The Hypertext Usenet Reader & Linker
HURL provides a searchable Web interface to Usenet news or
mailing list archives. For a sample archive system built with
HURL, see the HTML
Writers Guild mailing list archives or the fogo mailing list archives.
I am rewriting HURL from scratch in Perl, and will make it available
under the GNU GPL as soon as a first version is working. You really
don't want the old code, trust me; the new version is
infinitely better than the old one. (here's a sample of the new
version, still in development, but getting there.
See also: rough design notes.)
If you would like to be notified when the new version is
ready, please send me a short message
indicating your interest.
It should work on most Unixes and possibly also Windows NT or
something else, but I won't be able to test it on non-Unix
platforms.
- 18 Feb 2001
-
- added support for searching against arbitrary headers
(the list of headers is defined in the config file for each
installation)
- added a max-age to each generated page, to make the pages
more cacheable
- 12 Jan 2001
-
I made some good progress today:
- added support for mbox
format archives; until now it could only deal with MH-format
archives, i.e. one message per file
- improved the date parsing a bit: no longer relies on my
particular archive storage format, now it gets dates from
envelope sender headers
- added support for multiple archives in one installation,
so it's easy to maintain dozens/hundreds of archives on a single
system
Most of the other material on this page (and linked from it)
dates back to 1994 and hasn't been updated. The text above
this paragraph is the only part that is up to date.
(last updated $Date: 2002/07/25 04:06:23 $.)
Implemented Features
Query page (sample)
- Header searches:
Text matching against any header element (as defined by the
administrator)
- Searching with errors:
Text matching with anywhere from 0 to 8 errors can be specified,
thanks to agrep.
- Case selection:
Searches can be case sensitive, or case-insensitive.
- Exact / regexp matching:
Searches can match the text exactly as entered, or using perl
regular expressions.
Message list browser (sample)
- Allows for browsing of any number of messages in a list.
Article pages (sample)
- Browsing icons:
- Icons with links to the next or previous article by date, author,
or in the currently selected list of articles.
- Icons with links to the home page, query page, message list,
article filters, and help.
- Unavailable icons are ``dimmed'' and not linked, to prevent
user frustration.
- Intelligent header formatting:
- If the article was crossposted, the ``Newsgroups'' line is shown,
with links to the newsgroups themselves.
- The ``Subject'' line gets a link to a query for that subject
(rudimentary ``threading'').
- The ``From'' line gets a link to an ``author page'' for that author.
- Each article in the ``References'' line gets a link, only
if the article is present in the archive.
- If the ``Followup-To'' line is different than the ``Newsgroups''
line, it is shown with links to the newsgroups themselves.
- Links within articles:
- Article references and e-mail addresses are linked if the article
or author exists in the archive.
- URLs are automatically recognized and linked.
- Article filters:
- Any article can be reloaded with each word linked to various
dictionary gateways elsewhere on the Web.
Author pages
- Each author with an article in the archive has an ``author page''
with statistics about their posting history, links to a list of
their articles, to send them mail, or to finger them.
Coming Soon
- Full-text searching
(thanks to glimpse).
- Search caching (easy to do, and a big performance win).
- Better docs, both for the interface and the installation.
- Support for mailing lists. (done)
- Better support for text-mode users. (done)
- Much more; see the exhaustive feature list.
If you're a spammer, you're probably more interested in this infinite list
of e-mail addresses.
Last modified: $Date: 2002/07/25 04:06:23 $
Gerald Oskoboiny, <gerald@impressive.net>