RE: Any interest in automatic link-updating?

Should I just invent a general purpose protocol for "telling each other
that resources aren't working"

I'll call it Internet ADVISE protocol......and have it use port 666.

The "database" used at the both ends of the protocol will be a
formalized subset of HTML 1.0.

The client can read the HTML file, send mail for each entry in the file,
or if the bad link was the result of a spider, indexing or other
automated utility....it  can use the results to refresh its internal
database.

Right now there are hundreds of different forms to fill out for spiders
to index your site, ten additional "Meta" tags, and no way to tell them
something has changed except to request the wrath of re-indexing your
entire site.

The protocol will is simple and straight forward....maintainable by
hand...and build on top of  existing standards.

There is no attempt to duplicate work done by the HTTP protocol itself.

A. Protocol

1. A text/html file will be POST'ed to the offending server at location 
"http://advise.list" OR "https://advise.list" containing a single 
"application/x-form-content-encoding" submission

2. If this POST is Redirected to a mailto: URL, the advising agent will
use Internet Mail ... to mail the advise file to the new location.

3. If this POST is Redirected to an ftp: URL, the advising agent will
FTP-PUT the file the new location.

4. If this POST is Redirected to an https: URL, the advising agent will
attempt a SSL connection at the new location.  the agent can be expected
to produce the identical certification of the server for which it is
acting on behalf

B. File Specs (see HTML forms specification AND MIME specs FOR
application/x-form-content-encoding)

1. the following variables are defined, all are optional:

advise-agent=Name of Advise Software (SEE the User-Agent HTTP header)

contact-email
contact-name

company-name
company-address
company-locale
company-state
company-country
company-postal-code


4. the following variables can be multiply defined:

url= Correct URL
url-title: Correct URL Title (optional)
url-summary: Correct URL Summary (optional)
invalid-url=Offending URL  (optional)
referer=Referring URL (optional)

Additional variables can be suggested by contacting the hapless "Larry
Masinter" wherever and whenever he may be found.






>---------
>From: 	Larry Masinter[SMTP:masinter@parc.xerox.com]
>Sent: 	Tuesday, July 16, 1996 6:00 PM
>To: 	Erik Aronesty
>Subject: 	Re: Any interest in automatic link-updating?
>
>> GREAT!
>> should I forward this to IANA@isi.edu with you listed as the primary
>> contact or what?
>
>I think "or what". That is, this is something that does not require
>any changes to HTTP, but does require establishing a convention that
>webmasters might follow. How do you get webmasters to follow
>conventions?
>
>

Received on Wednesday, 17 July 1996 15:37:12 UTC