All of the following has not been implemented. It's just my design goals.
-- CP

HTML, XML, CSS and ECMASCRIPT PARSERS
=====================================

The idea with the parsers is to implement (at least) one for each version
of each  DTD or language version, and implement them as dynamically loadable 
modules ('plugins'). The user can then select which parser(s) should be used in 
the parsing of Web pages, or can have the browser choose at parse-time (using 
cues provided in the document). This way, the browser is extensible and can shed unneeded memory (i.e., no need to load support for ECMAScript if the user 
doesn't want it).

Ultimately, I'd like for developers to be able to write parsers in any of 
several languages. I intend to write one for strict HTML 4.0 and one for CSS 1 
in Perl. I do not intend to touch ECMAScript/JavaScript, because I think it 
sucks. Someone else can do that. :)


CACHEING/OFFLINE BROWSING
=========================

Every document a user downloads will be written to a file 
(~/.gzilla/cache/<domain name>/<file path>). (Maybe it should be gzipped for 
space saving?) Furthermore, the output of the parser will be written to another 
file (~/.gzilla/cache/<domain name>/.<file path>.out) to speed up display of 
pages the user has already seen.

Downloading an entire site or part of a site will be implemented as a special 
case of the integrated command shell (described below).


INTEGRATED COMMAND SHELL
========================

The 'address bar', the area where users type in URLs to be downloaded and 
displayed, will actually be a miniature command shell, with 'gzilla' implicit at the beginning. From here, users will have the ability to invoke any of Gzilla's 
command line options (described below), as well as access to external programs 
via piping and redirection. Examples:

	http://www.example.net/somefile.html | grep blarg > ~/blarg-results
	-dump http://www.example.net/somefile.html > ~/example-dump
	ftp://metalab.unc.edu/pub/Linux/kernel/patch.gz | gunzip - > \
		/usr/src/linux/patch
	http://www.example.net/somefile.html --no-css --use-<parser>
	...

Thus all the Unix goodies become available from within Gzilla, and you can do a 
lot of work without ever switching to a terminal window or mousing around for 
menu choices (e.g., redirection to a file is essentially 'Save As...'). Without 
file redirection, STDOUT is the browser window. In all cases, files are cached 
as described above.

The wildcard characters expand as normal when in the context of regular text, 
but file output(s) is/are written to the proper cache file and not displayed on 
STDOUT. Thus

http://www.example.net/*.*htm*

Writes all the files ending in .*htm* to the cache (e.g., for later offline 
browsing).

However, in the case of a lone wildcard, as in

http://www.example.net/*
http://www.example.net/section2/*

the entire site below the directory the * is in will be downloaded (the entire 
site, and all of section 2, respectively). Note that this is not the same 
behavior exhibited by the standard Unix shells.


COMMAND LINE OPTIONS
====================

Where command line options conflict with default or ser-defined preferences, 
they take precedence.

-h, --help
Display command line options.

-d, --dump
As in Lynx.

--no-css, --use-css
Do (not) use Cascading Style Sheets, if present in the document.

--no-ECMAScript, --use-ECMAScript
Do (not) use ECMAScript/JavaScript, if present in the document.

--use-<parser>
Parse the document with the given parser (overriding document's cues). This could give weird results. :)

-n, --new-window[name]
Open URL in a new window (with optional [name]).

--open-in<name>
Open URL in named window. If <name> does not exist, it is created.


GRAPHICAL INTERFACE (MENUS, BUTTONS)
====================================

MENUS
	Gzilla
		New Window		CTRL+N
		Open...			CTRL+O
		Preferences...	CTRL+F
		Close Window	CTRL+W
		---
		Exit Gzilla		CTRL+X
	Document
		Save (As)...	CTRL+S
		View Source		CTRL+V
		Copy Text		CTRL+C
		Select All		CTRL+A
		Find Text...	CTRL+F
		Print			CTRL+P
	Browse
		Back			ALT+LEFT
		Forward			ATL+RIGHT	
		Reload			CTRL+R
		Stop Loading	ESC
		Home			CTRL+H
		Bookmarks		CTRL+B
		---
		[linear history]
	Bookmarks
		Add Bookmark	CTRL+A
		---
		[bookmarks]
	Help
		Gzilla Home
		Gzilla Manual	# file:/path/to/manpage | groff -man :)

BUTTONS
	Back one step/back in history
	Forward one step/forward in history
	Reload
	Stop Loading
	Home
	Bookmarks






Raph's old stuff
* Better bookmark support, and improve bookmark file format and
parsing. [CP: Maybe use Randy's bm2html, or the like?]

* make autoconf find gtk libraries.

* implement ftp methods for bytesink