What are the diferences in this 2 scripts? What should I use?

Thanks...

Search.cgi supports the altavista style searching of +word -word +"some phrase" -"some phrase" "some phrase". It uses the Links SQL index tables, so is very fast. The limitation is that if you are using Big5, or GB encoding, or some other encoding where words can not be split up by spaces, the indexer does not work.

In this case you should use search-ni.cgi (the ni is for non indexed). This will just do a regular search on the database.

Cheers,

Alex

Ok, but, what is GB and Big5 enconding? I use characters like ã, é í, á... Am I able to use the indexed search?

Big5 or GB encoding is used to display some Asian languages. The problem with it is that words are not separated by spaces. So if you have 'aaabbbccc', there is no easy way to tell what the three words are from the string.

If you just have some accents, but your words are still separated by spaces, then the indexer will work fine.

Cheers,

Alex

Thanks for the quick response.... and for the info of course...

I'm having some difficulties with the search results. I have added a weight to the URL field to allow it to be searched on and reindexed the table. (I had to reindex through telnet and use the CTRL-D method outlined in gossamer-threads.com/scripts/forum/resources/Forum9/HTML/000527.html , all my paths appear to be set just fine...?)

Anyway, if you go to run-down.com/cgi-bin/runlinks/page.cgi?d=1 you can follow this example. If you do a search for http://www.bsim.org/ that link is returned in the results. However, if you search for bsim.org no results are returned. Why would this be?

At first I thought it might be related to the "space between words" problem, based on another example title/URL I was trying, but it doesn't appear to be that...

Dan

[This message has been edited by Dan Kaplan (edited February 16, 2000).]

Hi Dan,

This is because of how the indexer breaks up the word. If you have the word 'computer' in your database, and you search for 'puter' it won't match.

The indexer inserted 'www.bsim.org' as a word, so you need to search for 'www.bsim.org' to make a match. This is how most search engines work (in that you have to match the beginning words).

Does this make sense?

Cheers,

Alex

Hi Alex,

Thanks for the reply. That makes sense, and indeed you are correct. Searching for www.bsim.org works. However, Yahoo searches (or HotBot/Intomi or whatever you want to call it; I haven't tried it on any others) do not work that way. Leaving off the 'www' does not make a difference. Have they added additional criteria to allow for that?

The examples I used with Yahoo were:
www.dankaplan.com
dankaplan.com
kaplan.com

My link (old URL) shows up in the first two searches, but not the third.

Dan