gpmgroup - 7:50 pm on Jul 12, 2007 (gmt 0)

We currently handle well over a 100,000 non spam emails a year.

With hundreds of emails arriving every day our email database needs a good directory structure to stop it rapidly descending into an unusable quagmire.

(Unlike DMOZ we do not have the luxury of being able to ignore/queue these requests and therefore need to keep on top of things by dealing with them in a timely manner)

And yes like DMOZ we get far higher quantities of spam to also deal with too. But unlike DMOZ this is not allowed to pollute the areas where the work is done from.

The directory structure has grown over the last 10 years and like dmoz it's many layers deep. For the last few years we have had specialist search which enables the revrieval by any word(s) in any email body or header that returns matching emails usually within 5/100ths of a second or less.

All the users also have favourites pointing to directory paths they work on a on a daily basis.

Of the three ways to access the data; directory traverse is now used less and less. Search simply for speed and accuracy has replaced the directory traverse for data retrieval.

For day to day working favourites (directory path shortcuts) enables user to target groups of data.

On large data sets other than adding, editing, removing data a directory traverse is for the vast majority of tasks an extremely inefficient use of time.