🚀 go-pugleaf

In article <JONATHAN.92Dec11220406@speedy.cs.pitt.edu> jonathan@cs.pitt.edu (Jonathan Eunice) writes:
>Every time I use archie, I am amazed at (i) how very useful it is, and
>(ii) how geographically stupid it is.  It generally makese sense to
>use archives that are net-wise close to oneself.  Yet archie gives me
>many site-references in Europe, Japan, Australia, etc in addition to
>US sites.  And I'm sure my friends in Sydney and Munich don't care to
>use US sites any more than I want to use mine.
>
>Is there any way to limit archie's search only to "close" sites?  If
>it's in TFM, I haven't found it.  I can hack a script to filter out
>distant sites, but it seems to me archie would work faster, handle
>more requests, and be more useful were there a built-in 'give me only
>local sites' option.  This could be simply based on country-codes and
>still be right most of the time.  Thoughts?

Help is at hand (almost). The new release of the server
code (now in the hands of its licensees, although not yet
in production in most places) includes the ability to
specify both domain information on searches (for example
"prog pattern *.edu) which would return only examples of
"pattern" from hosts in the domains ending in .edu. In
fact, the operator can also define "pseudo-domains" such
as "europe", or "northamerica", which could be any set of
domains chosen by the operator. These pseudo-domains can
then be used in searches as well.

The new release also allows you to specify directory path
information to further limit searches and lots of other
new stuff, as well (such as a "shortform output format"
which returns less verbose results, allowing hopefully
more rapid scanning of results, since the system so often
generates _lots_ of excess hits.

Unfortunately, the new release of the Prospero protocol
(upon which all this new functionality depends) is not yet
frozen and the clients do not yet support it but it _is_
already supported in the server once the client authors
start upgrading. We'd hoped to have the protocol frozen and
clients usable RSN, although both Prospero development and
client development are to a large part out of our hands.

For what it's worth the new version is a completely
generalized data-gathering engine capable of gathering
files (not just filenames) and we're hoping to find
sponsors for additional interesting collections of
information in the coming months. Contact me for more
details.


				- peterd