Bazaar: Vendor + Buyer capabilty

[Maevenniia]

I did

[glenbois]

Would like to see trader and buyers standing in the same section - I think that would be a big improvement to start with - I do think it's too much hassle to go from one area to another for a lot of people.

I would also welcome a combined buyer/trader option too.

[Telorea]

You need to think outside the box more

The big thing that is needed is to make it so that you don't need to stay logged in when you want to sell stuff in /Trader mode or in /Buyer mode.

Not only would this reduce the masses of lag in bazaar and be environmentally friendly (by reducing the 1000's of PCs around the planet that get left running just to be AFK in bazaar) - but it would also be much simpler to impliment than trying to design a decent interface that can combine both buyer and trader mode.

The /barter interface is a tough problem to design properly - and the current design does not work well if there are a lot of buyers.

One way of implementing a better search function would be to search by item-type, similar to the list of item types in /trader search window. But even that would get tough to search if there were a very lareg number of buyers.

[Fonceur]

Actually, I wouldn't go into combo buyer/trader mode if it did exist simply because don't have enough bag space to sell all of the stuff I want to sell now. There is no way I would want to leave open spaces to buy with.

I agree there. Unless the buyer/trader could make use of the bank slots too...

[Faerwynd]

I would personally love to see something like how you buy items in FFXI..... why should we have to stay logged in to sell? Buying could be set up roughly the same way. IIRC, one reason they split up buyers and sellers was to help reduce lag, this would be the next step in that direction. I'd like to see houses as well, but thats another discussion

[Elmarnieh]

I would think that SOE would love to find a way to have us not be logged in and eating up bandwidth when we go in any mode.

I dont see why we couldn't get an in game NPC alliance to charge us 100 pp/rl day to hold say 50 items for sale and 50 items we want to buy and at what prices and then just pick up the money/items from them. Acts as a money sink which we all know SOE loves and uses less bandwidth at the cost for a little more code and perhaps and maybe a bit more processing power but that would be negligble.

I mean even if it was simply to test it for an upcomming game its a good idea.

[Chick-in-chainmail]

be nice to have both togeather, BUT I would really rather see them let you trade while off line. With getting booted off from ISP after a couple hours (or less) I can't just leave a trader/buyer up .

[Itek]

*sigh*

Saying "this would be nice" is great and all, but the problem is... someone has to actually make the code work. For a multitude of reasons this is non-trivial. "Off-line merchant agents" attached to each account sounds easy. Let me assure you it is not.

1) SOE doesn't get a discount on THEIR bandwidth expenses if you aren't connected. They buy it in bulk, and have to pay for it "lump-sum" at the "peak bandwidth" price. If they have 80,000 logged in Friday night their providers don't ONLY charge them for the 30,000 logged in Tuesday morning. Savvy ??

2) There has to be an "object" for a buyer/seller. Doesn't matter to the server if that object is interactive (a PC) or not (an NPC) so having your agent "logged in" is just as rough on their resources as having you logged in.

3) So, effectively, your "agent" is actually a FULL NPC Merchant. With all the concommitant expendatures of resources.

4) Now you want to put an "agent" for, in all possibility, EVERY logged out player into the bazaar. If you think this will lessen lag in the bazaar you haven't thought about it long enough.

5) Now you are faced with the option of splitting the bazaar up into multiple zones. Which kinda defeats the purpose of -having- a bazaar in the first place.

6) Then throw in a half-dozen software issues I haven't listed out since I don't work for SOE and don't have the specifics of how their software interacts from object to object. Suffice it so say that software engineers / designers are paid, usually well, to deal with these problems and the solutions usually take hundreds of man-hours to resolve.

7) Then learn the horrifiying news... Generally the productivity of software designers is ONE source-line-of-code-per-hour.

8 ) Now learn that there is generally one Customer-Reported-Unintended-Defect (read: BUG !!) per new source line of code. (Fixing things generally breaks things. This is why working on Legacy-Software, and EQ is in parts almost a decade old, is generally not people's favorite activity.

9) They certainly looked at this way WAY back when they developed the bazaar and barter interfaces. They didn't do it, so likely they had a -reason- for not doing it that way. (money or difficulty, or even just 'vision')

I mean come on, devs do PLAY this game, and most of them have at least HEARD of other games. (yes, that was SARCASM) While "agents" would be nice I'd rather they fix existing bugs than introduce more trying to insert them.

And yes, I too am unable to keep a bazaar mule up reliably. So this would be of interest to me, I just don't think it's worth the effort it would certainly take to implement.

[Kahmon]

Much as it would be nice to have a trader up when not logged in, here is the ultimate reason why it will never happen.

Many people have multiple accounts for the sole purpose of having full time vendors (and sometimes buyers too).

If Sony were to allow you to use a charactor slot as a vendor then log it off then many people would cancel their secondary accounts. In essence Sony would be spending money developing code in order to lose money in closed accounts.

[Bobaten]

What I would like to see is this:
Take the bazaar search window we already have.
Change the "Find Trader" button to "Browse Trader"
Clicking that button would open up the merchant window for that trader, without having to run across the bazaar to find them.

[Telorea]

Buyers-Vendors

Responses in line:

*sigh*
Saying "this would be nice" is great and all, but the problem is... someone has to actually make the code work. For a multitude of reasons this is non-trivial. "Off-line merchant agents" attached to each account sounds easy. Let me assure you it is not.

1) SOE doesn't get a discount on THEIR bandwidth expenses if you aren't connected. They buy it in bulk, and have to pay for it "lump-sum" at the "peak bandwidth" price. If they have 80,000 logged in Friday night their providers don't ONLY charge them for the 30,000 logged in Tuesday morning. Savvy ??

You are forgetting that not all players live in the same time zone. Hence peak bandwidth = peak playing time 'live' players + AFK buyers & traders from different time zones. Thus the peak numbers shall be reduced and SoE WILL get a discount if all those AFK traders are not logged.

2) There has to be an "object" for a buyer/seller. Doesn't matter to the server if that object is interactive (a PC) or not (an NPC) so having your agent "logged in" is just as rough on their resources as having you logged in.

3) So, effectively, your "agent" is actually a FULL NPC Merchant. With all the concommitant expendatures of resources.

4) Now you want to put an "agent" for, in all possibility, EVERY logged out player into the bazaar. If you think this will lessen lag in the bazaar you haven't thought about it long enough.

You are assuming that a not-logged-in trader would be represented by an NPC buyer/seller. That is the wrong assumption to make because that would be a very inefficient and more complicated way to do it - so I doubt SoE software designers would use your assumed bad design.

5) Now you are faced with the option of splitting the bazaar up into multiple zones. Which kinda defeats the purpose of -having- a bazaar in the first place.

I don't think anyone suggested that, why do you think that would eb needed?

6) Then throw in a half-dozen software issues I haven't listed out since I don't work for SOE and don't have the specifics of how their software interacts from object to object. Suffice it so say that software engineers / designers are paid, usually well, to deal with these problems and the solutions usually take hundreds of man-hours to resolve.

7) Then learn the horrifiying news... Generally the productivity of software designers is ONE source-line-of-code-per-hour.

This is simply not true. Once you count the full efforts of software project managers, designers, engineers, coders and testers then the number of lines of source code per man hour is very low - but it is not that low, even for safety critical systems.

8 ) Now learn that there is generally one Customer-Reported-Unintended-Defect (read: BUG !!) per new source line of code. (Fixing things generally breaks things. This is why working on Legacy-Software, and EQ is in parts almost a decade old, is generally not people's favorite activity.

9) They certainly looked at this way WAY back when they developed the bazaar and barter interfaces. They didn't do it, so likely they had a -reason- for not doing it that way. (money or difficulty, or even just 'vision')

I mean come on, devs do PLAY this game, and most of them have at least HEARD of other games. (yes, that was SARCASM) While "agents" would be nice I'd rather they fix existing bugs than introduce more trying to insert them.

And yes, I too am unable to keep a bazaar mule up reliably. So this would be of interest to me, I just don't think it's worth the effort it would certainly take to implement.

[Itek]

Clicking that button would open up the merchant window for that trader, without having to run across the bazaar to find them.

To the best of my knowledge there is a -reason- you have to be within a certain physical distance of a trader. (Something to do with an exploit or potential exploit, no I don't have specifics, no I don't want to know specifics.)

Hide all traders, use the "mystical path" from find and it's certainly workable, fast and reasonably bug free.

Telorea- Immediately go back and edit YOUR comments OUT of the "Quote = Itek" text as is customary. You aren't quoting me, so it should not appear that you are. (PS. It's generally accepted practice, also, to only quote the parts to disagree with, rather than the whole text, to save on bandwidth both for users and owners.)

They must have "peak bandwidth" well in excess of their actual useage. Squeaking by = NOT smart. Reducing the number of online accounts by AT MOST 750 on every server would not significantly discount their bandwidth price. (PS. I've been pricing bulk bandwidth for nearly 15 years. Don't tell grandpa how to catch rabbits.)

Yes, I do presume that there will be an "object" on the server with both buy and sell lines. (Pretty much equivelent to a NPC merchant, no?) The "not-logged-in" buyer/seller (a buyer who is also a seller is, effectively, an NPC merchant, actually a more complicated object as NPC merchants generally buy EVERYTHING where the new "agent" would not) MUST be represented as a software object (a resource used by the server) and must have some sort of visibility to ALL the players (and non-players) as well as the server.

See, if you want your "agent" to buy/sell for you while you are off-line the server still has to have the "agent" loaded in active memory. It must be searchable and useable by all logged in (and potentially non-logged in, if "agents" can make trades with each other, gosh, I actually have THOUGHT about designing such an economic/game model, perish the thought) objects. Even if you don't "draw" a toon-model it still takes up memory on the server and calculation time for the client. As there is no downside for "offline" people to ALWAYS leave an "agent" behind when they log off, it's presumptive they would do so.

a) "agents" take up server memory

b) "agents" take up "logged in" player calculation time (even if not rendered as a model)

c) "agents" take up no user resources for the "offline" person leaving them behind

d) effectively 80% or more of the population would choose to leave "agents" up given the possibility

You say my idea is bad, but offer none of your own. My idea clearly would work "out of the box" (NPC merchants currently work) while you can't say why (other than the fact that it would be a resource hog) it might not.

Your point is that "doing it that way would take a ton of resources" which, if you go back and read it again, is PRECISELY MY POINT !!! Not that you have a better way, just that you seem to WISH there WERE a better way.

"Why would they need to break up the bazaar?" (also, eb = be I think, I make and speak typo-ese as well)

Because, gee, they have a hard limit on the number of traders now, for a reason, and they used to have bazaar / barter split, for a reason, and if you dramaticly INCREASE the number of bazaar/barter mules with agents you will hit the same server memory limits? You can't load an infinite quantity of objects into a finite amount of RAM. And RAM is much more expensive than simply splitting the zone up. (It's actually cheaper to run two cheap servers than to buy expensive RAM to upgrade one server. Again, don't tell grandpa how to catch rabbits. He's been at this a LONG time.)

Contact James Collofello (Asst Chair Undergrad Comp Sci @ ASU) and argue with him about 1 SLOC per hour. Given his expertise, my personal and extended (family friends etc) experiences I'm sticking with 1 SLOC of NEW-CODE per hour. And SOE (the O is capitalized too, Sony Online Entertainment) certainly views the transactional security of traders to be mission-critical. For evidence of this just refer back to the "rollbacks" after the Beta-Vendors debacle.

1 SLOC / hour (new code, not code-reuse)
1 bug / SLOC (you had two typos in 4 lines of text, I had to edit out 3 of my own so far in my reply)
1 object in memory, regardless if it's "rendered" or not
Absolute transcational security requirements for bazaar mules (duping = BAD)

While more robust "agents" then have even been discussed in this thread are the subject of much of my own thought I don't see SOE wanting or needing such agents, simply because what they have works and is not a hinderance to the economy (such as it is) of EQ.

Wanting isn't the same as needing, and it's also not the same as having a workable design. If you have a workable design for a software object that takes no memory while providing services I'm sure the software engineering community would love to hear about it. You can't simply spec "I want a buy/sell agent, with no rendered model, which takes up no memory space, is searchable in precisely the same manner as existing objects, and allows full interaction with active players as other logged in buyer/seller objects." That's the same as saying "I want our client to be secure, and fast." ... Define "secure" and "fast" and we can talk.

There's a reason software engineers hate building use-cases and requirement-docs by talking to Legitimate-Users. Because L-Users have so little idea of what's possible, and are so amazed with what's been done already, they PRESUME that they can get it just by asking for it.

[Telorea]

My main point was that you are basing all your assumptions (on resources, number of traders, need to break up bazaar into several zones, etc) on the premise that there would have to be a one-to-one relationship between the logged out traders and and the number of 'agents' required to represent those traders.

You are absolutely correct that if you represent each logged-out trader by a separate 'agent'/'NPC'/whatever, then you don't free up much if any resources - which is why a design like that (as you were suggesting was the only way to impliment it) would be a very bad idea/design. You need to open your mind to alternative possibilities.

By thinking outside the box more, it doesn't take much imagination to junk the idea of limiting a solution to a one-to-one relationship and also produce more efficient methods of storing/accessing the data so that more data can be stored without the need for more RAM or spliting zones. Yes, there would still be limits to the number of traders/buyers, but given the extremely inefficient current implimentation it could be significantly higher than the current limits.

You don't need to represent every trader/buyer as a separate object (be it an NPC merchant or whatever) you could potentially represent every single trader on each server within a single database - think of the current bazaar-search window - why not have that as the single interface to all traders and buy straight from the window without having to run around anywhere.

The 1 SLOC per man hour is an out of date figure, it used to be true around 15 to 20 years ago, but certainly is not now. I would assume that it was some years ago that your lecturer told you that figure, but times do change - just because it may have been true 20 years ago does not mean it still is - this is why most professionals need to continously update their knowledge to ensure it remains current.

I am basing my knowledge on extensive *current* personal experience in a field that develops safety-critical systems, systems that if they fail could cause serious injury or death, so I certainly know what I am talking about.

Apart from anything else, there are many different types/classes of software development (UIs, Databases, Networking, Signal Processing, etc) that require differing levels of manual activity that it is entirely impossible to quote a single figure for lines of code per man-hour and have it be meaningful at all.

[olhoss]

The issue with barter is not the system...I think this works fine. The issue with barter is that the buyers' offers are generally so ridiculous that most people ignore the whole system, so they do not realize someone may be offering to buy their stuff. (While on barter for Treant Branches I would check /baz and often find them for sale cheaper than for what I was offering to buy them for).

I would expect to see someone offering less than market price for an item, since they are giving a seller the convenience of the instant sale. In some cases, like with MDS, this has been ok....buyers offering 15K in barter while the market was 20K for sellers...this is in the ballpark, if a little low, but doable.

The problem is that most buyers badly lowball their offeres....Discordant Scoriae going for 65K and buyers offering 10K. Blades of Carnage selling for 80K getting offers of 30K.

If barter is going to work for the community as a whole, we need to be realistic about what we offer for items. To me, 20% discount from the going price seems fair. Until those that buy regularly are willing to cough up something realistic for the items they want, they cannot expect the barter system to get a serious look from the public at large and to become a major part of the game.