Jump to content

Whats is the advanrage of a high traffic in sim?


bebejee
 Share

You are about to reply to a thread that has been inactive for 3472 days.

Please take a moment to consider if this thread is worth bumping.

Recommended Posts

No they don't get paid.  Traffic is used as part of the equation to determine the order your venue shows up in search.  The higher your traffic the higher you potentially are on the list, although other things figure into it also.  People want to be as high as possible on the list because they think when people search they will gravitate towards  and choose those venues at the top, thus leading to more people coming and therefore more potential customers.

 

Link to comment
Share on other sites

There are 2 search systems in the viewers. The LL viewer uses the only one that does a good job. In that system, traffic is *probably a very small factor in determining the rankings. The older search can be used in some 3rd party viewers and, in that one, traffic is the only factor in determining the rankings. So higher traffic counts make higher rankings in the old search that is still used by some people.

*I said 'probably' because the current search's predecessor did use traffic in a small way as a ranking factor. It's probable that the current one does too but we don't know for certain - at least I don't know for certain.

Link to comment
Share on other sites


Phil Deakins wrote:

There are 2 search systems in the viewers. The LL viewer uses the only one that does a good job. In that system, traffic is *probably a very small factor in determining the rankings. The older search can be used in some 3rd party viewers and, in that one, traffic is the
only
factor in determining the rankings. So higher traffic counts make higher rankings in the old search that is still used by some people.

Oh...

I didn't realize they still had the old system in the code for the TPVs to use.

This explains why I always see people saying SL search is broken, and yet I get such good results with it.

In the current one, I CAN use traffic when I search places, it is then one of the sort options - but not the default.

 

Traffic was a lot more important to a place back when shops were all online. Back then - if you could get bodies to your location, you could attract commercial venues, and charge higher rents to them. Now merchants get around that by selling on Marketplace, and/or an inworld "mainstore" that is a branded experience - where the brand is valued, but the traffic doesn't matter (I've even been in stores that try to keep down the number of people in them - they just want to be there as a showroom alongside a flickr group or something).

 

Now... the only benefit of having high traffic is social - people know "if I go there, I'm more likely to meet people also interested in that place's theme."

- For the actual venue and its occupants though, too much traffic just means more lag (thus why even some stores try to keep visitors down to a smaller 'managed' set).

So these days, "medium traffic" is probably ideal for socializing.

 

Link to comment
Share on other sites

bard1.JPG  bard 2.JPG

 


Pussycat Catnap wrote:


Phil Deakins wrote:

There are 2 search systems in the viewers. The LL viewer uses the only one that does a good job. In that system, traffic is *probably a very small factor in determining the rankings. The older search can be used in some 3rd party viewers and, in that one, traffic is the
only
factor in determining the rankings. So higher traffic counts make higher rankings in the old search that is still used by some people.

Oh...

I didn't realize they still had the old system in the code for the TPVs to use.

This explains why I always see people saying SL search is broken, and yet I get such good results with it.

In the current one, I CAN use traffic when I search places, it is then one of the sort options - but not the default.

 

A lot of the complaints you hear are from Merchants, Club Owners and related who for no explicable reason fall off the radar in Web Search.  One day they are in the top 10 results and the next day you can scroll through twenty pages of results and still not find them.   So when you are searching for say, a tail, you may not even realise that there are results missing.  They are refering to the SL Web Search, not the legacy search.

But lets take for instance something like people search which I posted screen shots.  Can't rememeber exactly how someone's name was spelled?  Which result do you think gives you a better chance of finding them?  The old legacy search with about 30 results or the new web search with it's measley 6?

Link to comment
Share on other sites

That's an interesting set of legacy results, Perrie. What viewer did you use for them?

The reason I ask is because I'm looking at the results for 'bard w' (no quotes) in Singulariy's legacy search and they are different. First it sent a message to say that the 'w' was dropped because it's too short, and that it was just searching on 'bard'. Then it showed my all the Bards in alphabetical order - and there are only 3 whose surnames begin with 'W'. So where all yours came from, I don't know.

In the main search four are listed with a surname beginning with 'W' and they are all different to the three in the legacy search. So between the two searches I came up with a total of seven.

Link to comment
Share on other sites

sing search 1.JPG

sing search 2.JPG

 


Phil Deakins wrote:

That's an interesting set of legacy results, Perrie. What viewer did you use for them?

The reason I ask is because I'm looking at the results for 'bard w' (no quotes) in Singulariy's legacy search and they are different. First it sent a message to say that the 'w' was dropped because it's too short, and that it was just searching on 'bard'. Then it showed my all the Bards in alphabetical order - and there are only 3 whose surnames begin with 'W'. So where all yours came from, I don't know.

In the main search four are listed with a surname beginning with 'W' and they are all different to the three in the legacy search. So between the two searches I came up with a total of seven.

I did do my search with Firestorm, same as Coby.  I just tried Singularity and it is odd that it drops the "W."  If I just search "Bard" with no "W" in Firestorm I get the same results as Singularity, every name that has "Bard" in it.  We'd have to ask the Singularity Devs why the "W" gets dropped.

I can only think that Web Search is picking up Fenn's profile because of the keyword "Bard" in it.  But quite frankly, though I know it is possible, I find it hard to believe that he would be the only person in SL who had the word "Bard" in their profile.   

Main point is, which is better?  In my example where you can't remember someone's exact name Legacy is far better.

I could go on and illustrate the superiority of Legacy Search when looking for events.  Web search is badly hobbled, I could say that actually it is "broke." It certainly is not meeting our needs the way that it both could and should.

Link to comment
Share on other sites


Bobbie Faulds wrote:

For the web search, you have to do just like you would for the internet. If you are looking for Bard Wxxxx but aren't sure of the spelling, you need to use the wildcard and search Bard W*. I did and got 76 results in people

 

 

Now that is very good to know. 

But here you have three experienced users who I am sure all of us know what a 'wild card' is and none of us thought to use it.  I'd be willing to venture that the majority of Users don't even know about 'wild cards.'

It's just like using Boolean terms in the Market Place.  How many people actually know them, much more even think to use them?

I know that back when I was using the Official Viewer 1 with what we now call Legacy Search that I never had to use a Wild Card.  In other words, Linden Lab has broke something.

Link to comment
Share on other sites

I'm pretty sure I have the answer :)

When they were using the GSA, they intentionally omitted people who had not logged on for a while. I think it was 30 days but I have an idea it may have been an even shorter period. It was to prevent a fair amount of bloat in the index; i.e. names that haven't logged in recently are not indexed, which means that millions of names are left out and, therefore, cannot be listed in the results. They are sure to have continued that idea with the replacement engine, which would account for the few results in the current search. They are the ones who have been active quite recently.

The legacy search doesn't do anything like that, of course. It just churns out every name there has ever been that matches the searchterm, which accounts for the larger number of results.

Perhaps searchers on names want people listed who haven't logged in for a while, in which case, for 'People' searches, the legacy search is better. But that's all it's better for. Generally, the current search is a heck of a lot better than the legacy search, which isn't even a search engine. The current search is, therefore, far more trustable to produce good results than the legacy search - except if you want all the currently inactive people listed.

It doesn't answer the problem you mentioned of places appearing to drop out of the index though. That was also a problem with the GSA. But it does explain why the two systems produce such different results for that People search.

Link to comment
Share on other sites

bard wasp.JPG

search 3.JPG

 


Phil Deakins wrote:

I'm pretty sure I have the answer
:)

When they were using the GSA, they intentionally omitted people who had not logged on for a while. I think it was 30 days but I have an idea it may have been an even shorter period. It was to prevent a fair amount of bloat in the index; i.e. names that haven't logged in recently are not indexed, which means that millions of names are left out and, therefore, cannot be listed in the results. They are sure to have continued that idea with the replacement engine, which would account for the few results in the current search. They are the ones who have been active quite recently.

The legacy search doesn't do anything like that, of course. It just churns out every name there has ever been that matches the searchterm, which accounts for the larger number of results.

Perhaps searchers on names want people listed who haven't logged in for a while, in which case, for 'People' searches, the legacy search is better. But that's all it's better for. Generally, the current search is a heck of a lot better than the legacy search, which isn't even a search engine. The current search is, therefore, far more trustable to produce good results than the legacy search - except if you want all the currently inactive people listed.

It doesn't answer the problem you mentioned of places appearing to drop out of the index though. That was also a problem with the GSA. But it does explain why the two systems produce such different results for that People search.

I'm really thinking of this from a users point of view.  In a sense I couldn't care what search engine they use.  What I want is results.  I do want them to be as accurate as possible, but results are still what I want.

If I know how someone's name is spelled, I get the one result in Legacy Search.  And while I understand that someone might decide to look up everyone with the name "Juran" (A few of us did that when we started out group "Juran Family Cousin's Club"), the more likely scenario is like the one I stated, they don't remember the spelling when looking for a specific person.

As to your theory about people being being dropped (not indexed????) if they haven't logged in for a while, the first result for "Perrie" is an account I own.  I snagged that the day  User Display Names went live.  It's been a long time since I last logged in with it. 

With Web Search, maybe we do need to get back into the habit of using Wild Cards.  We've all been spoiled by multiple services that don't require their use.  A goal any service should have is to make things as simple as possible for everyone.  I'll bet the percent that ever think to use them is miniscule.

Lastly, I don't think that Web Search from it's very start had anything to do with making search easier for us.  Search became a ba$tard child to what it's really all about, the revenues from advertising.  Nothing wrong with LL making money.  But the search should not have been its ba$tard child.

 

eta: corrected a term

Link to comment
Share on other sites

Well that was my theory, anyway :) I do know for sure that, with the GSA, they dropped all but recent logins from the index to reduce the index size. That's not a theory. That's what happened. And, to me, it seems likely that they continued it, although i don't understand you 'Perrie' thing. I don't know what you mean by "when user names went live". Do you mean display names?

You are probably right that the web search was never about making it easier for us, but I think you're mistaken about it being about LL revenues. They were getting those ad revenues before the GSA, and the GSA was extremely expensive to lease - and they leased 3 of them. Imo, the GSA came in simply because it's an actual search engine - they didn't have one of those - and, as such, it produced quality search engine results, which the previous system could never do. The results were orders of magnitude better than the previous system, given the limitations of using an actual search engine for SL.After that, they got an open source engine that they could tailor as best they could to suit SL. They could never do that with the GSA because they had no access to the source code. I haven't studied it to see if they've made it suit SL better than the GSA did, but I have no doubt whatsoever that, because it's an actual search engine, it produces good results that are orders of magnituded better than the legacy system's effort. The legacy system was never any good at producing results - unless you want People results to list every name that ever was that contains the searchterm, of course. It only ever listed stuff from a database, sorted according to traffic or alphabetically - no standard seach-egine-type criteria taken into account, excapt an exact match.

So I still disagree with you about which system is more trustworthy for people to find what they're looking for. An actual search engine produces much better results that a simple database listing, and that's what I'll go with.

Link to comment
Share on other sites

live 1.JPG

LEGACY SEARCH

 

live 2.JPG

WEB SEARCH

info 2.JPG

 

info 3.JPG

 


Phil Deakins wrote:

Well that was my theory, anyway
:)
I do know for sure that, with the GSA, they dropped all but recent logins from the index to reduce the index size. That's not a theory. That's what happened. And, to me, it seems likely that they continued it, although i don't understand you 'Perrie' thing. I don't know what you mean by "
when user names went live
". Do you mean display names?


I did mean to say "Display Names."  My bad and I'll fix so not to confuse anyone.

 


Phil Deakins wrote:

 

You are probably right that the web search was never about making it easier for us, but I think you're mistaken about it being about LL revenues. They were getting those ad revenues before the GSA, and the GSA was extremely expensive to lease....<snip>

 

I used the term "Web Search" because I was referring more to the 'format,' not the engine driving it.  But before "Web Search" where were the Ads displayed?  Web Search in the Viewer added value to the Ads and added value is what sells services.  In other words Linden Lab was saying, "Here is another reason you should buy classified ads with us."

 


Phil Deakins wrote:


So I still disagree with you about which system is more trustworthy for people to find what they're looking for. An actual search engine produces much better results that a simple database listing, and that's what I'll go with.

"Trustworthy?"  How can you say it is "more trustworthy" when the results are hobbled?  I am interested in "relevance."  If I'm searching for "blue shoes" I don't want to see "red shoes."  But if it is failing to show me all the blue shoes that are available when I ask it for it as in Bobbie's example when she used the wild card then it is not trustworthy.

Lastly, my biggest bugaboo with Web Search has been its Format.  I'm not as big on this as I used to be but I used to go to a lot of live music performances.  There were many musicians I enjoyed greatly.  With the old format I could open search and basically at a glance see who was performing and all the info about it.

Now with Web Search I need to start scrolling.  And scrolling and scrolling and scrolling.  And woe be if I want to look at all the info. I have to click not just once but twice and then wait for that new window to open in my Viewer.  And what ever the cause may be, I've had people tell me they've given up on that because of how long it can take for that last page to load.  My results vary with it.  But if I need to wait for 15 to 30 seconds for that page to load, and sometimes it is that slow, it starts to get very time consuming.

 

 

 

Link to comment
Share on other sites


Perrie Juran wrote: 

Phil Deakins wrote:

Well that was my theory, anyway
:)
I do know for sure that, with the GSA, they dropped all but recent logins from the index to reduce the index size. That's not a theory. That's what happened. And, to me, it seems likely that they continued it, although i don't understand you 'Perrie' thing. I don't know what you mean by "
when user names went live
". Do you mean display names?


I did mean to say "Display Names."  My bad and I'll fix so not to confuse anyone.

I grabbed Phil when display names came out and when I do a search for Phil now in the web search I get plenty of what appear to be display names; e.g. Phil (PhilipMax). I don't even see mine listed as just Phil like your Perrie is. Anyway, I don't know why the display name for an account that you haven't used in a long time shows up, but I still suspect that the index size is kept down by not indexing names that haven't logged in for a while, and that that is the reason for the large difference between the two systems in the numbers of listed people.

Phil Deakins wrote: 

You are probably right that the web search was never about making it easier for us, but I think you're mistaken about it being about LL revenues. They were getting those ad revenues before the GSA, and the GSA was extremely expensive to lease....<snip> 

I used the term "Web Search" because I was referring more to the 'format,' not the engine driving it.  But before "Web Search" where were the Ads displayed?  Web Search in the Viewer added value to the Ads and added value is what sells services.  In other words Linden Lab was saying, "Here is another reason you should buy classified ads with us."

Maybe I'm remembering wrongly, but I'm sure that people could pay for higher rankings before the GSA came along. In fact I think I did it. I seem to remember a few people paying extraordinary amounts of money for top rankings. So much that there was no way in the world that the money could be recovered through sales. I remember a discussion in the forum where one of them said that it was an experiment and that the money was simply lost - gone. That certainly happened but I can't be absolutely certain that it was before the GSA but I think it was.

ETA: Yes it was before the GSA but the payments didn't get you up the normal ranking. It was the Classifieds tab. So you may be right about LL seeing the GSA as a means of selling advertising. Even so, those GSAs cost a fortune to lease, and they leased three of them.

Phil Deakins wrote:

So I still disagree with you about which system is more trustworthy for people to find what they're looking for. An actual search engine produces much better results that a simple database listing, and that's what I'll go with.

"Trustworthy?"  How can you say it is "more trustworthy" when the results are hobbled?  I am interested in "relevance."  If I'm searching for "blue shoes" I don't want to see "red shoes."  But if it is failing to show me
all
the blue shoes that are available when I ask it for it as in Bobbie's example when she used the wild card then it is not trustworthy.

You can't get reliable "relevance" without a search engine, and, before the GSA, there was no search engine is SL. Search engines rank on relevance as best they can. All we had before the GSA was listing from a database, ranked on traffic. If the searchterm appeared in the place's description or name, then it got listed regardless of whether or not it was relevant. E.g. if you were searching for 'blue shoes' and that phrase appeared anywhere in the description, whether or not they were for sale, it would be listed. If you searched 'blue stilleto shoes' and that exact phrase didn't appear in the parcel's description or name, the parcel wouldn't be listed and you'd miss a lot of blue stilleto shoes. So the legacy search for blue shoes could
not
list "
all
" the blue shoes that are for sale. It could not do what you said you want it to do. The web search is much more trustworthy to do that job for you.

With a search engine, you can search for 'blue stilleto shoes' and get them listed even when that exact phrase does not appear in either the parcel's descritpion or name; e.g. 'red, white, blue, gold, and green stilleto shoes'. A search engine
can
do what you want it to do.

There is no doubt at all that a search engine produces relevant results far far better than a simple database listing, which is exactly what the legacy search is. That's not to say that a simple database listing doesn't list relevant results IF the exact searchterm is contained in the parcel's name or description, but what it can't do is list the place that sells blue stilleto shoes and puts 'red, white, blue, gold and green stilleto shoes in its description. A search engine can and does list such places.

Lastly, my biggest bugaboo with Web Search has been its Format.  I'm not as big on this as I used to be but I used to go to a lot of live music performances.  There were many musicians I enjoyed greatly.  With the old format I could open search and basically at a glance see who was performing and all the info about it.

Now with Web Search I need to start scrolling.  And scrolling and scrolling and scrolling.  And woe be if I want to look at all the info. I have to click not just once but twice and then wait for that new window to open in my Viewer.  And what ever the cause may be, I've had people tell me they've given up on that because of how long it can take for that last page to load.  My results vary with it.  But if I need to wait for 15 to 30 seconds for that page to load, and sometimes it is that slow, it starts to get very time consuming. 

I did the * Events search and the differences between the legacy and the web search results are:-

(1) Legacy lists only 200, even though it says there are more, whereas the web search says there are 6340 upcoming events. I don't know how many it will actually show but it'll be at least 1000. That's no biggie because you won't want events so far into the future.

(2) Legacy doesn't show any description in the listings whereas web search does, so legacy can fit more into a smaller space. In spite of that I think it favours the web search because it provides some decription.

(3) With the web search there are popularity indicators with the listings to show you how busy the event is. There is also another indicator to show the type of land the event is on - adult, moderate, etc. The legacy system doesn't do anything remotely similar to those. That's a very big plus for the web search.

(4) With the legacy you can TP to a place from the same floater after you've clicked a listing to get it's details. With the web search you can see some of the details already and you can click to TP to the event without needing any further clicks. I think that favours the web search too because to TP to an event, the legacy system requires one extra click to get to the TP button, but the web search requires no extra clicks because the TP button is right there with the listings along with a partial description which the legacy doesn't have.

(5) With both systems you can get the details of a likely-looking event (more details - longer description - with the web search). Both are instant (no waiting for a page to load), and both offer the same map and TP buttons.

(6) This is where the web search is worse. From the instant 'More info' expansion, you can get even more info from a button. That one opens a new floater with the full description, whereas you already got that with the legacy search. So that's one extra click in the web search to get the full description - IF the description part you got already got isn't enough for you but, to be fair, I do think it's sufficient when looking for an event you might enjoy.

(7) This is a big plus for the web search. In the new floater you have an option to set a reminder, so you carry on doing what you want and you'll be reminded when the event is due to start.

All-in-all, I think the web search is very good for events and far superior to the legacy listings. I don't understand your preference for the legacy system's events listings.

 

Link to comment
Share on other sites


Phil Deakins wrote:


Perrie Juran wrote:

 

I did mean to say "Display Names."  My bad and I'll fix so not to confuse anyone.

I grabbed Phil when display names came out and when I do a search for Phil now in the web search I get plenty of what appear to be display names; e.g. Phil (PhilipMax). I don't even see mine listed as just Phil like your Perrie is.

 

When I type Phil Resident in the legacy search (in Firestorm), the first name what shows up is Phil.

When I type just Phil in the search there are thousands of all kinds names which have the part "Phil" in them.

Phil Resident.jpg

 

In the web search typing either Phil Resident or Phil (Sort by, User Name: A to Z) it appears that plain Phil user name does not come up at all. There were 1814 results, I got tired of scrolling to see that was Phil included in the results. To make it easier there should be an option "search this word only, exactly as it is written and nothing else".

 

Anyway, it indeed looks like that user name Phil is already dropped off from the web search.

Link to comment
Share on other sites

You are about to reply to a thread that has been inactive for 3472 days.

Please take a moment to consider if this thread is worth bumping.

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
 Share

×
×
  • Create New...