Alex,
On the issue of validation, what about keeping a 'validation' database (so that the links database isn't changed) in which, the URL and the date_last_checked and valid variables are kept.
valid - is yes/no whether the link was validated on the date_last_checked.
That way, you could run a validate _only_ on links that failed, or links that haven't been validated in x-days, or since x-date.
It would speed up the process, and would allow separation of the validation routines from the regular links processes. The validate table could be updated either with each links modify, or at the time the validate routine is run (better option).
it reads through the links database, and makes sure each URL in the database is in the validate database, if not, it adds it. If a link ID and a URL disagree, the URL in the validate table is updated, and flagged as "unchecked" and set to be checked.
Make sense?
On the issue of validation, what about keeping a 'validation' database (so that the links database isn't changed) in which, the URL and the date_last_checked and valid variables are kept.
valid - is yes/no whether the link was validated on the date_last_checked.
That way, you could run a validate _only_ on links that failed, or links that haven't been validated in x-days, or since x-date.
It would speed up the process, and would allow separation of the validation routines from the regular links processes. The validate table could be updated either with each links modify, or at the time the validate routine is run (better option).
it reads through the links database, and makes sure each URL in the database is in the validate database, if not, it adds it. If a link ID and a URL disagree, the URL in the validate table is updated, and flagged as "unchecked" and set to be checked.
Make sense?
Real-time link validation