[ih] "How Gopher Nearly Won the Internet" Re: The Rise and Fall of the Gopher Protocol

[John Levine]

>But the most important ability was the CGI (IIRC that's what it was 
>called), the API and protocol which allowed a "document" to be retrieved 
>by calling some back-end program in the server, and even supplying 
>arguments to the call.

It's still called CGI -- my Apache server runs CGI scripts several
times a minute.

>Of course the "forms" interface also meant ...
>As far as I remember, there was no such capability with Gopher or WAIS, 
>or maybe I just hadn't found it there.

Gopher got about halfway there.  A gopher server returned a menu of
items, of which there were three types, menus, documents, and search.
A search item let you send along some search terms, and the server
constructed a menu on the fly to respond to it.  As far as I can tell,
all anyone ever did with it was to search lists of FTP sites or local
phone books or other gopher servers, but there's nothing in the
protocol that would prevent a general retrieval service beyond the
limitation that the result had to look like a menu.

The Gopher RFC talks about Gopher-to-FTP or Gopher-to-WAIS gateways,
but I never heard of anyone actually building them.  The gopher
protocol was designed to be extended (menus were type 0, documents
type 1, searches, type 7) but it wasn't.

R's,
John