[linux-elitists] A modest proposal to web designers of the world (was Re: Site feedback: fonts too small)

Don Marti dmarti@zgp.org
Thu Jun 26 10:28:33 PDT 2003


begin Rick Moen quotation of Wed, May 16, 2001 at 02:52:25PM -0700:

> The other item on my feature wish-list, however, is more ambitious:
> I'd like to be able to get access to an HTTP proxy's configuration
> in real-time from within a rendered page.  I want to be able to
> right-click on a rendered item, and specify either (1) proxy to oblivion
> all objects from this IP, (2) proxy to oblivion all objects from this
> FQDN, (3) proxy to oblivion all objects from a subdomain I specify,
> (4) give me the ability to specify some regex substitution right now,
> and implement it immediately, and (5) list all rules currently in effect
> and let me inspect/edit any of them.
> 
> It's unreasonable to have to build a general-purpose HTTP proxy into
> every Web browser, so what we should be looking for is a standard
> software interface browsers can talk to, and that proxies can take
> direction from.  I.e., a SWPI = Standard Web Proxy Interface. 

This may be the best way to implement "turning off filters" for
US libraries, which are now required to run web filters if they
receive Federal funds, but also required to turn off the filter if
a user 18 years old or older requests it.

It would also help build block lists for censorware
that accurately reflects community standards.
http://linuxjournal.com/article.php?sid=6807

The only configuration item needed for this would be one "SWPI URL"
line, which could be set to something like:  http://proxy/swpi
by default.

An SWPI request could simply be an HTTP GET containing exactly one
pattern and one command, and.   HTTP authentication can be used.

Patterns:
  URL:   a literal URL
  GLOB:  a shell-style pattern with * and ?
  REGEX: a regular expression

Commands:
  block:    request a block of any resource matching the identifier
  pass:     request the proxy to allow any resource matching the identifier to pass
  delete:   delete the rule
  alias:    request the proxy to replace any resource matching the
            identifier with a different resource
  inspect:  request an informational page

"New" field:
  If the command is "alias", the SWPI request must include a
  "new" field.  This is interpreted the same way as the pattern,
  as either a URL, a shell pattern, or a regex.

Valid SWPI response codes:

200 OK
201 Created: A new rule has been created
400 Bad Request: syntax error
401 Unauthorized: Give the user a chance to authenticate 
403 Forbidden: 

The actual content of SWPI "inspect" pages can be designed by the
individual proxy implementor.

Presumably

http://proxy.example.com/swpi.cgi?regex=.*&command=inspect

could give you a page with a number of links such as:

http://proxy.example.com/swpi.cgi?regex=www.%29%3Fa*&command=inspect

to let you drill down to rules for sites starting with "a".

-- 
Don Marti                Reform copyright law -- return abandoned works
http://zgp.org/~dmarti   to the public domain after 50 years:
dmarti@zgp.org           http://www.PetitionOnline.com/eldred/petition.html
KG6INA



More information about the linux-elitists mailing list