[david_c]

Rebound, on Mar 26 2005, 03:43 PM, said:

This reply is sufficient to pretty much address every concern I may have had regarding the changes. The way you describe it, Fred, it sounds perfectly fine to me :-)

Yeah, I agree There are a couple of things that I don't quite understand though:

Firstly, it seems that we will still be able to see everyone who is online. Which is great news for the user, but surely it's just as much of a burden on the server as the online list is at the moment. Now, at the moment, on average one in every n pieces of information sent is used to update the online list (I don't know what n is, about 20?). I suspect that this n is a constant, not depending significantly on the number of people online. If so, then as long as we are going to have a complete online list, we'll never be able to improve the situation by more than a factor of n, no matter how many mini-BBOs there are. And that means, if the membership of BBO increases by a factor of sqrt(n), which we hope will happen eventually, the servers will be just as busy as they are at the moment ...

Secondly, if there are lots of near-identical mini-BBOs, how would you prevent the situation where everyone joins just one of them? The bad way of doing it would be to put a limit on the number of people in each one - that's bad for very obvious reasons, but it's the way this is done on many other sites. Is there a good way?

[Walddk]

I have read enough to determine that Fred and Uday are on the right track. I am no technician, so as a simple layman I would like to state:

Continue as you see fit. I am sure that whatever you do, it will be in the best interest of BBO and its members. There will always be some who don't agree with your decisions, but that applies to anything you will or will not change.

Forget about pleasing everybody, perfectionism does not exist. Do what the majority would think is a good idea for as many as possible. I think that this thread has clearly indicated what that majority wants ......

... and you, Fred and Uday, have clearly told us what will be involved when the changes are implemented. I will not lose any sleep over it. The new look BBO, whenever that may come, will definitely be a place I would want to visit as often as I do now.

Who said 24 hours a day?

Roland

[scoob]

i should have read this thread before i posted in the other one

[DrTodd13]

Fred is totally aware of this issue I'm sure but for those who aren't and are suggesting it, one could really mess things up if you adopted the approach of "we'll only keep one area in synch at a time" where area is main lobby, tourneys, tables, vugraphs, etc. If I pop into the tourney area and have to get a full tourney list, that will take some time. As it is now, it takes essentially 0 time because we are constantly being updated. Then I decide to pop back to the main lobby. Now, assuming that I stopped receiving lobby updates when I was looking at tables or tournaments then now I have to get the whole lobby list again. If I pop back and forth a lot and do the obvious (but stupid) thing is registering for updates when I join an area and desregistering when I leave then you can really increase server load this way.

Personally, the main lobby being always in synch regardless of where you are in BBO is not a bad idea. You just have to define main lobby as your friends rather than everyone and you have to move the filtering function from the client to the server just like Yahoo does it. You could still get a list of everyone but that would be on an on-demand basis.

[rigour6]

Ok, a few cents again.

1) These (i.e growth) problems are good problems to have.

2) It is difficult for people to see the need for this now because the problem hasn't yet run us over. We had a taste of it when the rush to the Vugraph slowed things, but so far the system works so well, people don't see the issue. If the system maxes out, performance doesn't just decline - it plummets.

3) Right now, the lobby isn't much use because it's too full. So why have it? Dump that off somewhere and only provide that information on user requests. In other words, use the space for things that are useful.

4) In a similar vein, how useful is a Main Room with hundreds and hundreds of tables? How many are you going to play at? When the system fills, players will have frustrations because they'll go to sit down only to find the seat taken, on the next one same thing. Again, spin that off. Or spin off tourneys.
In other words I would much rather see a system that split players by what they were doing than where they lived - even though I know one of the main gains in response comes when you can get people choosing a localized entry point.
I say this because even though groupings by language are natural, to me one of the great advantages of the bbo community is its cosmopolitanism. My guess is the most natural seam would be North America / Everyone else. Without getting into a whole thing here, imho the world needs less things in which it's America / Everyone Else.

5) Now my one perhaps decent suggestion (it's your reward for reading to end): has BBO done any study into what users do? In other words, what are the migratory patterns of a "typical" BBO user. My guess is there won't be a single profile, but there will be identifiable groups. There'll be the users that ALWAYS go to the Main Club and stay there. There'll be the users that usually stay in the Tournaments.

Find out what the migratory pattern of fish are and you know where to have your fishing seasons and when. Find out the migratory pattern of BBOs, and you'll know where to put your seam. What if it turns out that 48.6% of all users log ONLY into the Main Room and stay only there? Now you've got a big increase in functionality if you put a seam there - sure people will move back and forth but you add a step there and they cross from one server to another.

Follow those fish and see where they go.






btw, 48.6% of all statistics are just made up on the spot.

[shoeless]

Hey Rigour6 as Stephen Leacock (humorist and professor of economics at McGill in Montreal) said - "There are lies, damn lies and statistics" Mostly I just want to add my add my support to Roland's comment and expand on it slightly - Leadership is about doing the right thing and since BBO in all it's aspects is coursing through your veins - I can't image you will be much off the mark whatever you do. Your report card so far looks pretty much like all A's. Go for it! Of the 10 people who quit because they don't like the change - 48.6% of them were going to quit anyway. My biggest concern on BBO to this point is: "What nick is De Falco using today?" Put a bell on him and I'll be happy.

[hrothgar]

shoeless, on Apr 14 2005, 08:04 PM, said:

Hey Rigour6 as Stephen Leacock (humorist and professor of economics at McGill in Montreal) said - "There are lies, damn lies and statistics"

I believe that the quote originated with Samuel Langhorne Clemens...

[rigour6]

Benjamin Disraeli

[hrothgar]

rigour6, on Apr 14 2005, 10:10 PM, said:

Benjamin Disraeli

This is the problem with growing up in the US...
We tend to ignore a lot that happened out side out continent

Up until a minute ago, I thought that Clemens had originated that quote

[shoeless]

I can now go to bed tonight secure in the knowledge that I can misdefend, misbid, misdeclare and misquote. Statistical analysis of the probability of this occuring in one 24 hour period to follow - likely 48.6% of the time.

[vincenzo]

I agree with mini bbo's idea.
Very inetresting for local federation !!
Very useful for technical reasons Uday has explained
And also all other points Fred and Uday have written !!

Vincenzo

[melviss666]

Well, for my two cents worth, I ain't convinced on these separate BBO's. If a separate BBO for Vugraph & private clubs, & even ACBL tourneys would fix the problem, then that's what I'm for. Maybe even add WBF or other point-awarding organizations as well to this other BBO. And furthermore...
I like the lobby list, irregardless of how unwieldly some might think it is. I surely met at least half my friends thru the lobby before EVER playing with or against them at a table. I've started chatting to folks I didn't know at the time because of something interesting in their profile & folks have started chatting to me by the same token. I like that. And furthermore...
All my other user names agree with me on this.
Best Regards,
melviss666

[DrTodd13]

This is an argument for being able to see everyone logged in and their profiles. It is not an argument for forcing everyone to have their lobby list constantly updated. The people who are suggesting getting rid of the lobby as it currently exists all say that you should be able to explicitly request the complete lobby list. This small inconvenience to you would hopefully gain a lot in terms of BBO scaling.

[adf]

I've read lots of posts on scaling and am not fully convinced of what is right. However, I am convinced that this is an important issue; more important in the long run than pretty much anything else discussed in the "suggestions" forum. Consequently, I would urge Fred and Uday to do the "right" thing, whatever they decide that is, rather than the "easy" (or at least not as hard) thing, even if that means a delay in getting some fun features in the system.

That being said, I don't consider it necessary to keep people at tables (playing, kibbitzing, vugraph, etc.) fully in synch. Maybe there is a subset of information that would be broadcast, such as the full list of table and player names; if a player navigates away from a table, their client would request to be fully resynchronized; in the mean time, at least some information could be updated on their screen. The server could also determine when it had some spare time and try to keep people updated even when they didn't ask for it (as it does now, but 100% compliance wouldn't be necessary), so the amount of information that needed to be sent in the event of a request would be minimized. It may also be that the server should know some of the client's settings (which it doesn't know now) in order to optimize just what needs to be sent. Yes, this is more work, but it may be worth it.