homepage Welcome to WebmasterWorld Guest from
register, free tools, login, search, pro membership, help, library, announcements, recent posts, open posts,
Become a Pro Member

Home / Forums Index / Google / Google News Archive
Forum Library, Charter, Moderator: open

Google News Archive Forum

This 87 message thread spans 3 pages: 87 ( [1] 2 3 > >     
What searches does Google fall down on?
Let's turn it around a little bit

 1:27 am on May 22, 2002 (gmt 0)

In the thread about Google Labs, I asked what searches people wanted to see. Now let me turn it around and ask--what searches aren't great on Google? Are there any times you start to head to Google and then think "You know, Google's not going to do well on that. I'll go somewhere else." ?



 1:40 am on May 22, 2002 (gmt 0)

that's a tough one.

Maybe if I want to find PURELY commercial pages only.


 1:40 am on May 22, 2002 (gmt 0)

Well, then there's always...
Nothin :)


 1:42 am on May 22, 2002 (gmt 0)

There just aren't too many times I can be coaxed off Google.


 1:45 am on May 22, 2002 (gmt 0)

Someone asked me how they would search for just New Zealand sites, and I have to say I don't think Google does that in this corner of the world yet.


 1:46 am on May 22, 2002 (gmt 0)


Excellent. Search by Region. Yeah, I would like that too.


 1:56 am on May 22, 2002 (gmt 0)

OK, I'll bite...

I think the Similar Pages feature could really use some improvement. I find it only works for larger (higher PR) sites, often doesn't return similar sites, and doesn't display _enough_ results.

Google rocks.. I think it's great that you care enough to ask these questions.


(edited by: Marcia at 6:42 am (utc) on May 22, 2002)


 2:11 am on May 22, 2002 (gmt 0)

Sometimes it seems that Google gives too much relevancy to page rank - or perhaps it is judging page rank by the MAIN url instead of that directory.
Examples are student pages on University sites - it seems to be using the PR for the site as a whole, not those pages, resulting in some pretty dismal results at times. Another example is this:
On a search <snip>, this site comes up 12th in Google, yet it is an empty page with nothing but a title.
<url snipped>
There again, it seems to be using the main URL for PR comparison.

<No specific site references, please. ~Marcia>

(edited by: Marcia at 6:48 am (utc) on May 22, 2002)


 2:18 am on May 22, 2002 (gmt 0)

>>>You know, Google's not going to do well on that. I'll go somewhere else.

Google's great for searching for hard information but it's not the best engine to use when shopping for products (for me anyways).


 2:23 am on May 22, 2002 (gmt 0)

Google can return a thousand non-answers to simple almanac-type questions.
Usually, this shows up with factual "how much" or "how many" questions that can be answered in a single sentence.


How many bridges are there over the Mississippi in the Saint Louis, MO metropolitan area?

How many solar flares occurred in 1999?

How many volcanoes are in the Ring of Fire?

Google ignores prepositions and question-adverbs in searches. In the case of almanac-type searches, these words are critical limiters (time, place, relationship, etc.)

I can find answers, but can a child (say age 10-12) find the information he/she needs for a school report? The information is likely somewhere on the Web, and Google probably knows the right page, but it can be a challenge to find the page with the factual answer you need.

Often, a good almanac or call/visit to the local reference librarian produces quicker results.


 2:58 am on May 22, 2002 (gmt 0)


Yeah, that's it. Shopping / Commercial sites only search would be my #1 wish.


 3:15 am on May 22, 2002 (gmt 0)

Just a tiny bit off topic, but why SHOULD or COULD google be something to everyone? I dont shop on the Web, so never have tested out Google's capability here. I guess if i was shopping iid go to Overture or MSN.

Google's strength in finding "hard", "objective" or "scientific" data is helped by their alogos especially their PR algo. But their algos are not good for finding the "best" commercial or shopping sites, - because PR is a terrible way to "rank" these when the reasons for other people hyperlinking to sites, and therefore higher page rank is far less driven by "authority" or "relevance". PPC and Adowrds actually does a much better job at finding good commercial sites.


 3:19 am on May 22, 2002 (gmt 0)

GoogleGuy, try finding general information about
the island of Java... There used to be an island
between Bali and Sumatra, but Sun Microsystems
seems to have removed it from your index (DMCA complaint? :-). I went through the first 200 Google search results on "Java" a week ago, and not ONE of them was about that island.

It would help if the Open Directory had a category
for the island, of course, but even without that
it should be possible to separate the pages with "Java" on them into clusters and offer a choice between "Java (programming language)" and "Java (island)".

Even the Google Glossary lists nine definitions [labs1.google.com] - one for the coffee, eight for the programming language, and ZERO for the island.


(edited by: danny at 3:25 am (utc) on May 22, 2002)


 3:23 am on May 22, 2002 (gmt 0)

Key_Master, 4crests

I agree. Along with the Web, Images, Directory, and Groups there should be a Shopping section. Perhaps a paid inclusion system that e-commerce sites can submit to.

The sites listed there wouldn't interfere or blend in with the "Web Results", but users can click on a tab that would show shopping sites listed by relevancy. The relevancy would be determined by Google's various algos. Not just PR, but perhaps by various sorting methods: PageRank, Freshness, Popularity, and Editor-assigned topic/keyword categorization.

PageRank: This would allow people to view the sites that are deemed valuable by existing PR methods.

Freshness: This would allow people to view the sites that are updated frequently so that users can find sites that might offer better deals or special sales.

Popularity: This would allow people to view the sites that are visited the most by Google users.

Editor-assigned topic/keyword categorization: This would allow people to view sites for specific content that was found by the editors. (Would keep out the sites that people could artificially influence in the above factors.


 4:17 am on May 22, 2002 (gmt 0)

Hi Danny.. youve hit on an interesting point! There are several examples of tech terms that actually are "borrowed" from elsewhere completely outshining the original.

However Java is part of Indonesia, same as Victoria in Australia, surrey in england, and borneo (which includes parts of Indonesia, Malaysia, and all of Brunei). I would use "Java Indonesia" and you should get a much better result. Other possibles are "Java island" "Javanese" etc. So we all have to "help the SE's" a bit to find what we are looking for as it is not all magic!


 4:48 am on May 22, 2002 (gmt 0)

Translation. Google fails to do good Japanese translation into english. When I am at a Japanese webpage viewing it fine with all the neato charectors I select from my Google Toolbar the Translate into english then I see a nice page filled with ????

At that point I go to AltaVista and utilize Bablefish.


 5:53 am on May 22, 2002 (gmt 0)

Chiyo, you're right about it being possible to get around the Java roadblock, but that only works for people who a) know something about Java already (what happens if someone is trying to find out that Java is an Indonesian island?) and b) are reasonably adroit at using search engines...

Practice and a reasonably good understanding of Google's algorithm makes it easy for me (and probably most people in this forum) to think of good searches, but most people don't have that skill. And Google has always tried to make things easy for people.



 6:55 am on May 22, 2002 (gmt 0)

Try a search sometime for IT training [google.com] and then compare it to the same search at Fast [alltheweb.com] or Alta Vista. [altavista.com]

Fast and AV have no problem returning results from sites that offer information technology training. Google returns relevant ads, but the main results don't have anything to do with information technology.

It just doesn't seem right that I can't find an IT training course at the search engine most commonly used by the IT community.


 7:00 am on May 22, 2002 (gmt 0)

I agree with Danny's points - refined searches should be more intuitive for the casual user. How about a "Not" box for topics? e.g. Search "Java"
Not "Computer"
(If I use advanced search at the moment I can knock out a single word but not a topic area.)

Paid inclusion for shopping? One of the best commercial and PR advantages that Google has is its reputation for helping David against Goliath. Although that isn't strictly true, this would be a step which would damage that reputation.

I don't have a problem searching for items to buy on Google. Various people who want to sell me items might have a problem because they don't come as high as they might wish...

Robert Charlton

 7:18 am on May 22, 2002 (gmt 0)

GoogleGuy - This is one I've written to Google about, but the problem persists... Do a search on San Francisco public relations and note what comes up. On each of the first two pages, four or so of the sites returned have nothing to do with public relations... they're just very well-linked for "San Francisco" and happen to contain the phrase "public relations" somewhere on the page.

I've noticed this kind of thing a lot in competitive searches with "San Francisco" as part of the query... must happen for other cities too.


 7:21 am on May 22, 2002 (gmt 0)

what about


 7:30 am on May 22, 2002 (gmt 0)

;) Just kidding [google.com]


 7:36 am on May 22, 2002 (gmt 0)

1.Books and Publications [webmasterworld.com]
I still use scirus.com, or good old conventional library facilities for (scientific)Journal/Publications and Amazon for books. There must be copyright issues in indexing the headings or titles of publications, otherwise I cannot immagine why all those academics at Google would not have added this feature a lot earlier. There must be a lot of potential advertising for a seperate "book" tab (put in an option for out of publication/still available).

2. Fulltext indexing/searching of out-of-date copyright books. (search for a quote of Oscar Wilde, did Shakespeare ever use the word ....)

3. Search function for looking for similar images [webmasterworld.com]. (that is you copy a picture into the search field). Do not ask me what technical implications this would have.

4. Add the map(quest) function for foreign (non-us)addresses. BTW, why is this function disabled with other language Google versions?

5. Ixquick/Copernic for all sites that do not have a link from a site indexed by Google. (buy more servers with your IPO money, so what if they have a Pagerank of 0, you can still find these sites if you type in their company name with some kind of address/identification combination).

6. Donate the Google search funtion to Brett for internal site searches (the only time I go elswhere, to Google instead of WebmasterWorld ;).


 7:51 am on May 22, 2002 (gmt 0)

Id like to see Google showing all the backward links of a page when using the link command. Currently its better to use Alltheweb therefore.


 8:01 am on May 22, 2002 (gmt 0)

On each of the first two pages, four or so of the sites returned have nothing to do with public relations... they're just very well-linked for "San Francisco" and happen to contain the phrase "public relations" somewhere on the page.

I come across variations of that problem all the time.

I really hope the powers that be down at the Plex have taken the time to read about the concept of Topic Sensitive PageRank [www2002.org]:)


 8:11 am on May 22, 2002 (gmt 0)

I think you should get the number one slot sorted out for the search "googleguy"...it doesnt do you much credit. ;)


 8:18 am on May 22, 2002 (gmt 0)

Following the lead of Zips, I did a Google search for GoogleGuy and happened across his picture here:



 8:20 am on May 22, 2002 (gmt 0)

I'd like to see the books and publications search as well. Titles can't be copyrighted so that shouldn't be an issue. It's not the bestsellers that I want to see though, I want those big meaty tomes of info that people rarely buy and those obscure and esoteric publications. One of the reasons I want XML to take off is because BookXML and HistoryXML and TheologyXML would pretty much guarantee the content would be relevant.

3. Search function for looking for similar images.
Trying to beat the Gizmo Quiz? ;)

Then I'd like to see a notification system so I can send advertisers the SERP of their sometimes poor choice of keyword selection. If I do a search on Gap Theory does the merchant really think I'm interested in a new pair of jeans? Am I going to shop for a "theory" at Ebay?

The save search feature is a good one, how about one for emailing search results as well? :) Top it all off with a Descriptionary and use the OED instead of the rag-tag collection of dictionaries that dictionary.com uses and...

people will complain that the interface is too cluttered.

As for falling down, I think the searchers do that more often than Google does. A friend of mine hardcoded the word tees as a substitute for teas and tease at his golfing site.



 9:00 am on May 22, 2002 (gmt 0)

One thing that annoys me. I often encounter "dead" web sites with just the homepage left and maybe an outdated announcement for a relaunch, and these homepages are top ranked because they have the keyword in the title and hardly any text on the page. I guess it is not easy for Google to filter these "dead" sites out.


 9:25 am on May 22, 2002 (gmt 0)

I second Robert Charlton and Web Guerilla very strongly on the importance on getting those longer queries more relevant. One or two very popular words in a 3 or 4 word query and the other word(s) do not really seem to be given much weight. This is the key improvement i would like to suggest.

And I really look forward to the day, as WG says, when link relevance refines PR to a major extent. It will make for even more relevent SERPS and discourage this crazy PR "buying". I get sick of emails from people requesting reciprocal links... and most of the sites are unrelated in content as well.

This 87 message thread spans 3 pages: 87 ( [1] 2 3 > >
Global Options:
 top home search open messages active posts  

Home / Forums Index / Google / Google News Archive
rss feed

All trademarks and copyrights held by respective owners. Member comments are owned by the poster.
Home ¦ Free Tools ¦ Terms of Service ¦ Privacy Policy ¦ Report Problem ¦ About ¦ Library ¦ Newsletter
WebmasterWorld is a Developer Shed Community owned by Jim Boykin.
© Webmaster World 1996-2014 all rights reserved