🚀 go-pugleaf

peterd@pjd.dev.cdx.mot.com (Peter Desnoyers) writes:

>It should be simple to extend the site-specific commands to allow users
>to request descriptions of files, etc. The problems I see are:

>- keeping a lot of file descriptions up-to-date is a pain in the rear.

>- the user interface is atrocious - "quote site describe <file>" isn't a
>  very intuitive way to get information on a file. A better FTP client
>  would fix that.

Yes, these are the two best reasons, why extending ftp won't lead to a
very comfortable solution.

My vision is this:

Let's define two terms:

     a file       this is want we know from ftp, FTAM, the UNIX filesystem,
                  Andrew,... . It is just a byte sequence with a very
                  limited number of additional attributes, e.g. a
                  filename from a very restricted characterset, a date
                  and time of last modification, a owner ID, access rights, ...

     a document   this is all a file contains plus additional attributes,
                  e.g. a one-line-description, a more detailed description,
                  keywords, a describption of the file format, the compression
                  method, a version number, an author, a worldwide unique
                  identifier, ...

A well designed document server would also have ftp (and FTAM :-) access,
but with these access methods you only see files. If it were possible
to establish a standard for "documents", then each information object
published has only once to be stored in a "document server" with all
it's additional attributes. Then every manager of an other "document server"
could get a copy of a complete document with a kind of "document access
protocol". So the additional attributes only have to be entered once
in the first document archive. All other document archives will just
copy them as easily as this is done today with ftp.

The ftp users won't see anything from this new development, because
each document contains also everything a file contains. But a user
of a "document access user agent" will get a very comfortable universe
of documents for browsing.

E.g. a document might contain the following attributes (of course a
user agent would have to present a subset of them in a nicer way):

commonName:              ISODE non-Consortium Release
suggestedFileName:   *   isode80.tar.Z
version:                 8.0
shortDescription:        A wonderful OSI development environment for UNIX
longDescription:         <the text of README>
keywords:                UNIX|OSI|networking|FTAM|QUIPU|PP|...
lastModification:    *   199204311000Z
fileFormat:              <object identifier for *.tar.Z>
author:                  Steven Hardcastle-Kille|Marshall Rose|Collin Robins|...
size:                *   8765432
legalStatus:             <object identifier for "public domain">
uniqueID:                <c=GB, o=ucl, ou=cs, ou=document store, ...>

A ftp user would only see the * attributes.

In addition, this document would have a unique name in the world wide
X.500 naming tree. This unique name could then easily be used e.g.
in a hypertext news article and everyone interested in this document
would just klick a button in the hypertext article and his user agent
would use X.500 to search the nearest copy of the document and would
then automatically download it. (What "nearest" means has to be defined
by the document user agent administrator). ISO standards like SGML/HyTime
offer powerful hypertext referencing mechanisms for text documents.

Last week, I read the latest drafts of the OSI group communication standards
X.gc and X.acc. They use DFR for access to a documents store that belongs
to a news group. So I was wondering, whether DFR might not in general
be a nice protocol for the "document access protocol" I described above.

I hope a reader of this group could post a short description of DFR
and related standards (e.g. what is a DOR?), because I'll have to wait
several weeks on the ISO documents (why are they not available in a
document store??? @!%# ).

Markus :-)

---
Markus Kuhn, Computer Science student -=-=- University of Erlangen, Germany
Internet: mskuhn@immd4.informatik.uni-erlangen.de  |  X.500 entry available
Don't worry over what other people are thinking about you. ----------------
---------- They're too busy worrying over what you are thinking about them.