On 20/12/06, Steven C. Barr(x) wrote:
>> Is there a medium ground then?
> Well, since you're designing (and as such, defining) the project...
> that's YOUR call! It would be possible (if you have a good typing
> finger, anyway...?) to create, or try to, the "ultimate" database...
> from which, by using well-designed queries, you could then take
> the necessary data to serve as an "(Almost) Complete Classical
> <format(s)?> Record Dating Guide...
> In any data-related project...be it database creation, cataloguing,
> useful HTML page(s), usw...the data entry is always the "Killer"
> task. It's slow, tedious, and very hard to automate to any useful
> extent...and worse yet all but demands a fairly complete knowledge
> of the material involved, if only to avoid the sort of screamingly
> obvious errors (using titles for composer names or vice versa,
> entering data that couldn't possibly exist, and so on...) that
> are immediate tipoffs the work was done by folks who had no idea
> WHAT they were data-entering!
As can be seen any time you try to find a classical CD on Amazon.
I think it is best to set up as many fields as you can, even if for most
recordings the majority will be "not known". For example, the
microphones used are stated for the Mercury CDs and for some Decca
"Legends" CDs. Those should be included, even if only 1% of recordings
have the data available.
[log in to unmask]