Definition of Theft?

I wonder if the issue with not finding bnb and pillow talk under poses is the words used in search. I believe that technically they sell animations and not poses for the most part. It may be that they come up very high, due to relevence, under the things they have the most of and things they sell that have those specific words in the title. Im pretty sure pillow talk comes up under "pose pillows."

Also, i use all search pretty much exclusively. The reason is because i can search for a certain term under places, people or groups alone and get no results returned. But, i find that i can use the exact same word in search under all search and get results that include both people AND places. It seems to capture the results differently and with a finer gauge net so not as many things get passed over. I do find that i have to be more specific just like i do when using google on the web. For chandelier, i might have to type sculptie chandelier to get one that isnt just flat prims. Or maybe crystal lighted chandelier in order to get one that has a light script etc. i know for my own stuff....it returns different results for blanket than it does for blanketS. if folks have not thought of that when naming their items or descriptions, then things wont come up in search all if someone is one letter off what they did use.

With that said...i doubt either bits and bobs or pillow talk are worried about where they rank since their stuff is so well used and loved all over sl. items being out in the world go a long way to promote the store as people see them in use and click to find the creator. So much less motivation for them to work to be on the first page....

Wouldn't that be kind of self-fulfilling, though?

The more you sell, the more you sell!

coco

I'd expected this response and was all prepared to say that it doesn't matter: it's the effect that matters, and if the standard in-world search mechanism doesn't turn up those two in a prominent position, it is simply not going to be useful. And I was all ready to agree that this was an acceptable *algorithm* but that Search is more than the algorithm, it's the whole system including how well people are educated in using it--and so, for Search All to be successful, it must be the case that all merchants know everything there is to be known about how it works. And then the fix might be as simple as making completely public all the details of anything that can affect ranking and all the tricks of the trade for affecting those

However, I've been able to get in-world now, and played around with our sample search string. Now I'm beginning to think that GSA itself is just crap, at the algorithm level (or conceivably the configuration). One of our test cases, Bits and Bobs, comes up number 2 with the search term "couple pose" but is nowhere on that front page with "couples pose", "couple poses", or "couples poses". Is it common knowledge that stemming is that weak?

Did you know that, with the search term "low prim sofa", your shop is buried deep in the results?

I thought the search algorithms would automaticly try variations, like adding S in the end of words with plurals and vice-versa, and I have a fain t memory of reading somthings written by a Linden mentioning the search engine in SL now did that...

You are correct here is the blog posting URL:

http://blog.secondlife.com/2008/03/26/tips-to-improve-your-search-ranking/

QUOTED FROM PAGE:

perhaps I was remebering it worng then, I was thinking that if somthing with shoe instead of shoes was ranked higher it would show before any results for shoes, I thought it would make the variations be treated as the same word, or at least not be ranked worse than the worse exact match

But the information is all there Qie. Even published by LL. So the fact that most merchants do not optimize their parcels for search, cannot be blamed upon the search engine. If I put up the best website ever about a subject, I will have to undertake some action to make sure people can find it. What many do forget, is that a big part of virtual business is making sure people can find you.


You do have a good point there, they should come up as well. That is why I said long before, that the current engine could use some tweaking. The GSA is not crap, but the implementation of it could be improved.


Yes I know your question is to Phil, but I can answer that one as easy. You simply, as a merchant, cannot be found under every keyword people type in. Phil can type in low prim sofa now, but tomorrow someone searches for low prim table and again he might be low ranked. We have to try and optimize for the keywords we think are most used. When I look to my own parcel, maybelow prim lamp and low prim fireplace are not resulting in high positions. Low prim on its own, does result well though. It is a matter of choices.

Well poke me with a stick then, because relevancy is the answer.

The fact that new search can be influenced means it is just like every other way of search. No search engine can find out by itself how relevant a website really is, if the owner does not optimize for it. If owner A had the best one but did no optimizing, and owner B did, then owner B will end up higher. And you can see a parcel owner just similar to a website owner. It is part of the business to make sure people find you. After that, it is up to the people to buy or not buy.

The GSA is a round hole and SL is a square peg. I think everyone has pretty much agreed with that from ages ago. That said, it is still the best search system that SL has, and LL did give everyone the main details of how to optimise for it before it was launched, but they should also put those details in the KB and/or wiki - maybe they've done that - I don't know.

If people don't or won't optimise their parcels for it, then it's their own fault if they don't show as porminently as they think they should. No search engine can determine the places that *should* rank highly. It's up to the land owners to do something about it, and the search system can't be blamned, but that's what some people here try to do. Search systems are 2-way - we help it and it will produce. They can't do it on their own.

BnB and Pillow Talk are "the top 2" in Mickey's book for "poses", but they are not the top 2 in other people's book, and there's no reason why those two should be the top ranked ones. I've heard of Pillow Talk but I don't know their stuff, and I don't think that BnB sell poses. They sell animations. Mickey is searching on the wrong thing. The conclusion that the search system is crap because it fails to rank at the top what *I* think should be at the top, is a mistaken view. No search system can do that. Search systems can only rank what's given to them, and if parcel owners don't do it, then the parcels won't rank highly. We've been through it all before, and the only type of system that can do what Mickey wants is a human edited directory.

Stemming:
According to LL, the GSA does do stemming, but Google's stemming isn't what we may think it is. It doesn't treat all words with the same root as equals. It includes pages with plurals and such in the results set, but it biases the rankings towards exact word matches, and rightly so. So the results when searching on "pose" will be biased towards those with "pose" in their text rather than those with "poses" in their text. Stemming means that there is no need to include the word variations in the text to have the page included in the results set, but there is every need to include the variations to rank higher for each of the variations. LL didn't mention that because they probably don't know it.

[correction]
I've just seen, from Lord's post, that LL does understand the Google stemming, and they wrote about it.
[/correction]

I didn't know - I've never looked. It may be surprising but I don't do a great deal of optimisation for search, but maybe I'll do something about that phrase

Quiet, let grandpah speak for a moment!
In ye slightly older days of the interweb, we're talking ~1995 now, one of the mayor search engines actually wasn't completely automatisized.
Yahoo ran with hundreds, if not thousands of volunteers for ages, who manually visited websites and assigned scores to them, based on the keywords presented and their respective relevance. This basically meant that a sex site that had "chairs" in its keywords would get no score for chairs, not even if they had a model on their main page sitting on a chair, simply because the relevance was 0.
A site selling furniture would get a very high score, while a site specializing in chairs would receive the maximum score.

One didn't just become a Yahoo volunteer out of nowhere at first, so volunteers' scores were pretty accurate. But after a while, the internet grew so exponentially that it was basically impossible to keep up with it, the restrictions on becoming a volunteer lowered, fake volunteers entered, giving fake scores, so even Yahoo went for the automated search results after a while.

Well, Yahoo is still a known name on the internet, but for some it's probably hard to imagine that that site used to be as big back then as Google is right now.

But it does tell us one thing: to know what the good places are, you better rely on human results than on search results, so help out the good places, and tell your friends about them!

I didn't arrive on the web until 1998, and I didn't know that Y! editors were volunteers. I always assumed they were employees.

Then we'd have shops paying people to buy stuff.

What I feel about shops in general, are the search keys used.

When I search for a specific item, all sorts of shops pop up on the first results, that after checking, don't even sell those items, or have one poorly generated item that kind of match my searches.

It reminds me of the old google fight, getting on top regardless. It's annoying, but I can't do anything about it.

True, keyword spamming is something that is hard to filter out. But I think such shops hardly can have a long life, since they do not attract the customers they want.
If I attract a customer, I want to be sure that I deliver what he/she searched for. Maybe none of my items appeal to this particular customer, but at least they will not feel cheated.
What I could do, is optimize my keyword listings for some sex items, something people do search for as well. Now I can hardly believe that someone searching a prim appendage, will walk out m store with a couch But they will remember me as the store not delivering what they promised.

Well, but this is all very "search-centric"--it assumes that the mechanism provided for finding stuff in-world should be based on text search. Now, I'm not dismissing that whole proposition, but I'm saying that these two places are almost surely among the top few that people would want to visit if they type in that search term. If people don't get that result, they're not being well-served by in-world search, however in heck it's implemented, and regardless of whose fault it is. The end result is that Search is not as effective for advertisers, either, because people just give up on it and switch to a third-party provider.

Also, I just must say that the Linden-supplied information about the new search is barely even the tip of the iceberg. Remember how much trial and error it took to figure out what was required for an account to have a page, so its Profile Pick could count? Remember how much longer after that before somebody finally quantified (correctly?) that it was 30 days of not logging in until an account quit counting? I mean, these and countless more are all decisions somebody made in the GSA implementation--all stuff we don't get to know, except empirically. That's just stoopid.

(Mind you, absolutely none of this matters to me as much as the fact that LL has--for no apparent reason--lobotomized its own domain such that in-world scripts can't access search.secondlife.com. That means scripters can't help make the results of Search be more accessible or useful in-world. My jira: http://jira.secondlife.com/browse/SVC-3122.)

Well, I call them poses....you call it couch, I call it sofa.....you call it stool, I call it ottoman.....you call it lounger, I call it chaise lounge.....you call it bookshelf, I call it bookcase.....you call them drapes, I call them curtains.............

I'm tired of guessing and assuming how people shop and search, so I took a poll of my store group, and the notecards are flying in. Well, duhhhhh! I'm tweaking everything and revamping everything according to what they told me.

Since I have no "Influence".....I won't go into detail, and I'm not going to exploit my group's kindly participation and effort into that......but I will say, that if you put any more effort into analyzing any of the above, you are wasting your time. Except for the original concept posted, that Bots Suck....you might want to pay a little attention there.

And if you're a store owner that does not use bots, and believes that B2B referrals and Influence works, and if you treasure a drop in the bucket sale from a customer, on any given day, and if you appreciate every customer that walks in your door, or sends you a new customer.....well, you just give me a call, and I'll get you hip to the process, and send you a few drops in the bucket.

*pokes you with a stick... repeatedly* I warned you .

You're the very last person who should be arguing that search results are "relevant" since you're manipulating search left and right. The mere fact that you have a direct impact on how your store ranks contradicts that search results are ordered by "relevancy".

The same goes for Marcel when he started his picks incentive campaign and literally jumped ahead by a few pages. He didn't change a single thing about his store, he was as "relevant" the day before as the day after but he tricked search into considering him more "relevant" than he was and true relevancy of the search results worsened as a result.

When all you need to do to become more "relevant" is to play games with search then you do not have a "relevant" search, you have a manipulated search where whoever has the biggest bag of tricks ranks before everyone else.

Or it could be that your definition of "relevancy" is simply "phrase matching" but that would be about the lowest possible expectation to have from any search. If I download all the places pages onto my puter and search for "kitty ears" in file contents all the same results will turn up.

BTW-your whole quote is a contradiction by itself... if I can't find what I want by using search it is by definition not returning results that are relevant to me.
And of course there's also the matter of perspective. Your one and only concern is that people find your store and you couldn't care less whether the results really are productive or not, just as long as you can manipulate them to your liking.

QFT

Oh lookie! A pointy stick!!

Poke, poke, poke

I do that with the fashion blogs in general (or individual creators' blogs) .

I'm still trying to fill the gap left by StyleDisorder though . Although their main focus was on fashion, they'd also mention new useful "gadgets", modelling poses, or just fun "everyday for everyone" things that popped up.

For things like furniture, general knick-knacks, plants/trees, landscaping, etc I haven't found anything though . Just tp'ing around to stores I like and look at where they got things from so far seems to be the best way to find anything.

The major appeal of a blog is simply that it's so much more time-effective to skim through all the new posts and catch up. No need to tp, rez, look around, go "meh", move on, rez, look around, etc.

*gasps*

I'd never use a pointy stick, baaaaaad Sling .

I call then:- sofas, curtains, etc. However, poses are static and animations are animated. You can call them what you want, but that's what they are. If you go looking for poses, don't be surprised or disappointed if you don't find animations.

I'm sorry that you are not as influencial as you perhaps think, Mickey, but that's also the way it is. I don't need to give any thought as to whether or not bots suck. I'm perfectly happy for you to hold that opinion. You are not alone. I don't think they do, and I too am not alone.

That's what relevant search results are. You search for "kitty ears", you get results for "kitty ears". If the store isn't selling "kitty ears" then that result shouldn't be in there because it isn't relevant.

What Phil is doing is optimising his parcel so that those who search for a product he sells will find his parcel first. That doesn't make his parcel irrelevant.

AMEN!!!

That has to be one of the MOST RELEVANT and most concise, well said statements I have heard on that topic.

You are dead wrong, Kitty. I optimise my place for the very things that I sell. All my optimisation enhances the relevancy. When my place appears in the search results, it's *because* it is relevant to the search term. I make sure of that.

You said it youself, Kitty - the results are not relevant to *you*. I just don't understand the mentality that insists that search results must be relevant to *me* personally - they must be what *I* want or they are crap. Never mind that they contain what I searched for - chandeliers, for instance - the top ranked places didn't have the sort of chandeliers that *I* want, therefore the search engine is crap. It doesn't make any sense whatsoever.

It might have been if it wasn't so wrong.

Thank you for the correction, Phil. When I do my notecard referral system on bot-free stores that make quality products, that will go out to 100's of shoppers, (perhaps 1000's, if they start handing those out to their friends), you know, that drop in the bucket system that doesn't work.....I will be sure to use the word Animations. Very handy info. Thank You.