On 3/17/2000 at 5:47 PM, l(_dot_)wood(_at_)eim(_dot_)surrey(_dot_)ac(_dot_)uk
(Lloyd Wood) wrote:
I believe it's worth changing the default to conform to this
behaviour, since it offers search advantages. Generally, you wouldn't
search your own archive from e.g. altavista.
Does it really offer significant advantages? Per the example you gave:
+url:www-nrg.ee.lbl.gov/ -url:thrd +jain +juha
Even if someone knew what they were looking for to the point of using
this syntax, a successful match would just turn up a threaded index
containing those two terms. It doesn't indicate that two people were
talking with each other, just that their names appeared in the same list
within the same list-specific unit of time. And even where they did
contribute to the same thread, finding it on a large index can still be
awkward.
Meanwhile, changing MHonArc's default file-naming conventions would
break every inbound link the first time an archive admin rebuilt their
collection after upgrading.
However, while I'm on the subject, I've got a widget in development that
stores MHonArc message attributes in a SQL database (currently only
MySQL, but I intend to open it to all databases supported by Perl DBI).
Support for thread recognition is a high priority.
It won't have an impact on altavista or the other general search
engines, but it should allow for fine-grained search implementations by
archive admins.
-nat
