Re: Why is From: limited?

At 11:38 AM 3/27/95, Roman Czyborra wrote:
>Albert Lunde:
>
>> In real life the Unix UID that "owns" a document may have nothing to
>> do with the author, for example on a CWIS where everything is posted
>> by a few people.
>
>But those few people will have the write permission on the documents
>and know who to forward the criticism to.

This is BOGUS. It makes this feature of NO value. Suppose "bin" owns all
WWW documents? Do you think a system admin wants to spend all day
forwarding stuff to the appropriate person? Using the group or owner is
simply a hack that only works in *some* situations on *some* hosts running
*some* operating systems. It is an implementation detail of a protocol that
should be implementation independent.

>> There are already HTML solutions in wide use for
>> indicating/contacting the author of HTML docs.
>
>What do you do about non-HTML documents served on HTTP?  Do you put
>comments on all of your GIFs?

No, that is what copyright notices and artist/photographer attributions are
for. There are mechanisms in HTML 3.0 for doing this, and that is why I
pointed this out as a bad thing for HTTP to worry about.

>
>> Assuming that a feature in the protocol is easy to implement based
>> on the effort to hack it into Unix is not a valid measure.
>
>Assuming that an optional header cannot be generated on all machines
>is not a valid reason to rule it out for those who could.

I didn't say it might not be a good idea. I said that it couldn't be
implemented in an efficient, usable fashion across multiple platforms. The
proposed "technique" was a hack and it didn't even meet the needs of all
Unix users. HTTP servers ultimately need to sit on databases (object
bases), not file systems. When this happens, things like this will be easy
to implement. In the meantime, the value of such protocol candy is minimal
when other, better techniques exist in HTML for doing the same thing.

>> Third, there is no correlation between an operating system-specific
>> "owner" of a file and the author of the file
>
>For the documents at our server, there is a very strong correlation.

But not at mine. Now, whose site is "right"? Extrapolating that all sites
must be like your site is faulty logic...

>> this field has no place as a required header field
>
>As I said, I was asking for the From: in the response to be as
>optional as the From: in the request.

If it is optional, then it is useless. And if it is required, then it is an
unreasonable invasion of privacy for authors that require or desire
anonymity. You need to understand that there are better mechanisms for
representing authorship via WWW than a header field in the HTTP protocol.
Please read through the URI/URN standards and the HTML 3.0 standard and I
think you'll find plenty of stuff to sate your appetite for features.

>> If it's not required, then it is of little value.
>
>LINK REV="made" is not required either and still it is very valuable
>in the documents that provide it.

"in the documents!!!" Not "in the protocol."

>> The bottom line is that "authorship" of content has no place in the
>> transport protocol.
>
>I find it very convenient that my RFC822 mail transport automatically
>places my name in the header so I don't have to type it into every
>message.  I am envisioning the same convenience for HTTP.

E-mail is for person to person communication. Of course you want to know
where the message came from. WWW is for anonymous publisher to anonymous
reader communication. See the difference? (PLEASE read the proposals for
the URN stuff. It is the appropriate mechanism for doing what you desire.)


-----------------------------------------------------------------------
Chuck Shotton
cshotton@biap.com                                  http://www.biap.com/
cshotton@oac.hsc.uth.tmc.edu                           "I am NOT here."

Received on Monday, 27 March 1995 07:54:26 UTC