Screenshots

Just ASS("U" and "ME") that the system is not manipulated, not only MAHATHIR has been dropped, even BADAWI has not been seen either...

Maybe all of us are just too fed up of all this nonsense!

Resorting to daily than cumulative popular search term may be worthwhile but the search results is not as.

Clicking on "siti nurhaliza" resulted in only one solitary news. That is all. 2004 news is not shown for archives from 2003 to 2004.

Google revealed three pages from NST. And it's free. And Wiki has her too.

It's fun having it there in NST, to gauge the interests of readers, though this is a bit early to adapt; only six terms were searched. The site is not 2.0 anyway.

Btw, Siti's homepage is here: http://www.sitizone.com/, not certain if it is the official website, though it is claimed to be so.

looks like they just did a reset on the counter. a not unreasonable reaction after a tabloid picked up a non-issue and blew it all out of proportion.
if the previous result reflected real search pattern, it's just a matter of time before the pattern re-emerge. so let just wait and see what are malaysian fetishes.

Ok, if anyone not satisfied with NST result, go to google.com and add "site:http://www.nst.com.my" behind your keyword to dig out nst.
example
mahathir site:http://www.nst.com.my

ok, its purely a tech thing, so i don't see why its blown out of propotion.

the way a tag cloud works is that any term keyed into the search feature gets thrown into the cloud. "even if there is 0 results". And tag clouds have time frames which they work on, so if the term is no longer being searched within the frame it is dropped.

The odds of that 3 terms being the most searched within a 2-3 hour period is not very hard to understand, as you can see that suki (one in a million) is somewhat popular today and odds are that you might see "siti nurhaliza rape suki" later today. Thats just how the clouds work, and it is not based on the 'amount of content on the nst site' but the 'number of times a particular keyword has been searched'

Tags clouds are independent keywords, and the odds of a clouds having a keywords that combine to create a sentence always happens, just smile. Seriously, don't blow things out of propotion when it really is nothing more then a coincidence.

JEFF OOI says: I watched from the sideline as I must say I am amused at the way Screenshots readers just TEMBAK IKUT DAN, so speak Kedahans. If you have a site that runs Eurekster.com like of keyword search tags and Screenshots uses this pioneer small apps on this blog -- the clue I gave away right from Day I on September 22 -- you wuld know that the the webmaster/site owner MUST define (or select) keywords of his own choice.This gives room to human bias, preferences,and yes, manipulation. Once the keywords are chosen and tagged, the engine, by way of a small script, will take over to determine which one should be bold and big. And I said-- right from Day One -- that technology itself is innocent. Other variables? I am not so sure. Bottomline, I can see how the layar butuk among Screenshots readers just shot from the hip. You know who they are, and I am not pinpointing this particular commenter I conveniently dropped my response. Common guys, write your own keyword tag script or just intall Eureskter.com to learn for yourself before you jerk again!

I did some tests. I searched the word "sucked" like 10 - 20 times and it came out in the list. Then, I searched "this is a test" like 50 times and it came out in the list. Then, I searched "script working" like couple of hundred times, it came out in the list. I did it using a automated web test tool. The thing I found out was the list is not in any particular order, but random. So, I guess that "mahathir ... rape ... siti ..." was just coincidence. The only thing I do not understand is why those words were highlighted as larger font size?

Just found out that the larger the font size means the more popular it is.

yes, larger fonts means more searches done with the word tags. and it is in a random order, that is why u see scertain words larger and in bold and not in order.

i stil believe its an innocent Jscript.. simple

yes, it can be manipulated. A few days ago I typed SUKI in the search box and pressed enter quite a number of times (>20 times!).. today I noticed that "SUKI" has a bigger font!!!!
So if we will to type a few words....... what will come out?

Coincidental?
Try yourself.

Sheep, you cannot make any phrase you like to appear there. If the search does not return any articles, it will not register. Tested.

which demonstrate my point. any keyword searched that has results will appear on the list, there is no such thing as a "webmaster/site owner MUST define (or select) keywords of his own choice" any keyword (with results) searched a preset number of times will appear on the cloud - and the more searches it has the larger it becomes. Every refresh moves the keywords around, and that particular result was just a coincidence. Having select keywords defeats the purpose of having a popularity cloud. (see flickr and technorati's search clouds if you still need proof)

i've got better things to do then TEMBAK IKUT DAN on this blog.

JEFF OOI says: To those who are prone to TEMBAK IKUT DAN, and prone to disinform and misinform the public with half-past-six google-quotes, open up and accept the fact that in site and web searching, filters and parameters CAN BE defined and updated by publisher hosts (read: human interference of technical neutrality) --thus technological neutrality can be manipulated for good or for bad.

The closest model of this so-called search-tag is again, I quote, Eurekster.com/Swicki technology that this blog has been using -- and testing -- for almost a year now. URL: http://www.eurekster.com/products.

The keywords that appeared on the frontpage of Screenshots WERE indeed _chosen_ and _determined_ by me. Hence the keywords/tags appeared THE WAY I HAD INTENDED. In other words, I can influence, manipulate and dictate the type of tags I want appear to suit my specific scope of research. (Yes, I run Eurekster.com/Swicki script on this blog to study readers search behaviour over an extended period of time. There is something scientific and methodological about the research approach.)

If you want to talk web search technology, then let's get rid of as many blind spots you may have as possible. So as NOT to misinform and disionform the unassuming public.

By the way, I am not ridiculing you, but "swicki" is an expression to refer to the new trend for "search wiki". What NST uses is a form of this new trend -- search wiki. Being a Swicki, HUMAN (webmaster) INTERFERENCE is not a remote impossibility though technology, in itself and by itself -- I repeat -- is innocent .

Need proof? Need no proof. Technorati and Flickr.com that you mentioned are two different animals. I have personally tested the two applications (I have a paid Flickr Pro account and this blog/this blogger is ranked constantly by Technorati -- so I know the subject matter that you raised as so-called "proof" well enough. But, frankly, I don'tknow if YOU are being ranked by Technorati as intimately as I am?) The former is a blog directory while the latter is a community-generated content cum P2P portal -- and both use their own indexing and algorithm methodologies to network content creators and users via search-active platform. They are only a search service at the secondarily level as all Web2.0 apps will acceed to "collective intelligence" and the "string-able" search functions form the scritical base for mass usability -- aka their primary relevancy.

Whatever said and done, to understand the context of this blog entry, just be humble to explore available web search technologies -- for eg install the Eurekster.com/Swicki -- and you will a closer-to-home perspective. Don't rely on stupid gut feel. It may cheat you at times. To help you, here's the URL: http://swicki.eurekster.com.

i doubt it will make any difference today, buy i'm adding this for future reference. I might be wrong, if so i'm sorry, but i'm going to state the facts none the less.

There are tag clouds, there are search+tag clouds and there are search clouds to the best of my knowledge. There is no official definations of search clouds but they exist in different forms (a couple months and it will be clear)

The most popular is the tag cloud, where content has got relevant tags, and these tags are then used to generate the clouds. The most clicked on, the most visited tags grow in size, while unpopular tags dissapear into tiny unreadable dots. These tag clouds are popular in blogs (where content is tagged accordingly, or there are a bunch of categories that all content would fall under) These clouds basically show the user in one glance whats the most popular topics in the site/blog.

Search+Tag blogs are what eurekster is, where specific tags are set and searches made and results returned from categories/tags will determine the size of the tags on the cloud. Its a compromise of telling a general user what people are looking for, without revealing what exactly the search term that was most popularly searched really was. If there were 1 million different search queries done today, you woudn't expect eurekster an external site to keep track of each of those queries and track it, hence the selective tags.

However in a pure Content Management System environment, like newspaper sites, searches are conducted by the site itself, and all internal search queries can be stored on the internal database. Polling out these search phrases and displaying them as is "without having to relate it to a tag/word" is what a search cloud does. No tags, no categories, just the exact term keyed in by the visitor. Yes, it is a risky concept because you're revealing what exactly a user is searching for, and as long as there is related content on your site, it appear on the cloud. Friendster once practiced a similar concept and got a lot of bad publicity because there were some very explicit searches conducted by some members - which turned up in the "what my friends are searching for" list. Back to my point, if you look at the cloud now, it displays things like "one in a million", "this is a test" and so on. Would these terms have been manually keyed into the database? They are there because people searched for them. A simple test would debunk your tags are manually entered theory. If one of the earlier commentors searched for "mahathir malaysia" instead of just mahathir, it would also appear on the tag cloud - something which would not happen if the cloud was based on manual tags.

I'm not for or against anybody here, but there is an innocent guy out there who might loose his job for putting in a simple script with no malicious intent whatsoever, and luck has it that those words appeared in the way they did and he is made the scrapegoat.

JEFF OOI says: You spoke a lot and you have drifted as much from the original context. Period.