>The original product also uses SQL.
>
>I found the same problem as you with blogs... Most of the information is personal in nature, with only a small portion being useful. That is where I focused on improvement. The Better Log application I wrote allows one to select a category for the entries they make (which can later be searched upon). Also, the person making the entry can assign keywords to the entry for the purpose of searching. The traditional blog engine does not support searching... The Better Log application offers extensive search on a number of different criteria. I have also included a number of flags (source code included, examples included, etc...) that can be used on the search screen. More than anything I tried to make the Better Log more useful as far delivering information.

Yes, having a default category for each entry, when such criteria is used in a search, it also speeds the search as a filter is then applicable.

I also find those blogs to have those entries to be lost in space after a while. There seems to be a lot of rotation that occurs, such as a monthly summarization, and sometimes you might thing there is nothing in there where the information is hidden.

Your approach has given me a lot of ideas. It also provided me a more solid experience on why this should be avoided. I don't see much flexibility to make a data centric evolution out of this. We will try to see what advantage we could gain by having that in the project and we will see about alternatives.