--On Thursday, November 11, 2004 5:50 PM +0100 Danny Ayers <[EMAIL PROTECTED]> wrote: > > Empty content and summary could be used to signify that an entry was > no more.
Let's not do that, even though it would sorta work. In our HTTP spidering, we do see documents with no content, but we have different log messages for "200, no title or summary" and "404". Documents like this would be binaries served as text/plain or PDFs which are scanned images with no metadata. Quite common, really, and entries representing those would be empty, too. We are sure to see software problems where entries are generated with no content, and I'd really like to flag those as what they are instead of mistakenly deleting lots of stuff. Should we add a requirement that there must be content in at least one of the entry fields? I think that would be a good idea. Otherwise, clients are required to store null entries. wunder -- Walter Underwood Principal Architect, Verity
