Re: Search Help Features within Long Threads and many Posts/Replies
You are 100%correct in that it is all about server load. I've dealt with this at the programing level on a single site basis. The content search text used is applied to a field (user entered post) that is relatively large, contains text combined punctuation. The backend routine has parse out special characters such as dashes, commas, semi-colons, dollar signs, slashes, etc. or you would be less likely to find content.
There are also "illegal" characters that would conflict with the SQL command language or that could allow execution of undesirable programming code fields.
Short strings can be searched but there is more to it than just that. There are different kinds of searches. Do you want to find a string (word) that begins with the 3 characters, contains the 3 characters, etc.
The begins with search is relatively straight forward; you search every post for those 3 characters appearing after a space and stop. Those results are moved to an index, and then you search the next post, and so on.
The contains means means you have to search every 3-characters in the entire post. which is a lot CPU and disk IO. The service hosting the site likely has many different databases from different organizations with 10 of thousand potential users. If a single server is being use that is a big load. So it's also a matter of hardware, communications bandwidth, etc.
Finally, if you want to really screw up a site, you can load it up with searches that involve very few characters and automate that process to repeat the inquiries by generating random characters. You'll end up with a denial of service at some point. This is the main reason the searches require a minimum number of characters.
You have to live with the results as well. You will get exponentially for results each time you reduce the minimum number of characters so the search will likely be less useful. Some sites offer the ability to do an initial search, then a "Search on Results" which is very helpful. The NCRS site is very well structure with a lot of good search / filter options.
I did post an answer to the 3-character search limit in case it was missed. you can simply type in your 3-characters followed by an asterisk and you will get your results. Search for L88 is entered as L88* Once you get the page results, by pressing Control-F or Command-F on a MAC, you will get a page search and you can find additional content but you would be limited to what appears on the page you are viewing.
Yes it can be done, but I would be surprised if the service provider would allow it.
You are 100%correct in that it is all about server load. I've dealt with this at the programing level on a single site basis. The content search text used is applied to a field (user entered post) that is relatively large, contains text combined punctuation. The backend routine has parse out special characters such as dashes, commas, semi-colons, dollar signs, slashes, etc. or you would be less likely to find content.
There are also "illegal" characters that would conflict with the SQL command language or that could allow execution of undesirable programming code fields.
Short strings can be searched but there is more to it than just that. There are different kinds of searches. Do you want to find a string (word) that begins with the 3 characters, contains the 3 characters, etc.
The begins with search is relatively straight forward; you search every post for those 3 characters appearing after a space and stop. Those results are moved to an index, and then you search the next post, and so on.
The contains means means you have to search every 3-characters in the entire post. which is a lot CPU and disk IO. The service hosting the site likely has many different databases from different organizations with 10 of thousand potential users. If a single server is being use that is a big load. So it's also a matter of hardware, communications bandwidth, etc.
Finally, if you want to really screw up a site, you can load it up with searches that involve very few characters and automate that process to repeat the inquiries by generating random characters. You'll end up with a denial of service at some point. This is the main reason the searches require a minimum number of characters.
You have to live with the results as well. You will get exponentially for results each time you reduce the minimum number of characters so the search will likely be less useful. Some sites offer the ability to do an initial search, then a "Search on Results" which is very helpful. The NCRS site is very well structure with a lot of good search / filter options.
I did post an answer to the 3-character search limit in case it was missed. you can simply type in your 3-characters followed by an asterisk and you will get your results. Search for L88 is entered as L88* Once you get the page results, by pressing Control-F or Command-F on a MAC, you will get a page search and you can find additional content but you would be limited to what appears on the page you are viewing.
Yes it can be done, but I would be surprised if the service provider would allow it.
Comment