I would like to know 2 questions. If I am searching on this site, or
the largest subject directory, and I want to ensure that whatever I am
looking for brings back results NOT LISTED IN THE UNITED STATES then
what do I need to do. In each search I use 2 words as well as my
desire to not retrieve U.S. results. My other question is how can I
ensure that when I do a search using 2 words that I can retrieve
results for A SPECIFIC COUNTRY.

Thank You,
Joel

It's impossible to search for specific countries on the Internet.  The
problem is that websites are hosted on "servers".  These "servers" are
located all over the world.  A webmaster in Africa may very well use a
server in the United States, for example.  You may be able to tell
where the server is located, but this is not always an accurate
indicator of where the website is located from.  A lot of websites
merely contain information and they're not country specific.  There
are, however, three methods that can be used for semi-accurate
country-specific searches.  They are not real accurate, but they're
all the options you have.



Method #1 - Use a Country-Level Domain Search

Many countries have their own domain name.  France has ".fr", Canada
has ".ca" and the United Kingdom has ".uk".  An entire list of these
country or "top level" domains can be found at the ICANN website:
http://www.iana.org/cctld/cctld-whois.htm

Most search engines allow you to refine your search to a specific
country code.  While this allows you to only search for a certain
country, it really is not a very good search tool because almost
everyone uses the .com domain extension.  A lot of French, Canadian
and English websites use .com.  It's more than likely a majority, but
since I haven't seen specific data backing up my belief, I will leave
it at "a lot".  By requesting only ".fr" websites, you're really
missing a lot of French websites that use the .com extension.

Almost American websites use the ".com" extension.  You can specify to
search only non-dotcom websites, insuring that you will mainly recieve
foreign websites, but you must understand that you will be missing
lots of foreign websites by doing that.

I will give you instructions on how to do this and it's the only real
way to search by country per your requests, I'm just warning you that
it's not a real advisable way to go about it because of the amount of
data you'll miss.

Google Specific Country Search:
Search Google using the "site:" extension.  For example, search for
"web searches site:.fr".
://www.google.com/search?&q=web+searches+site%3A.fr

Google Remove US Search:
Search Google using the search "web searches -site:.com" (without
quotes).  This will remove any .com websites from your search.  You
will probably also want to remove any .net and .org searches.  To do
this, search for "web searches -site:.com -site:.net -site:.org". 
This can be extended to .us (the official but not yet widely used US
extension) and .edu as well.
://www.google.com/search?&q=web+searches+-site%3A.com+-site%3A.net+-site%3A.org+-site%3A.us+-site%3A.edu



Method #2 - Use a Specific Language Search

Many search engines allow you to search for a specific language.  This
is probably the best indicator of a specific country.

Google Specific Language Search:
://www.google.com/advanced_search?hl=en
Look at the first option in the white box underneath the light blue
box.

Altaista Specfiic Language Search:
http://www.altavista.com/sites/search/webadv
First option in the white box underneath the yellow box.



Method #3 - Search a Country-Specific Directory
While this is the most "accurate" means of restricting your results to
a specific country, I would recommend the language search first if
possible.  Search engines have far more websites than any directory
could ever have.  Examples of country specific directories are:

German Portal:
http://www.web.de

UK Pay-Per-Click Engine:
http://www.espotting.com

Yahoo Spain Directory:
http://dir.yahoo.com/Regional/Countries/Spain/  (take off the "/spain"
to see all countries listed in the Yahoo directory)



Thanks for the question and if you require more information, don't
hesitate to ask for clarification.  And if you find this answer
satisfactory, please feel free to rate it. Thank you!
      
xemion-ga

---

Request for Answer Clarification by joel1357-ga on 14 Jun 2002 01:25 PDT

I'm getting a little tired so I'll wait to rate (thanks for the
responses though) this but the approach I have been trying tonight has
been slight variations of this..."secondhand widgets" -com -net -org
-gov -edu -uk -ca +mx

This has worked pretty well but both this site and the largest subject
directory only allowed me to use 10 words. I would like to use 2-3
times that many.

Thanks,
Joel

---

Clarification of Answer by xemion-ga on 14 Jun 2002 09:09 PDT

I'm confused...I thought you were trying to use TWO search terms.  As
for as I know, there's no way to search for more than 10 terms on
Google.

xemion-ga

A site that may interest you lists several search engines reported to
be in a certain country can be found here :

http://www.searchenginecolossus.com/

This is only a list of other search engines that try to keep their
listings country specific. As stated in the answer, there is really no
way, at the search engine level, to filter out all answers not in a
specified country, unless you check each one. Programmically you could
check for an address of origin or the whois listing of a website, and
thus get a cleaner responce.

Hope this helps you out. 

webadept-ga

A couple of directories which may be useful under Methods #2 & #3,
both compiled by the Open Directory Project (dmoz.org).

Open Directory Project

World Directory (in 71 languages)
http://dmoz.org/World/

Regional Directories
http://dmoz.org/Regional/

~L

An excellent comment by webadept (as always).  I would highly
recommend the SearhcEngineColussus website.  I wish I had included it
in my answer, but alas, I didn't know about it until just now.

He is also correct on a search according to the whois info.  I had
considered that possibility, but did not include it in my answer
because I knew of no large search engine that had that feature and
even if it they did, it would not highly reliable.

xemion-ga

Sometimes, a very simple technique can work well.  For example, if I'm
trying to locate UK companies from a particular sector, I put in
search terms describing the sector and also the term "UK".  That
avoids the problem of missing out on UK companies with a .com rather
than a .co.uk domain.  It's not perfect, but it does help to reduce
irrelevant results to some extent.

In the same way, you could use - (minus) to exclude reference to a
country e.g. NOT USA.  Obviously, this runs the risk of losing a lot
of pages from non-US where reference is made to the USA. It really
depends on how exhaustive you want your search results to be.

AltaVista advanced search offers (a) search by domain and country
(with the possibility to open up a listing of these for reference);
(b) search by region, eg North America, Asia, Middle East etc; (c)
search by language.  It also offers the possibility to use a Boolean
search expression, which will allow you to form very precise
questions.

I don't think this is all that hard in general you can search lik
this:

ATG Dynamo -site:.com -site:.org


Will bring up all the sites which don't end in .com or .org,
realistically that get's you most things which are not in the US

ATG Dynamo site:.de

Will bring up only German sites

ATG Dynamo site:.fr

Only french

Hope that helps
Stuart

Boolean logic (which is the basis to all search engines and databases)
has a NOT command - which is implemented differently on different
databases.

The Google variation of this is its 'Without' field which you use in
the Google Advanced Search.

Good luck! (and dont forget to use the " " around multiple word search
terms)

I think Joel was referring to not searching for something with more
than 2 words, but rather limiting by more than the 8 countries that
will fit within the 10 word limit.

if you're trying to do a search like...

"web search -us -ca -mx -fr -de -bz -fd -gdf -sdf -sdf -er -gs -wer
-sd -df -fe -se -gw -hy -yj -as -wer -hy -sd" (I'm making up country
codes here)

...it's not going to work.  You can only search for 10 terms total.  I
don't think there's anyway around it.  10 words actually a pretty big
limit.  For example, http://www.wisenut.com, a Google competitor, only
allows seven words.  I'm sorry.

xemion-ga

I think one method would be to search for your terms, and include a
requirement for the country you're searching for or it's abbreviation.
I think most companies would have some mention on their site where
they are located, or their address.
example  from google:
used widgets mexico OR MX
using the advanced search box, I entered used widgets in the All of
these box
and mexico mx in the at least one of box
which yielded that search term. You could add more terms to the at
least one of box, for example other abbreviations for Mexico.

It actually is would be very easy to find out what server you are on.
Go to www.download.com and download VisualRoute and then when you go
to a website and you want to make sure you are not on a foreign server
use this tool to trace where in the world the orgin is. You can also
use search specific method, for example: I want to find information on
a dog on an server located in the U.S.
I would type: +"dogs" +"us server"               that may do it so go
ahead and try

Hi,
you can find all what you need to know on how to use search engines on
the following Very interesting link --> http://www.searchlores.org
-- Eye --.

My answer is probably familiar to some of you. It's also messy,
recursive, and doesn't work all the time -- but nevertheless I thought
I would go ahead and share it since it wasn't mentioned already.

Inspect your target web site and extract a unique text string that is
both unique to the country and commonly used accross the country.

For example, let's say you wanted to find web sites located in France.
You could locate a web site in Toulouse France and another site in
Paris France. From those two web sites, you could extract their phone
numbers "tel.: 01 53 35 05 40" and "Tel : 01 45 03 21 92" and then you
could infer what they have in common "$el$:$01$" and try it out on
Alta Vista.com. Again, this approach is far from perfect, my search
didn't catch all the sites in France and for some reason it also
managed to grab a couple of web sites from Italy, but in any case you
get the idea. This technique can also work for other strings like zip
code/addresses, foreign charatères, and probably a bunch of other
examples I don't know about.

Steph
PS: Another potential solution would be using the Google API. For a
fee, Google allows you to programmatically access its database, so it
should be possible to refine the search further programmatically.