HTTP Working Group                                     Koen Holtman, TUE
Internet-Draft                              Andrew Mutz, Hewlett-Packard
Expires: January 28, 1998                                  July 28, 1997

                          Feature Tag Scenarios

                 draft-ietf-http-feature-scenarios-01.txt


STATUS OF THIS MEMO

        This document is an Internet-Draft. Internet-Drafts are
        working documents of the Internet Engineering Task Force
        (IETF), its areas, and its working groups. Note that other
        groups may also distribute working documents as
        Internet-Drafts.

        Internet-Drafts are draft documents valid for a maximum of
        six months and may be updated, replaced, or obsoleted by
        other documents at any time. It is inappropriate to use
        Internet-Drafts as reference material or to cite them other
        than as "work in progress".

        To learn the current status of any Internet-Draft, please
        check the "1id-abstracts.txt" listing contained in the
        Internet-Drafts Shadow Directories on ftp.is.co.za
        (Africa), nic.nordu.net (Europe), munnari.oz.au (Pacific
        Rim), ds.internic.net (US East Coast), or ftp.isi.edu (US
        West Coast).

        Distribution of this document is unlimited.  Please send
        comments to the HTTP working group at
        <http-wg@cuckoo.hpl.hp.com>.  Discussions of the working
        group are archived at
        <URL:http://www.ics.uci.edu/pub/ietf/http/>.  General
        discussions about HTTP and the applications which use HTTP
        should take place on the <www-talk@w3.org> mailing list.

        A HTML version of this document can be found at
        <URL:http://gewis.win.tue.nl/~koen/conneg/>.

ABSTRACT

   Recent Internet applications, such as the World Wide Web, tie
   together a great diversity in data formats, client and server
   platforms, and communities.  This has created a need for various
   kinds of negotiation mechanisms, which tailor the data which is
   exchanged, or the exchange process itself, to the capabilities and
   preferences of the parties involved.

   Extensible negotiation mechanisms need a vocabulary to identify
   various things which can be negotiated on.  To promote
   interoperability, a registration process is needed to ensure that
   that this vocabulary, which can be shared between negotiation
   mechanisms, is developed in an orderly, well-specified, and public
   manner.

   This document discusses requirements and scenarios the registration
   of this vocabulary, which is the vocabulary of feature tags.
   Feature tag registration is foreseen as an ongoing, open process
   which will keep pace with the introduction of new features by
   software vendors, and other parties such as standards bodies.


TABLE OF CONTENTS

   1 Introduction

   2 Basic concepts and definitions
    2.1 Areas of negotiation and feature tags
    2.2 Complexity of negotiation
    2.3 The result in an area of negotiation

   3 Extensible negotiation mechanisms
    3.1 The need for extensible negotiation mechanisms: the case of the Web
    3.2 Extensible negotiation mechanisms for the Web
    3.3 Extensible negotiation mechanisms for other Internet protocols

   4 Feature tag registration
    4.1 IANA registration procedures
    4.2 Examples of parties who would want to register feature tags
    4.3 Feature tag registration scenario
    4.4 Volume considerations
    4.5 Danger of excessive registration

   5 Security considerations

   6 Acknowledgments

   7 References

   8 Authors' addresses


 Introduction

   Recent Internet applications, such as the World Wide Web, tie
   together a great diversity in data formats, client and server
   platforms, and communities.  This has created a need for various
   kinds of negotiation mechanisms, which tailor the data which is
   exchanged, or the exchange process itself, to the capabilities and
   preferences of the parties involved.

   Extensible negotiation mechanisms need a vocabulary to identify
   various things which can be negotiated on.  To promote
   interoperability, a registration process is needed to ensure that
   that this vocabulary, which can be shared between negotiation
   mechanisms, is developed in an orderly, well-specified, and public
   manner.

   This document discusses requirements and scenarios the registration
   of this vocabulary, which is the vocabulary of feature tags.
   Feature tag registration is foreseen as an ongoing, open process
   which will keep pace with the introduction of new features by
   software vendors, and other parties such as standards bodies.


 Basic concepts and definitions

2.1 Areas of negotiation and feature tags

   Something which can be negotiated on is called an `area of
   negotiation' in this document.  Examples of areas of negotiation
   are:

     * the MIME media type of the data which is transmitted
     * the language of the text document which is transmitted
     * the color depth of the screen on which something is to be
       displayed
     * whether the recipient supports the `floating 5 dimensional
       tables' feature
     * the fonts which are available to the recipient
     * whether a Web user prefers speed over graphical detail
     * whether the recipient is capable of displaying graphical
       content
     * whether the user prefers a blue background with purple dots over
       a green background with pictures of small furry animals, except
       on Fridays.

   A feature tag identifies a single area of negotiation.

   It is expected that the majority of feature tags will identify new
   areas of negotiation, in which the object of negotiation is to
   decide on the presence or use of some new feature in a software
   product.  This explains the name `feature tag'.

   It is recognized that there is continuous growth in the number of
   areas in which some form of negotiation is desirable.  To keep up
   with this growth, extensible negotiation mechanisms are needed,
   which refer to the feature tag vocabulary to identify new areas of
   negotiation, rather than relying on hard-coded knowledge about a
   few areas.

   To avoid the duplication of work, and to promote the interoperable
   naming of areas of negotiation across protocols and applications,
   the feature tag namespace should not be bound to a particular
   protocol or negotiation mechanism.  Also, there should be no prior
   restriction on the areas of negotiation which may be identified by
   a feature tag, other than that it must be conceivable to negotiate
   in these areas in the context of some Internet application.


2.2 Complexity of negotiation

   Negotiation processes can often be complex.  Two frequent sources
   of complexity are:

    1. An area of negotiation may be inherently complex.  For
       example, negotiating on the use of a particular media type is
       inherently more complex than negotiating on the presence of a
       single feature, because there are more possible outcomes.

    2. There may be complex of interdependencies between the choices
       in different areas of negotiation.  For example, if the
       following versions of a document are available on a Web server:

           * text/html, English
           * text/plain, French
           * audio/x-wav, German

       then the content negotiation mechanism cannot treat the areas
       of `MIME media type negotiation' and `language negotiation' as
       separate.

   It is recognized that extensible negotiation mechanisms will often
   differ in the type and amount of complexity they can handle.  Thus,
   though negotiation mechanisms share the feature tag namespace, it
   will not be the case that every tag is usable in every negotiation
   mechanism, or that every negotiation mechanism will be able to
   handle all possible interdependencies.


2.3 The result in an area of negotiation

   During negotiation, negotiation mechanisms will often need to
   transmit (canonical representations of) the possible results in
   various areas of negotiation over the wire.  Also, at the end of a
   negotiation process, the mechanism may need to return (a
   canonical representation of) the result to the application which
   invoked it.

   In many areas of negotiation, there will be a natural, canonical
   representation of the result.  For example, in the area

     * whether the recipient supports the `floating 5 dimensional
       tables' feature

   the canonical representation of the result is a boolean value (yes,
   the feature is supported, or no, the feature is not supported).  In
   the area

      * the MIME media type of the data which is transmitted

   the canonical representation of the result will be a MIME media
   type identifier like text/html or application/postscript.  In some
   areas of negotiation, the result could be a compound value (e.g. a
   coordinate in a 3D space).

   To promote interoperability, the registration entry of a feature
   tag can include a definition of the canonical representations of
   the possible results in the corresponding area of negotiation.


 Extensible negotiation mechanisms

   We call a negotiation mechanism extensible if the set of areas
   on which the mechanism can negotiate is extensible, instead of
   hard-coded inside the mechanism.

3.1 The need for extensible negotiation mechanisms: the case of the Web

   HTTP [2] has traditionally recognized four areas of negotiation:

     1. MIME media type
     2. Charset
     3. Language
     4. Encoding

   HTTP provides support for these areas of negotiation by defining
   identifiers (Accept headers) for these areas, and defining
   canonical representations of the possible results in these areas.

   Experience with the Web has shown there is a great need to
   negotiate on things which do not fit the four areas above.  This
   need has shown itself in a number of ways:

     - Web servers routinely use (abuse) other headers than the Accept
       headers for negotiation purposes.  In particular, the HTTP
       User-Agent header has been widely used by web sites to detect
       the presence of certain features in browsers, and by browsers
       to trigger certain features in web sites, even though such use
       is error-prone, and conflicts with the original purpose of the
       User-Agent header.

     - Web servers routinely use `dynamic URLs' and cookies to encode
       negotiation related information like user preferences.  This
       can be cache-unfriendly, in particular in the case of `dynamic
       URLs'.

     - During the standardization of HTTP, several proposals for
       additional Accept headers, matching additional areas of
       negotiation, were made.  These proposals have been rejected in
       favor of developing an extensible negotiation mechanism.

     - There has been pressure to extend the MIME media type parameter
       mechanism to allow for the naming of, and negotiation on, new
       features in media types already registered, something which is
       explicitly disallowed in the MIME type registration rules.  It
       was recognized that this pressure would be better addressed by
       creating a new namespace independent from the MIME media type
       space.


3.2 Extensible negotiation mechanisms for the Web

   In the IETF HTTP working group, it has been recognized that the
   number of areas for Web content negotiation is growing so rapidly
   that the IETF would never be able to keep up with this growth by by
   continually revising a negotiation mechanism with a hard-coded set
   of areas.  Instead, a number of extensible content negotiation
   mechanisms have been proposed.  All of these mechanisms share the
   need for an external vocabulary to identify areas, a vocabulary
   which can be updated quickly enough to keep pace with the
   introduction of new features by Web software vendors.

   The proposed extensible content negotiation mechanisms are
   transparent content negotiation [2], and negotiation mechanisms
   based on various forms of "active content".  In "active content"
   negotiation, the web server returns an executable program which is
   run by the browser.  This program then accesses a database of
   negotiation related settings inside the browser, and chooses and
   renders the most appropriate content.  Note that there are several
   existing and proposed forms of active content.

   To tie everything together, a browser architecture with a common
   internal database for negotiation related information has been
   proposed.  The database would be filled to reflect the capabilities
   of all browser components, both native components and plugins, and
   the preference settings made by the user.  Feature tags would serve
   as the keys under which the database entries for different areas of
   negotiation are stored.  Individual negotiation mechanisms could
   access the central database through some API.  The architecture is
   shown in the following picture.

   +-----------------------------------------------------------------+
   | BROWSER                                                         |
   |                                                                 |
   |  +------------------+  +----------+ +----------+  +-----------+ |
   |  | Native browser   |  | Plugin 1 | | Plugin 2 |  |User       | |
   |  | rendering engine |  +----------+ +----------+  |preferences| |
   |  +------------------+     |               |       +-----------+ |
   |               |           |               |             |       |
   |               V           V               V             V       |
   |        +------------------------------------------------------+ |
   |        | Common internal database for negotiation information | |
   |        +------------------------------------------------------+ |
   |               |         |         |                |            |
   |              API       API       API              API           |
   |               |         |         |                |            |
   |               V         V         V                V            |
   |         +---------+ +---------+             +-------------+     |
   |         | Java    | | JScript |             | transparent |     |
   |         | based   | | based   | ..etc       | content     |     |
   |         | active  | | active  |             | negotiation |     |
   |         | content | | content |             | engine      |     |
   |         +---------+ +---------+             +-------------+     |
   |                                                                 |
   +-----------------------------------------------------------------+


3.3 Extensible negotiation mechanisms for other Internet protocols

   Extensible negotiation mechanisms for Internet printing and
   Internet fax are currently under investigation in the IETF.

   It has been proposed to make multipart/alternative negotiation in
   Internet mail more extensible, in particular if the mail client is
   part of a Web browser, by adapting some of the protocol elements
   developed for transparent content negotiation [2] to Internet mail.

 Feature tag registration

4.1 IANA registration procedures

   Examples of IANA registration procedures can be found in [1].

   There has been some confusion over what the IANA will register.
   Jon Postel told us that:

     The IANA is pretty good at keeping lists.  It is not so good at
     evaluating the merits (or lack thereof) of the requests to add
     things to lists.  [...]  So, yes the IANA would keep the list of
     "feature tags" provided that that there is either a very simple
     way to decide if requests pass muster, or a designated someone
     else will be available to make that decision.

   So two types of registration namespaces can be created:

      a) a space with feature tag review process performed by the IETF

      b) a space with very basic registration rules which do not take
         the merit of the feature tag into account.  To quote [1],
         this type of registration "does not imply endorsement,
         approval, or recommendation by the IANA or IETF or even
         certification that the specification is adequate."

   If extensible negotiation mechanisms are to keep up with the speed
   of development in the Web, a type b) registration process for
   feature tags seems necessary, if only because the IETF does not
   have the manpower required for a review process which can keep up
   with the speed of Web development.

   It is proposed that feature tag registration closely mimics the new
   MIME media type registration rules in [1], providing both type a)
   and b) namespaces.  This proposal is based on the observation that
   the rules in [1] seem to be working nicely.


4.2 Examples of parties who would want to register feature tags

   Feature registration allows for the quick introduction of new areas
   of negotiation in extensible negotiation mechanisms.  In a Web
   context, examples of parties which might want to introduce new
   areas of negotiation are:

     1. Browser and browser component vendors, when inventing and
        implementing new features or components.

     2. The IETF or some other standards body, when creating a new
        standard for a content format, or when identifying a new type
        of user preference (for example a preference for
        representations without animated gifs).

     3. Content authors, when identifying a new type of user
        preference and creating new content to match.

   A fast registration process is mainly needed for registration by
   group 1 and 3.  For 2, a slower process would suffice.


4.3 Feature tag registration scenario

   Below is a scenario, in a Web context, for the registration of a
   feature tag which succeeds in being generally used.

    Time    Action
   (months)

    t+0    Browser vendor A invents the new feature XY.

    t+1    Vendor A starts implementing XY, and completes a
           feature tag registration form for the `g.xy' tag.

    t+2    Vendor A submits the form and the IANA registers the `g.xy'
           feature tag.

    t+2.1  Vendor A releases a new browser version, which
            a) implements the feature XY
            b) has an entry under `g.xy' in its negotiation database,
               which tells extensible negotiation mechanisms that this
               feature is present

    t+2.5  `Early adopter' content authors start making content
           representations which use XY.

    t+3    Vendor B starts implementing XY in their own browser.

    t+3    The `g.xy' tag appears in lists of useful tags maintained by
           third parties.

    t+3.5  Vendor B releases a new browser version, which
            a) implements the feature XY
            b) has an entry under `g.xy' in its negotiation database,
               which tells extensible negotiation mechanisms that this
               feature is present

    t+3.5  Many content authors start making content representations
           which use XY.

    t+4    Vendor C starts implementing XY, and invents the extension
           XY_COLOR.

    t+4.5  Vendor C registers the `g.xy_color' feature tag.

    t+4.5  Vendor C releases a new browser version, which
            a) implements the features XY and XY_COLOR
            b) has appropriate entries under `g.xy' and `g.xy_color'
               in its database

    t+10   90% of all deployed browsers support XY. Content authors
           start using XY it without bothering to provide an alternate
           representation.


4.4 Volume considerations

   Feature tag registration which will have to keep pace with the
   introduction of new features by vendors.

   In particular in the Web domain, vendors are moving fast, and this
   will inevitably lead to a feature tag namespace which contains a
   lot of tags.  Also, a lot of tags will be `dead' tags, tags related
   to features which failed to take off and gain widespread use.
   Compare this to the situation in the USENET newsgroup namespace.

   Like a list of all MIME media types, a list of all registered
   feature tags will generally be too long to be useful to any content
   author.  Third parties could filter the feature tag namespace and
   compile short lists of useful tags.  In the Web domain, Web
   indexing robots could, while traversing the web, gather statistics
   about actual use of feature tags; these statistics could help in
   compiling lists.


4.5 Danger of excessive registration

   One danger for the feature tag namespace is the emergence of
   excessive registration as seen in the internet domain name system
   (DNS) namespace.

   We speculate that the various forces which contributed to the DNS
   registration problems are absent for feature tags: feature tags
   will not be very visible to end users, and registration of a
   feature tag does not mean you get exclusive use.

   We therefore do not expect excessive registration to occur.  We
   note it has not occured (so far) in the MIME media type namespace.
   Of course it is possible to update the registration procedure if
   excessive registration _does_ occur.  A necessary precaution is to
   reserve a part of the feature tag namespace for future use.


 Security considerations

   When used, negotiation mechanisms usually reveal some information
   about one party to other parties.  This may raise privacy concerns,
   and may allow a malicious party to make more educated guesses about
   the presence of security holes in the other party.


 Acknowledgments

   The idea of creating a vocabulary of areas of negotiation, which is
   maintained in a central open registry, is due to discussions on
   extensible negotiation mechanisms in the IETF HTTP working group.
   The authors wish to thank Larry Masinter and Graham Klyne for
   contributing to discussions about feature tag registration.


 References

   [] N. Freed, J. Klensin, J. Postel, Multipurpose Internet Mail
       Extensions (MIME) Part Four: Registration Procedures.  RFC
       2048, BCP 13, Network Working Group, November 1996

   [] R. Fielding, J. Gettys, J. C. Mogul, H. Frystyk, and
       T. Berners-Lee.  Hypertext Transfer Protocol -- HTTP/1.1.  RFC
       2068, HTTP Working Group, January, 1997.

   [] K. Holtman, A. Mutz.  Transparent Content Negotiation in HTTP.
       Internet-Draft draft-ietf-http-negotiation-03.txt, HTTP Working
       Group.  July 1997.


 Authors' addresses

   Koen Holtman
   Technische Universiteit Eindhoven
   Postbus 513
   Kamer HG 6.57
   5600 MB Eindhoven (The Netherlands)
   Email: koen@win.tue.nl

   Andrew H. Mutz
   Hewlett-Packard Company
   1501 Page Mill Road 3U-3
   Palo Alto CA 94304, USA
   Fax +1 415 857 4691
   Email: mutz@hpl.hp.com


Expires: January 28, 1998