my profile | register | search | faq | forum home
  next oldest topic   next newest topic
»  COMBATSIM.COM Forum Archive   » Game Discussions (Genre)   » Helos   » Hope Richard Give me HOPE

   
Author Topic: Hope Richard Give me HOPE
Rush
Member
Member # 60

posted 10-25-1999 01:49 PM     Profile for Rush   Author's Homepage     Send New Private Message   Edit/Delete Post   Reply With Quote
I really have limited time online.
But I want to believe that I can have a CO-Op in Apache Havoc.....Please say YES
Im on cable in Hollland and my play mate is on Dial up in Canada....
Im getting sick of Sims that dont deliver online....Too many to mention

Regards

Rush


Posts: 35 | From: Amsterdam | Registered: Sep 1999  |  IP: Logged
Blaze
Member
Member # 198

posted 10-25-1999 02:49 PM     Profile for Blaze   Author's Homepage     Send New Private Message   Edit/Delete Post   Reply With Quote
I'm not Richard and therefor I can't give you hope. ;-)
AH's multiplay is really weak, it probably won't change (if at all) until the release of Comanche-Hokum.

Some claim to get successful games but in the end "successful" is a pretty subjective term.

Blaze


Posts: 1556 | From: | Registered: Sep 1999  |  IP: Logged
Rush
Member
Member # 60

posted 10-25-1999 03:02 PM     Profile for Rush   Author's Homepage     Send New Private Message   Edit/Delete Post   Reply With Quote
Thanks Blaze
Ive just tried Fleet Command.....its hopeless confusion, all you do is chase dots around not knowing what youre doing, man those guys know how to take our money........

Posts: 35 | From: Amsterdam | Registered: Sep 1999  |  IP: Logged
Richard 'Flexman' Hawley
Member
Member # 55

posted 10-25-1999 05:24 PM     Profile for Richard 'Flexman' Hawley   Author's Homepage     Send New Private Message   Edit/Delete Post   Reply With Quote
I *AM* Richard but I still can't give you hope.

It's as Blaze said. Maybe in Comanche-Hokum it will work.

It's kind of interesting as I was as confused as much as the developers as to why it performed so badly. Then they pointed me as some interesting data they just discovered about Internet protocols and the shitty performance the net generally delivers.

The most interesting nugget of info (and this applies to anyone wanting to write net games) was the TCP/IP which is *supposed* to gaurantee that packets are delivered in order...doesn't.

As much as 50% of you packets can go missing, and any packets coming in after a missing one will be out of order. This means a delay while the missing packet is re-sent. And sometimes that can get lost too. So the *net* effect (pun intended) is that deliveries of data can be delayed by as much as 50 seconds in tests done by "Totaly Games". So TCP/IP is complete *pants* for high traffic games over the internet.

It seems their solution was to use UDP packets instead of TCP/IP and add their own error correction scheme. They also "piggy backed" the previous packet onto the back of the new packet just in case the previous one went missing, this avoids the whole resending process thus cutting lag at the expense of a little more traffic. But given the nature of UDP packets there's little point in sending small packets anyway.

What does this mean? Well C-H (and hence A-H) may work a lot better. But ultimatley it's still going to use the permissable server/client model and it needs to send a lot of data. There are going to be some changes I think which reduces the number of "permissions" a client makes to a server thus reducing traffic.

That's as much hope as I can offer.

But as it stands, TCP/IP is at the mercy of the internet environment, meaning that traffic may by good or bad at any given time. This explains why some people have good connections and other are no hopers.

Hope that was of general interest,


Regards,
Richard 'Flexman' Hawley


Posts: 396 | From: West Yorkshire, United Kingdom | Registered: Sep 1999  |  IP: Logged
Enforcer
Member
Member # 85

posted 10-25-1999 11:30 PM     Profile for Enforcer   Author's Homepage     Send New Private Message   Edit/Delete Post   Reply With Quote
Pretty interesting Post Flexman...

Odd isn't it that a two years old sim is doing quite fine on Inet.

Oh! the name ? Longbow 2...

Enforcer. (without malice)


Posts: 342 | From: Brossard, Québec, Canada | Registered: Sep 1999  |  IP: Logged
Rush
Member
Member # 60

posted 10-26-1999 01:58 AM     Profile for Rush   Author's Homepage     Send New Private Message   Edit/Delete Post   Reply With Quote
Thanks very much Richard for that.
Its a big concern really, do they bother testing online when they beta test, if so they would find this out and be able to try and compensate for it in some way.
Or do they just ignore it and not bother about the 10% who go online..or at least try to.
LB2 seems to be very good online (havent tried it myself) is that down to good planning or just luck?
Im following the progress of B17-2, and I have to say that Im very concerned about the Multiplay aspect of it, with todays envirement who knows how disappointed we could all end up online.
As I mentioned with AH and Mig Alley to name just two MPlay is just non existant and we the online Sqds basically just waste good money on those Sims that advertise as being online capable, but are infact non-capable.
It seems to me that Graphics and MPlay dont get on too well.
Its all very frustrating.

Regards

Rush


Posts: 35 | From: Amsterdam | Registered: Sep 1999  |  IP: Logged
goon
Member
Member # 457

posted 10-26-1999 03:13 AM     Profile for goon     Send New Private Message   Edit/Delete Post   Reply With Quote
So how is it on a LAN? I've had it since it came out, but have only just got sufficient power to run it. It's still slow on low/low terrain and models, on a P266/160RAM/Voodoo2. Is this the norm? Er...back to the topic, is there any point in trying to persuade my friends to play on a LAN? Or do these issues affect this also.
While I'm here I think it was Flexman who wrote the excellent strategy guide I printed yesterday and contributed to the best session I've had on the game last night. That is to say I understood what the hell was going on for the first time! It seems AH has great potential, so is there any reason the scene is so quiet?
Thanks for any info....

------------------
I'm the famous Eccles!


Posts: 124 | From: Stafford, England | Registered: Oct 1999  |  IP: Logged
Richard 'Flexman' Hawley
Member
Member # 55

posted 10-26-1999 05:37 AM     Profile for Richard 'Flexman' Hawley   Author's Homepage     Send New Private Message   Edit/Delete Post   Reply With Quote
Rush,

The secret of the net-success of LB2 is down to the fact they *didn't* use DirectPlay or TCP/IP type packets (instead using their own datagrams) plus it has very very small traffic requirements due to the game universe restrictions (that and a good design). Having a whole game based on the principle of "low and slow" really helps.


Microsoft should share the blame for the recent rash of poor perfoming apps. They tote DirectX (of which DirectPlay is part). As programmers have realised, it's less than stella (pants more like). Someone at Microsoft has agreeed with them, DirectPlay 8 is being completely re-written as a result, sadly it won't be available for some time.


Each and every game is *different*, and require different approaches to handling communication. There's no free-lunch with this stuff. Each kind of comms model has pluses and minuses. It's up to the designer to decide which fits their game and data requirements best. And having a reliable network gaming API would help too.


When it comes to beta testing, it's difficult at times as the betas may not work correctly, or seem crippled somehow. Even scarier is the realisation that what you THOUGHT you were testing suddenly isn't. An example is a dev' studio testing internet play by having several office machines dialling into the net, they play around, it looks fine then they move on. Then weeks later they discover that it doesn't work, all they had done is test the POP performance (think about it - several machines from the same place dialing into the same ISP, the packets go up the pipe and then turn around and come right back - hardly testing the internet is it?)

So you do have a right to be concerned. As games get bigger over time, so do the traffic requirements. As 'net popularity increases and users play these ever more traffic hungry apps, game performance will suffer even more. "Just when you thought the internet couldn't get worse...it does".

I've seen betas so crippled that the term beta becomes a loose term. But there are still new games that are managing to handle net traffic. Contrary to popular opinion, programmers do learn from mistakes (it's mostly a matter of being ridiculed by ones peers <g> ) so we may see less ambitious games featuring internet play or better network models. If *only* the internet let you do multi-casting (sending the same packet to different destinations) it would be a different story.


Now, I think "goon" posted a question about LAN performance...(these boards don't let you see messages as you compose). A-H works OK, I'd use IPX and the interpolate play position switch though. Only issue with it is flying on autopilot (if a player is using the autopilot their world position is not updated via the network, as the AI is effectively flying).

Anyway, the reason the A-H scene is so "quiet" is that it's a better solo play game than an internet game. For reasons above and many more unknown, the scale of the campaign combined with traffic volume, net environment and design makes it a *bitch* to get working online. More luck than anything else, no one was more dissapointed than the guy who wrote the network code...except me perhaps (feck and bugger). Still, there's a lot of strategy in the game.

When it comes to frame rate, A-H is not texture rich, it's mostly loads of irregular polys and therefore CPU/FPU intensive. Any CPU which pipelines FPU instructions should *vastly* improve the performance of any high-poly sim. It makes the Athlon very attractive in this regard.

(And Longbow 2 is older than two years - and while I'm not disagreeing with Enforcer. LB2 is not the flawless internet app he believes, of course he can only comment from his point of view. Having a rose-tinted cable modem where brute force alone covers a multitude of data resending sins. If LB2 *was* so reliable, then net-squadron members would not have had to resort to email campaigns or flying missions solo when connections fell over. Yes it was good, but not perfect. JCN - same game, different server, complete pants. There are better performing games for low bandwidth users.)


Regards one and all,

Richard 'Flexman' Hawley


Posts: 396 | From: West Yorkshire, United Kingdom | Registered: Sep 1999  |  IP: Logged
Scout
Member
Member # 657

posted 10-26-1999 08:01 AM     Profile for Scout   Author's Homepage     Send New Private Message   Edit/Delete Post   Reply With Quote
Flex,
What do you make of the following? (Taken from B-17II FAQ over at bombs-away)
By the way, taking a UDP and adding error correction turns it from connectionless to connection-oriented protocol. That, in my book, is Diet-TCP. It doesn't of course matter how you call it, what counts is the performance.
----------QUOTE OFF------------
We are supporting multi-player through both DirectX and our own IP sockets stuff. There are effectively two exes; the game, which can be a server (more in a minute) and the relay.
The relay is the bit which is best hosted on a fast server somewhere on the net. It does not host games at all, merely acts as a disribution point for all the object status packets which are sent. It knows which objects are closest to you, and prioritises packet transmission to your machine on that basis. (You get more frequent updates for the objects 'in your face')
Anyone can host a game from his machine at home. When other people join, the objects (planes, crewmembers, AI's etc) get distributed to the most sensible PC using a similar mechanism to that described for the relay, and run from there.
The objects are all distributed and properly client server. So, after a short period, there is no longer really a server; everybody participates in hosting various parts of the game. If a player disappears, control of that object reverts to the original server until it gets re-distributed.
That is roughly how it works. The current plan is to provide the relay in the box, so anybody can host either a relay or game. I don't believe that a pay for play system is on the cards. However, I cannot speak for our publishers, so this might change in the future.
-------------QUOTE OFF-------------------

Cheers,
Scout


Posts: 715 | From: Israel | Registered: Oct 1999  |  IP: Logged
Richard 'Flexman' Hawley
Member
Member # 55

posted 10-26-1999 01:04 PM     Profile for Richard 'Flexman' Hawley   Author's Homepage     Send New Private Message   Edit/Delete Post   Reply With Quote
Diet-TCP, I like that term :-) very descriptive, mind if I nab it? But yeah that's what we're talking about. Real-time games *are* connection orientated applications.

What you quote sounds like the done thing these days. Avoiding DirectPlay like the plague and writing your own error handling (required when using UDP for this sort of app).

The "client bubble" server sounds classy. It's the sort of technology that's needed for large scale internet warzones. Reduces traffic and I suppose delivers a form of "dynamic multicasting" for want of a better term (if you get my meaning).

Still, if you have 20 attacking AI planes all in the vacinity of a bomber, crewed with five or clients (players) then that's still be a lot of traffic for anyone to handle.

Suppose it depends on how they script the AI and how the "bubble" code works out.

But it's sound theory. I guess the biggest problem is how the AI handles the long periods when there is no data coming in. Of course a clever designer would fudge it so that no more than x number of AI objects approach y players and swamp the connection. Sort of like scripting some of the AI planes break up and "go around" in threes or fours.

There are many subtle ways smart designers can work a game to fit the limitations of the internet. Some of them are not that smart however.

I saw B17II TME in London last month and was well impressed by nearly everything demonstrated (the silk sheet effect landscapes are a bit hard to grasp in contrast to the aircraft detail).

Certainly they have put a lot of work into it.


Posts: 396 | From: West Yorkshire, United Kingdom | Registered: Sep 1999  |  IP: Logged
Scout
Member
Member # 657

posted 10-26-1999 02:54 PM     Profile for Scout   Author's Homepage     Send New Private Message   Edit/Delete Post   Reply With Quote
Flex,

Yeah, I think this relay system may really rock, but then again it may really blow. The big if is the design of course, done right it will rock, done slightly not so right it will blow. So if them Wayward boys found the right formula, WOW. Personally I don't mind not seeing huge formations of bombers in the air all filled up to the tailgun with on-line simmers. I will be happy if I see 3 piloted by humans plus 3 escorts on each side with no lag at all. This will make me happy as hog.

P.S Blanksheet terrain? Have you seen the latest screenshots? www.bombs-away.net - goto screenshots section. I've seen the VHS video Wayward sentme, and let me tell you, this one is really unbelievable! The athmosphere is there, the wholeness of the experience,as in the little things that make up the whole.


Cheers,
Gripes


Posts: 715 | From: Israel | Registered: Oct 1999  |  IP: Logged
Rush
Member
Member # 60

posted 10-26-1999 03:24 PM     Profile for Rush   Author's Homepage     Send New Private Message   Edit/Delete Post   Reply With Quote
Gents

Thanks again for your comments.
B17 certainly looks the part, but the above mentioned will certainly take some implementing for it to work as desired.
Ive also got the VHS pre alpha video and it looks impressive...I must say that the Wayward crew are really a great bunch of guys...and I wish them all the success they deserve.
Anyway my point is, and has been that to get B17 to work in multiplay you have to be looking for realistically 3 online in each bomber X 4 bombers = 12...at this stage I would like to add that Wayward are optimistic about having ONE plane host the online players...4 online escort pilots ..say in a P51 = 16, and a attack group of 8 in 109`s for example.
The grand total adding upto 24.
Now 24 online players in a sim this georgous in 16 calculated real time objects is a LOT.
As I said they are optimistic, and I want to believe them, but the fact is can that possible run stable most of the time, with players running all different connect systems?
I will refrain myself from answering this,as I am pro Wayward, but it certainly gets my mouth watering if they can.
We will have to wait and see, but release is at least 5 months away and that in real terms is a lifetime...........

Regards

Rush


Posts: 35 | From: Amsterdam | Registered: Sep 1999  |  IP: Logged
Richard 'Flexman' Hawley
Member
Member # 55

posted 10-26-1999 07:14 PM     Profile for Richard 'Flexman' Hawley   Author's Homepage     Send New Private Message   Edit/Delete Post   Reply With Quote
Thanks for the site pointer. My goodness, I see what you mean. The terrain has come a long way since I last saw it.

I was gobsmaked when I saw the planes and cockpits being put through their paces but that cultural backdrop is really quite fine.(still looks a little flat at low level though.)

The networking stuff you just described sounds awfully ambitious, mind you Warbirds managed it. Well, if they don't get it right first time...there's always the follow up patch(es).

Bags I get to be the radio operator.



Posts: 396 | From: West Yorkshire, United Kingdom | Registered: Sep 1999  |  IP: Logged
Richard 'Flexman' Hawley
Member
Member # 55

posted 10-26-1999 07:16 PM     Profile for Richard 'Flexman' Hawley   Author's Homepage     Send New Private Message   Edit/Delete Post   Reply With Quote
Thanks for the site pointer. My goodness, I see what you mean. The terrain has come a long way since I last saw it.

I was gobsmaked when I saw the planes and cockpits being put through their paces but that cultural backdrop is really quite fine.(still looks a little flat at low level though.)

The networking stuff you just described sounds awfully ambitious, mind you Warbirds managed it. Well, if they don't get it right first time...there's always the follow up patch(es).

Bags I get to be the radio operator.


Posts: 396 | From: West Yorkshire, United Kingdom | Registered: Sep 1999  |  IP: Logged

All times are MST (US)  

   Close Topic    Move Topic    Delete Topic next oldest topic   next newest topic
Hop To:

Contact Us | COMBATSIM.COM Home

© COMBATSIM.COM, INC. All Rights Reserved.

Powered by Infopop Corporation
Ultimate Bulletin Board 6.04b

Sponsor
© 2014 COMBATSIM.COM - All Rights Reserved