From: gopher-bounce@complete.org Date: Mon Aug 4 04:56:49 2008 Subject: [gopher] Re: Item Type Suggestions Yes, lets focus on taking advantage of what Gopher already has, before we worry about extending things. On the Server Side, note that Windows still lacks any kind of Gopher Spider/Crawler, GopherMail Server (i.e., a way to return Gopher pages via email), or Gopher Document Search Engine. These would make excellent Gopher projects for Windows programmers. --- On Sun, 8/3/08, Mate Nagy <k-zed@hactar.net> wrote: From: Mate Nagy <k-zed@hactar.net> Subject: [gopher] Re: Item Type Suggestions To: gopher@complete.org Date: Sunday, August 3, 2008, 3:23 PM Hulloh, > A possible workaround: get the parent menu (using the same method I > mentioned before - I don't remember anything stating the selectors are > supposed to be built with a slash between each directory name, I need to > read the rfc again (even if gopher is supposed to be hierarchical, what > prevents some server from using dots instead of slashes for directory > separators?)), check the mime-type (for the mime-in-the-menu approach), > or check if the entry has gopher+ data (then retrieve such data). Exactly - you can't rely on '/' being the separator, or that the last section cut off yields the valid parent dir, or even that there are directories. At this point I'd much rather drop this protocol extension thing, and focus on using the existing technology to the fullest.. whatever that means :) ... For that matter, I'd love a client for Windows Mobile 6.. also for the iphone 3G :) On the server side: full text search engine. I have some experience with using the HyperEstraier engine. I have about 16 gigabytes of text indexed; I can search and view results over Gopher (with a python "mole"), it's pretty fast and convenient. It also supports boolean operators and phrase search. It doesn't have a Gopher spider currently per se, but it's fully open source; maybe we could fit a spider to it? (Or even get the existing Veronica spider to talk to it?) (Setting up and initializing the search db (from local data) is very easy, so the required effort to play with it is minimal.) Regards, Mate