PDA

View Full Version : I've got a theory on why .50cal is synched now



gx-warspite
01-18-2006, 07:23 PM
Most servers out there insist that players run 28.k or 56k connection settings, in order to save on bandwidth.

However, with such a small packet size it's difficult to track multiple objects. By synchronizing the 50s, and sending an entire volley across as a single object (rather than each gun or pair of guns firing its own bullets as individual objects), this will reduce the amount of data that needs to be sent across.

gx-warspite
01-18-2006, 07:23 PM
Most servers out there insist that players run 28.k or 56k connection settings, in order to save on bandwidth.

However, with such a small packet size it's difficult to track multiple objects. By synchronizing the 50s, and sending an entire volley across as a single object (rather than each gun or pair of guns firing its own bullets as individual objects), this will reduce the amount of data that needs to be sent across.

VW-IceFire
01-18-2006, 07:32 PM
*IceFire points to the 12 .303 gunned Hurricane Mark IIB*

FritzGryphon
01-18-2006, 07:58 PM
Indeed, the game tracks individual bullets.

Furthermore, hit detection is client side, not server. The order, or number of bullets makes no difference for data transfer.

The data sent to the server is simply that the player has pulled the trigger, and again when it's released. It's why weapon firing (or bomb dropping) is out of sync with client and server, yet damage is allocated according to hits that were observed by the client.

Stackhouse25th
01-18-2006, 08:08 PM
*pulls up chair, gets a bag of popcorn*

Lordbutter4
01-18-2006, 08:33 PM
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by FritzGryphon:
It's why weapon firing (or bomb dropping) is out of sync with client and server, yet damage is allocated according to hits that were observed by the client. </div></BLOCKQUOTE>

Are you sure about that? I have recorded many online sorties were I clearly remember hitting planes more then is shown on my recording. Try it...much more noticable when using 50's or 303's.

Jetbuff
01-18-2006, 10:08 PM
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Stackhouse25th:
*pulls up chair, gets a bag of popcorn* </div></BLOCKQUOTE>
Pass me some...

Tator_Totts
01-18-2006, 10:35 PM
I got mine. http://home.carolina.rr.com/squad/popcorn.gif

fordfan25
01-18-2006, 10:40 PM
.50s are porked be sure

*add's a log*

Aaron_GT
01-19-2006, 01:43 AM
My theory is a bad code merge. In a number of instances we've had features/bugs come and go between patch versions. There are apparently over 1 million lines of code. So my theory is that developers have copies of the sim, make there changes to some modules to get the appropriate behaviour, and then merge these versions to form the new version to be shipped, but that there have been a few slips in this process. I am guessing, but it's not beyond the realms of possibility that the changes to the 50 cal dispersion code/data were tested initially on an older code base which had 'salvo' 50 cals (whereas the latest shipped patch did not have 'salvo' guns) and when the merge of the new code was done back into the main code base the 'salvo ' setting crept back in. We've had other things which suggest this - e.g. sniper AI gunners coming and going, and so on.

This isn't a dig at Oleg's team - it's a huge code base and the number of show stopping bugs has been immensely small for a code base this size. I know a bit about software testing, and the lack of bugs is a real achievement. But a few incorrectly executed code merges introducing non-showstopping issues isn't beyond the realms of possibility.

BSS_CUDA
01-19-2006, 03:25 AM
/me pulls up his chair and waits for the

if you cant hit with 50's you suck and learn to aim line

jimDG
01-19-2006, 03:57 AM
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by gx-warspite:
Most servers out there insist that players run 28.k or 56k connection settings, in order to save on bandwidth.

However, with such a small packet size it's difficult to track multiple objects. By synchronizing the 50s, and sending an entire volley across as a single object (rather than each gun or pair of guns firing its own bullets as individual objects), this will reduce the amount of data that needs to be sent across. </div></BLOCKQUOTE>

Only quite stable a/c have their guns unsynced. less stable a/c (stang/p47) have synced guns. Therefore, having unsynced guns in 4.02 would have probably led to enormous wing rocking due to the recoil, in some a/c - they way the 4.02 FM is.
And the p-40M was still given synced gun even though its quite stable, so that it doesnt feel excluded from its brethren http://forums.ubi.com/groupee_common/emoticons/icon_wink.gif

LStarosta
01-19-2006, 04:18 AM
I always thought it was Oleg's anticapitalist bias.

WTE_Ibis
01-19-2006, 05:11 AM
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by LStarosta:
I always thought it was Oleg's anticapitalist bias. </div></BLOCKQUOTE>
+++++++++++++++++++++++++++++++++++++++++++++++++

Now that can't be 'cause everyone's equal.

Although some may be more equal than others. http://forums.ubi.com/images/smilies/16x16_smiley-wink.gif

.

AustinPowers_
01-19-2006, 07:43 AM
Oleg is a logical guy
For Oleg to synch the .50 cals for the reason of lag is uncomprehendable.
He just has it wrong.

Tully__
01-19-2006, 07:50 AM
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Lordbutter4:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by FritzGryphon:
It's why weapon firing (or bomb dropping) is out of sync with client and server, yet damage is allocated according to hits that were observed by the client. </div></BLOCKQUOTE>

Are you sure about that? I have recorded many online sorties were I clearly remember hitting planes more then is shown on my recording. Try it...much more noticable when using 50's or 303's. </div></BLOCKQUOTE>
Recordings are not an entirely accurate representation of what actually occured.

Tully__
01-19-2006, 07:51 AM
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Stackhouse25th:
*pulls up chair, gets a bag of popcorn* </div></BLOCKQUOTE>
You and your fellow munchers are risking holidays for spam violations http://forums.ubi.com/groupee_common/emoticons/icon_wink.gif

Daiichidoku
01-19-2006, 09:42 AM
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Aaron_GT:
My theory is a bad code merge. In a number of instances we've had features/bugs come and go between patch versions. There are apparently over 1 million lines of code. So my theory is that developers have copies of the sim, make there changes to some modules to get the appropriate behaviour, and then merge these versions to form the new version to be shipped, but that there have been a few slips in this process. I am guessing, but it's not beyond the realms of possibility that the changes to the 50 cal dispersion code/data were tested initially on an older code base which had 'salvo' 50 cals (whereas the latest shipped patch did not have 'salvo' guns) and when the merge of the new code was done back into the main code base the 'salvo ' setting crept back in. We've had other things which suggest this - e.g. sniper AI gunners coming and going, and so on.

This isn't a dig at Oleg's team - it's a huge code base and the number of show stopping bugs has been immensely small for a code base this size. I know a bit about software testing, and the lack of bugs is a real achievement. But a few incorrectly executed code merges introducing non-showstopping issues isn't beyond the realms of possibility. </div></BLOCKQUOTE>

perhaps this explains the re-appearing fuel leak bug in 190?

tigertalon
01-19-2006, 09:46 AM
I did a test using only .50cals online, typing &lt;gunstat and checking my stats on several servers. On average I needed 50 .50cal bullets per kill. Most of the kills were made against fighters (approx. half LW and half IJA/IJN) plus a few twinengined planes. I got a lot of kills from damaged planes that RTBed.

And I still have 50 (!!!!!) .50 cals needed to get a kill. In ETO there was approximately 20 hits with .50 cals needed to down a fighter.

Go figure.