of those."

Though, to be accurate, you can't unconditionally state that cloud services are all hosted on clusters of many servers. You don't know that, you can't know that, the whole point of 'cloud' is that you don't care. A cloud provider sells (well, leases) you a service. That service might be hosted on a state of the art geographically distributed cluster of hundreds of servers. Or it might be on a single Dell in a closet because you're the only subscriber for that service. It's up to the provider to figure out what technology to use to meet the requirements they promised as part of the service. That's why I say that 'cloud' is merely a politically correct synonym for 'outsourcing'.

Every time I go away for a while and come back, I end up reading about how someone has rearranged the solar system and computer technology took a giant leap sideways.

Just because this is something I've talked about recently (like in public, on a stage, not here) and I don't care if I'm replying to future people posting from jetpacks, back in 2000 someone asked me what "cloud computing" was and I said pretty much the same thing: its a marketing buzz word to sell you more hosted services.

I've changed my tune a bit, in that while it might have started off that way, the mindset of cloud computing has generated real differences, particularly in the mindset of "run anywhere, on anything, at any scale."  Even though the pieces are the same and the technology is the same, there are genuine differences in how cloud computing clusters are designed, and even in how cloud computing people think.  A few years back I was bouncing cluster ideas off of a guy I know that is more of a cloud guy, whereas I was at the time more of a virtualized hosting person in mindset, and we got into a weird discussion about network reliability.  It basically went something like this: we were talking about rack servers with multiple 10 gig ethernet trunks.  I mentioned in my systems I designed the network to use the 1 gig ethernet ports as emergency backup to the main 10 gig trunks in case of a failure in the 10 gig segments.  He said he didn't bother to connect those at all, even though he had them and there were lots of ports to connect them to.  My thought was: use them, it can't hurt.  His thought was: don't use them, they can only hurt.

To me, having an entire server network-partitioned meant all the load on that server went dead: slow was better than nothing.  To him, having an entire server dropped down from 20 gig trunked to 2 gig meant that load was hyperconstrained: better to fail the node and let the workload migrate or restart on a fully operational node than be crippled.

Take that conversation, multiply by about a thousand, and that's the difference between cloud computing and conventional computing.  It starts off as mostly a philosophical difference, but then that philosophy gets turned into implementation, and pretty fast the cloud guys are living in a completely different world.  But only if the entire stack buys in, from the applications upward.  Otherwise, its still marketing.  But in a clusterized containered world with virtualized storage and appropriate switching, it can be night and day.

brothermutant

  • New Efforts # 14,000!
  • Elite Boss
  • *****
  • Posts: 309
Re: New efforts!
« Reply #13923 on: January 05, 2015, 09:52:53 PM »
"The" PFF problem was (as you mentioned) that PFF was outside block (I have no idea if it still is or not).  Given the mechanics of block, that one problem made all other issues practically irrelevant, because PFF was essentially worthless against anything that had super attacks that were designed to be blocked (even if you didn't use block for anything else).  The stats thing was just annoying by comparison.

Actually, "the" PFF problem was the devs explicitly stated that that was the mathematically balanced correct behavior for PFF, and to this day I have no idea how anyone could think that was possible, but it made it impossible to argue for a fix.  That was when I decided to give up analyzing CO, because it seemed pointless to me to calculate the stuff that was 10% out of balance when there were massive order of magnitude balance errors the devs thought were correct.  That would be like in CoH trying to calculate the proper recharge for Fire Blast powers when the devs thought all Fire Blast powers should do Scale 8 DoT.

Also, most players who were in Beta and rolled PFF and decided to do that for launch were instantly vaporized on launch without even knowing why.  It wasn't until testers like myself and others actually tested PFF carefully that the block behavior was discovered (that was one of the launch day nerfs, and I'm pretty sure it was a stealth one).
The only reason people hated how they screwed up PFF was because this was the only game to ever really appreciated how a PFF SHOULD work. A self contained field that bled thru a MINOR amount of the damage to your health with most of the damage going to the PFF. When the field was "broken", you were on your own. Sadly, since there was little in the way of getting a broken shield back up to full WHILE in battle, it was considered borked (yes, "borked" is a word).

I for one loved the PFF and had it on many toons. They FINALLY added a minimal layer of protection to it. its been so long I forget what it was but blocking was now part of the layer of protection applied to the amount of damage before being directly subtracted from the field (I think dodge/avoidance were too, but you really needed great gear to get those numbers into something realistic).

The argument I remember was really two-fold:
1) why does PFF "heal" itself so slowly, especially after being broken? This was a real concern as Regen got better at healing the more damage you took, PFF was the reverse of this.
And, 2) why take PFF that needs constant care and the use of a couple of OTHER powers to help maintain it when I can just grab Defiance/Regen/whatev and pretty much just run with only that one power choice and be (reasonably) fine?

To me, these were the real reasons PFF is overlooked and hated by many. But I still feel like this is the only game that has come close to making a PFF that seems right for a hero (CoH had a DUMA$$ PFF that gave HIGH def and a decent resist bonus to it BUT you can't attack with it on!? And DCUO has a lamesauce shield sucks up X amount of damage or # of attacks before going away. Lame.)

Vee

  • Elite Boss
  • *****
  • Posts: 2,376
Re: New efforts!
« Reply #13924 on: January 05, 2015, 10:29:31 PM »
While I really liked Fly... only one of my toons actually had fly. The rest of my toons all had ss/sj. Heck even my toon with fly had cj and ss. I usually would put a stealth IO in cj.

For future reference (fingers crossed, right?) you could put the stealth IO in sprint leaving CJ open as a LOTG+recharge mule and SS open for a BOTZ -KB if needed with maybe another BOTZ for the ranged defense. Putting it in sprint also allows you to use whichever of unbounded leap and celerity is on hand/cheaper.

LaughingAlex

  • Giggling like an
  • Elite Boss
  • *****
  • Posts: 2,019
Re: New efforts!
« Reply #13925 on: January 06, 2015, 01:24:59 AM »
"The" PFF problem was (as you mentioned) that PFF was outside block (I have no idea if it still is or not).  Given the mechanics of block, that one problem made all other issues practically irrelevant, because PFF was essentially worthless against anything that had super attacks that were designed to be blocked (even if you didn't use block for anything else).  The stats thing was just annoying by comparison.

Actually, "the" PFF problem was the devs explicitly stated that that was the mathematically balanced correct behavior for PFF, and to this day I have no idea how anyone could think that was possible, but it made it impossible to argue for a fix.  That was when I decided to give up analyzing CO, because it seemed pointless to me to calculate the stuff that was 10% out of balance when there were massive order of magnitude balance errors the devs thought were correct.  That would be like in CoH trying to calculate the proper recharge for Fire Blast powers when the devs thought all Fire Blast powers should do Scale 8 DoT.

Also, most players who were in Beta and rolled PFF and decided to do that for launch were instantly vaporized on launch without even knowing why.  It wasn't until testers like myself and others actually tested PFF carefully that the block behavior was discovered (that was one of the launch day nerfs, and I'm pretty sure it was a stealth one).

Yupe, pretty much what was happening was PFF was placed before any kind of resistances or dodging during the overall damage calculations.  It was more like "PFF absorbs damage first, then the damage left over goes through resistances, then dodging", which ensured PFF broke right away.  They had fixed that, but even then it doesn't provide enough resistance or recharge fast enough to be effective for the frosticus rampage.  I had proposed that PFF shouldn't absorb 100% damage and should absorb damage based on the % of it left, with resistances still applying, to make it useful during the entire course of a fight.  Course, that won't happen at this rate :(.
Currently; Not doing any streaming, found myself with less time available recently.  Still playing starbound periodically, though I am thinking of trying other games.  Don't tell me to play mmohtg's though please :).  Getting back into participating in VO and the successors again to.

Rejolt

  • Elite Boss
  • *****
  • Posts: 512
  • Infinity Server 2005-2012
Re: New efforts!
« Reply #13926 on: January 06, 2015, 02:56:18 AM »
For future reference (fingers crossed, right?) you could put the stealth IO in sprint leaving CJ open as a LOTG+recharge mule and SS open for a BOTZ -KB if needed with maybe another BOTZ for the ranged defense. Putting it in sprint also allows you to use whichever of unbounded leap and celerity is on hand/cheaper.

That was always my plan. Hover usually got a lotg if I didnt have a microfiliment on hand.

Edit: lord, all my 100s of Hamidons gone ... And I never liked the new raids. Eh, there's always the S... Er Liberty tf!
Rejolt Industries LLC is now a thing. Woo!

Sinistar

  • Elite Boss
  • *****
  • Posts: 1,045
Re: New efforts!
« Reply #13927 on: January 06, 2015, 03:33:38 AM »
Edit: lord, all my 100s of Hamidons gone ... And I never liked the new raids. Eh, there's always the S... Er Liberty tf!

I share your pain.
In fearful COH-less days
In Raging COH-less nights
With Strong Hearts Full, we shall UNITE!
When all seems lost in the effort to bring CoH back to life,
Look to Cyberspace, where HOPE burns bright!

Vee

  • Elite Boss
  • *****
  • Posts: 2,376
Re: New efforts!
« Reply #13928 on: January 06, 2015, 03:45:24 AM »
only hami i used a lot was the one with the ddr bug that they finally fixed to the great dismay of my widow and shield toons.

so i had a thought (probably dumb) and am in dire need of some rampant insane speculation or for someone to tell me i'm stupid - what do we think would be the state of the market on a disc image? i have no idea what the image would involve but i wonder if a) there'd be nothing on the market at all and we'd have to rebuild the economy as we rebuild our toons or II) if all the stuff and bids would still exist despite the accounts they correspond to no longer existing and thus all our new toons with their low budgets would have to pony up 400 million for purps. i'd assume the former but i figure we'll all have a ton of work to do either way (which i'm kind of looking forward to actually - i left for 6 months once before and gave away everything. was kind of fun to have to rebuild it all).

Waffles

  • Elite Boss
  • *****
  • Posts: 387
Re: New efforts!
« Reply #13929 on: January 06, 2015, 05:03:09 AM »
Hami's were great for defense toggles when short on slots to spare. >_>

AmberOfDzu

  • Boss
  • ****
  • Posts: 179
Re: New efforts!
« Reply #13930 on: January 06, 2015, 05:04:40 AM »
so i had a thought (probably dumb) and am in dire need of some rampant insane speculation or for someone to tell me i'm stupid - what do we think would be the state of the market on a disc image?

I would expect a disk-image restart of CoH to begin with an empty market. All the items on the market were posted from and linked to user accounts, so they're almost certainly part of the user databases that would not be on the disk image. So, we'd have to build a new market up from the beginnings. The good news is that all those years of inflation would be wiped away, and everyone would be at a fresh start. Who knows? Maybe the pricing would be reasonable for a while. ^_^

Vee

  • Elite Boss
  • *****
  • Posts: 2,376
Re: New efforts!
« Reply #13931 on: January 06, 2015, 05:12:27 AM »
That's what i figured but wasn't sure. I expect we'll all need to forget there are vendors for a while to get things going again nicely and hope that the supply problem doesn't kickstart the inflation back to where it was too fast. And maybe it'll keep more people doing tfs for the merits instead of ae farming everything again.

Waffles

  • Elite Boss
  • *****
  • Posts: 387
Re: New efforts!
« Reply #13932 on: January 06, 2015, 05:17:00 AM »
I would expect a disk-image restart of CoH to begin with an empty market. All the items on the market were posted from and linked to user accounts, so they're almost certainly part of the user databases that would not be on the disk image. So, we'd have to build a new market up from the beginnings. The good news is that all those years of inflation would be wiped away, and everyone would be at a fresh start. Who knows? Maybe the pricing would be reasonable for a while. ^_^

I'd argue the lack of badass super-farmers IOed to the brim would drive prices technically higher.


I'm worried about how i'm going to afford stuff, I made serious dough off selling purples, one sale could fund an entire character's set ;_;

Vee

  • Elite Boss
  • *****
  • Posts: 2,376
Re: New efforts!
« Reply #13933 on: January 06, 2015, 05:30:46 AM »
Be tough for them to be as high just because there won't be that much circulating influence at the beginning. could easily be just as cost prohibitive though.

Sinistar

  • Elite Boss
  • *****
  • Posts: 1,045
Re: New efforts!
« Reply #13934 on: January 06, 2015, 05:30:50 AM »
I'd argue the lack of badass super-farmers IOed to the brim would drive prices technically higher.


I'm worried about how i'm going to afford stuff, I made serious dough off selling purples, one sale could fund an entire character's set ;_;

Or a sale of one +3 DEF IO from the PVP set could fund one or two character builds.

As to the market, it should be starting off an zero when the game returns unless the new owner/devs seed the market a bit with some salvage and IO's.

I don't see the market getting crazy with prices for the first month or two, after that the insanity should begin, after about 6 months I'd expect some of the old price gouging to be back.  However IO conversions did help offset some of the market prices for certain IO's and I expect that would be the same again.

Also perhaps a once per month reset of the market wouldn't hurt? All posted items for sale are revoked back to the poster along with any inf spent to post the item, the market is reset then perhaps seeded a bit to help out players as they rebuild their characters.

After about 6 to 12 months such seeding could likely be stopped.
In fearful COH-less days
In Raging COH-less nights
With Strong Hearts Full, we shall UNITE!
When all seems lost in the effort to bring CoH back to life,
Look to Cyberspace, where HOPE burns bright!

Vee

  • Elite Boss
  • *****
  • Posts: 2,376
Re: New efforts!
« Reply #13935 on: January 06, 2015, 05:33:54 AM »
i hope the seeding is possible, not sure what sort of gm tools will be available. i shudder at the idea of unsold returns though, as i've never been so annoyed by a game feature as i am at swtor's market.

Burnt Toast

  • Elite Boss
  • *****
  • Posts: 548
  • Tastes Good!
Re: New efforts!
« Reply #13936 on: January 06, 2015, 07:30:18 AM »

I never run sprint. I always have CJ on so my stealth is always there. If I want to move faster than normal... SS. SS+stealth=invis. I used to always place the stealth IO in SS, but moved it to cj because I found it more effective since I don't always have SS on. As far as the set bonuses... I sometimes use the BOTZ set for KB and the ranged bonus....

For future reference (fingers crossed, right?) you could put the stealth IO in sprint leaving CJ open as a LOTG+recharge mule and SS open for a BOTZ -KB if needed with maybe another BOTZ for the ranged defense. Putting it in sprint also allows you to use whichever of unbounded leap and celerity is on hand/cheaper.

Void Huntress

  • Minion
  • **
  • Posts: 39
Re: New efforts!
« Reply #13937 on: January 06, 2015, 07:44:20 AM »
To me, having an entire server network-partitioned meant all the load on that server went dead: slow was better than nothing.  To him, having an entire server dropped down from 20 gig trunked to 2 gig meant that load was hyperconstrained: better to fail the node and let the workload migrate or restart on a fully operational node than be crippled.

The first time I encountered this mindset was when I started investigating erlang as a venue for developing some projects. The principle is: If it's broken, kill it, and start over somewhere else, or restart this one if the environment is still viable. In that view, attempting to gracefully fail on any particular node is attacking the wrong problem, because once 'reliable' assumptions have been violated, you can't trust any error recovery to actually address the issue, so suiciding and starting fresh is more likely to maintain uptime.

Depending on what you're actually implementing, this is sometimes a dubious outlook, but for "100% availability" applications, such as the cloud paradigm, it's a surprisingly effective approach, despite being counterintuitive to traditional viewpoints.

Ankhammon

  • Elite Boss
  • *****
  • Posts: 676
Re: New efforts!
« Reply #13938 on: January 06, 2015, 03:07:31 PM »
The first time I encountered this mindset was when I started investigating erlang as a venue for developing some projects. The principle is: If it's broken, kill it, and start over somewhere else, or restart this one if the environment is still viable. In that view, attempting to gracefully fail on any particular node is attacking the wrong problem, because once 'reliable' assumptions have been violated, you can't trust any error recovery to actually address the issue, so suiciding and starting fresh is more likely to maintain uptime.

Depending on what you're actually implementing, this is sometimes a dubious outlook, but for "100% availability" applications, such as the cloud paradigm, it's a surprisingly effective approach, despite being counterintuitive to traditional viewpoints.

I'm realizing that there is some merit in this approach with my new job. Even though it goes against my every computer trained instincts, it is a faster approach to getting the end user going in many instances.
To me, it's basically taking the intelligence out of debugging issues.
Cogito, Ergo... eh?

Codewalker

  • Hero of the City
  • Titan Network Admin
  • Elite Boss
  • *****
  • Posts: 2,740
  • Moar Dots!
Re: New efforts!
« Reply #13939 on: January 06, 2015, 03:23:41 PM »
I'd probably take that approach with any highly virtualized workload, whether it's heterogeneous (cloud) or not. Cheap, generic, easily replaceable (and by extension disposable) processing hosts is one of the big selling points of things like vSphere Enterprise's vMotion / DRS.

The last thing you want is severe performance degradation as your failure mode; better to migrate everything off. I've run into a similar situation where a particular host had hardware issues that caused all the network ports to negotiate down to 100MBps. Needless to say we migrated everything off ASAP and took it down for maintenance. The problem is that most monitoring tools only check if a given link is up or down, not that it's running at the speed it's supposed to be.