Protocol-A.texi

\input texinfo                         @c -*-texinfo-*-
@c
@c FIXME: Explain how the garb works with nice and keep-commented
@c FIXME: Make all types clickable in HTML (and info?)
@c FIXME: Move predefined aux-items to separate chapter
@c FIXME: Move misc-infos to chapter?
@c FIXME: Check for too much text before @menu.
@c FIXME: error-reply defined twice (conflicting).
@c FIXME: add link to Name Expansion from relevant reqs.
@c FIXME: Name Expansion collate table info is outdated.
@c FIXME: Overview -> Preface
@c FIXME: Move syntax from overview to somewhere else
@c FIXME: sentence-end-double-space!
@c FIXME: "Foo bar (bar baz.)" or "Foo bar (bar baz)."?
@c
@c $Id: Protocol-A.texi,v 1.135 2001/05/01 20:30:10 ceder Exp $
@c %**start of header
@setfilename protocol-a.info
@settitle LysKOM Protocol A
@setchapternewpage odd
@c @afourpaper

@set PROTOEDITION 10.4
@set PROTOVER 10
@set VERSION 1.0.4



@c
@c Define a few strings that contains characters from ISO Latin-1.
@c

@iftex

@set IAM Inl@"a{}gg @}t mig
@set Pell P@"ar Emanuelsson
@set Kent Kent Eng@-str@"om@penalty-10000
@set presconf 27HPresentation (av nya) m@"oten

@end iftex
@ifnottex

@set IAM Inlägg @}t mig
@set Pell Pär Emanuelsson
@set Kent Kent Engström
@set presconf 27HPresentation (av nya) möten

@end ifnottex



@c
@c Index setup
@c

@c Index for asynchronous messages.
@defcodeindex am

@c Merge all indices.
@syncodeindex am fn
@syncodeindex tp fn



@c
@c Special markup for Protocol-A.texi
@c

@c @req{login} is used for protocol requests.

@macro req {n}
@code{\n\}
@end macro

@c @reqlink{login} is used for protocol requests, that shuld be linked.

@ifhtml
@macro reqlink {n}
@code{@ref{\n\}}
@end macro
@end ifhtml

@ifnothtml
@macro reqlink {n}
@code{\n\} (@pxref{\n\})
@end macro
@end ifnothtml

@c @reqdlink{login}...@linkhere{} is used for protocol requests, that
@c shuld be linked, where the link should occur a few words after the
@c request itself.

@ifhtml

@c HTML version: make the word a link.
@macro reqdlink {n}
@code{@ref{\n\}}@c
@unmacro linkhere
@macro linkhere
@end macro
@end macro

@end ifhtml

@ifnothtml
@ifnottex

@c Info version: create linkhere as a macro that contains a reference.

@macro reqdlink {n}
@code{\n\}@c
@unmacro linkhere
@macro linkhere
(@pxref{\n\})
@end macro
@end macro
@end ifnottex

@iftex

@c TeX version: this produces output similar to the Info version.

@tex
    \gdef\reqdlink#1{%
        \code{#1}%
        \gdef\linkhere{ (\pxref{#1})}%
    }
@end tex
@end iftex
@end ifnothtml

@c @aux{mx-allow-filter} is used for aux-item names.
@macro aux {n}
@code{\n\}
@end macro

@c @async{async-login} is used for async names.
@macro async {n}
@code{\n\}
@end macro

@c @asynclink{async-login} is used for async names, that shuld be linked.

@ifhtml
@macro asynclink {n}
@code{@ref{\n\}}
@end macro
@end ifhtml

@ifnothtml
@macro asynclink {n}
@code{\n\} (@pxref{\n\})
@end macro
@end ifnothtml

@c @asyncdlink{async-login}...@linkhere{} is used for async messages,
@c that shuld be linked, where the link should occur a few words after
@c the async name itself.

@ifhtml

@c HTML version: make the word a link.
@macro asyncdlink {n}
@code{@ref{\n\}}@c
@unmacro linkhere
@macro linkhere
@end macro
@end macro

@end ifhtml

@ifnothtml
@ifnottex

@c Info version: create linkhere as a macro that contains a reference.

@macro asyncdlink {n}
@code{\n\}@c
@unmacro linkhere
@macro linkhere
(@pxref{\n\})
@end macro
@end macro
@end ifnottex

@iftex

@c TeX version: this produces output similar to the Info version.

@tex
    \gdef\asyncdlink#1{%
        \code{#1}%
        \gdef\linkhere{ (\pxref{#1})}%
    }
@end tex
@end iftex
@end ifnothtml

@c @type{Conference} is used for LysKOM types.
@macro type {n}
@code{\n\}
@end macro

@c @lt{Conference} is used for linked types.  The link is only present
@c when it is unobtrusive.  It should only be used inside
@c @example...@end example.

@ifhtml
@macro lt {n}
@ref{\n\}@c
@end macro
@end ifhtml

@ifnothtml
@macro lt {n}
\n\@c
@end macro
@end ifnothtml

@c @priv{wheel} is used for privilege bits.
@macro priv {n}
@code{\n\}
@end macro

@c @conftype{rd-prot} is used for conference type bits.
@macro conftype {n}
@code{\n\}
@end macro

@c @misc{bcc-recpt} is used for misc-info names.
@macro misc {n}
@code{\n\}
@end macro

@c @rarg{passwd} is used to reference arguments of requests.  This is
@c *only* used in the nodes that defines the requests.
@macro rarg {n}
@code{\n\}
@end macro

@c @aarg{text-no} is used to reference arguments of asynchronous
@c messages.  This is *only* used in the nodes that defines the
@c relevant asynchronous message.
@macro aarg {n}
@code{\n\}
@end macro

@c @errorcode{no-such-text} is used to mark error codes.
@macro errorcode {n}
@code{\n\}
@end macro

@c @field{ident-user} is used to mark fields of struct types.
@macro field {n}
@code{\n\}
@end macro

@c @reqexample is used immediately before Protocol-A example
@c fragments.
@macro reqexample
@need 2000
@i{Example:}
@end macro

@c %**end of header

@ifinfo
This is the LysKOM Protocol A specification, edition
@value{PROTOEDITION}.  It specifies version @value{PROTOVER} of the
protocol.  It corresponds to version @value{VERSION} of lyskomd.

Copyright @copyright{} 1995-2000 Lysator ACS.

Permission is granted to make and distribute verbatim copies of this
specification provided the copyright notice and this permission notice
are preserved on all copies.

@ignore
Permission is granted to process this file through TeX
and print the results, provided the printed document
carries a copying permission notice identical to this
one except for the removal of this paragraph (this
paragraph not being relevant to the printed manual).
@end ignore

Modified versions of this document may be redistributed with the added
condition that all modifications not cleared with the LysKOM development group
are clearly marked and that the entire modified work be redistributed under the
same conditions as the original.

Permission is granted to copy and distribute translations of this manual into
another language under the same conditions as for modified versions.
@end ifinfo

@c FIXME: What does this do?
@iftex
@parindent 0pt
@font@tensltt=cmsltt10
@tex
\global\def\rett#1{{\let\t\sltt\tt #1}}
\global\def\sltt#1{{\fam\ttfam\tensltt\let\t\rett #1}}
\global\let\t\sltt
@end tex
@end iftex

@titlepage
@title{LysKOM Protocol A}
@subtitle{Protocol version @value{PROTOVER}}
@subtitle{Edition @value{PROTOEDITION}}
@subtitle{(lyskomd @value{VERSION})}
@author by the LysKOM Developers

@page
@vskip 0pt plus 1filll
Copyright @copyright{} 1995-2000 Lysator ACS

Permission is granted to make and distribute verbatim copies of this document
provided the copyright notice and this permission notice are preserved on all
copies.

Modified versions of this document may be redistributed with the added
condition that all modifications not cleared with the LysKOM development group
are clearly marked and that the entire modified work be redistributed under the
same conditions as the original.

Permission is granted to copy and distribute translations of this manual into
another language under the same conditions as for modified versions.

@end titlepage
@ifnothtml
@contents
@end ifnothtml

@dircategory LysKOM
@direntry
* Protocol A: (protocol-a).             The LysKOM Protocol A specification.
@end direntry

@ifhtml
@html
<a href="http://www.lysator.liu.se/" ><img
src="http://www.lysator.liu.se/icons/lyslogo.gif" alt="Lysator"
width=40 border=0 height=40 align=left ></a>

<a href="http://www.lysator.liu.se/lyskom/" ><img
src="http://www.lysator.liu.se/icons/lyskom.gif" alt="LysKOM" width=40
border=0 height=40 align=left ></a>
<br clear=all>
@end html
@end ifhtml
@ifnottex
@node Top
@top LysKOM Protocol A

This document specifies version @value{PROTOVER} of LysKOM Protocol A.
This is edition @value{PROTOEDITION} of the specification.  It
corresponds to version @value{VERSION} of lyskomd.

The most up-to-date version if this document can always be found at 
@uref{http://www.lysator.liu.se/lyskom/protocol/}.
@end ifnottex

@menu
* Preface::                     
* Introduction::                
* Fundamentals::                
* LysKOM Data Types::           
* Protocol Requests::           
* Asynchronous Messages::       
* Error Codes::                 
* Aux-Item Types::              
* Name Expansion::              
* LysKOM Content Types::        
* The User Area::               
* Writing Clients::             
* Importing and Exporting E-Mail::  
* Future changes::              
* Protocol Version History::    
* Document Edition History::    
* Index::                       
@end menu

@node Preface
@chapter Preface

@iftex
This document specifies version @value{PROTOVER} of LysKOM Protocol A.
This is edition @value{PROTOEDITION} of the specification.  It
corresponds to version @value{VERSION} of lyskomd.

The most up-to-date version if this document can always be found on
the web at @uref{http://www.lysator.liu.se/lyskom/protocol/}.
@end iftex

LysKOM is a conferencing system@footnote{Or in modern terms, enabling
technology for Computer-Supported Cooperative Work (CSCW).}. Similar
systems were QZ-KOM and PortaCOM@footnote{Also known as ``PottaKOM'' and
``BortaKOM.''}. The LysKOM system is copyrighted by Lysator Academic
Computing Society and distributed under conditions of the GNU Public
License. LysKOM and its documentation is provided ``as is'' without
warranty of any kind.

Anything described here as ``unspecified'' is liable to change in
future protocol versions.

This specification is the work of several people. The main contributors have
been
Per Cederqvist @email{ceder@@lysator.liu.se}, 
David Byers @email{byers@@lysator.liu.se},
@value{Pell} @email{pell@@lysator.liu.se},
Thomas Bellman @email{bellman@@lysator.liu.se}, 
Lars Aronsson @email{aronsson@@lysator.liu.se},
Linus Tolke @email{linus@@lysator.liu.se} and
@value{Kent} @email{kent@@lysator.liu.se}.

The LysKOM developers can be reached by email to
@email{lyskom@@lysator.liu.se}.

@node Introduction
@chapter Introduction

This chapter introduces the concepts used in LysKOM, such as articles,
conferences and sessions. 

@menu
* Articles::                    
* Conferences::                 
* Persons and Sessions::        
* The Misc-Info List::          
* The Aux-Item List::           
* Security::                    
* Membership and Reading::      
@end menu


@node Articles
@section Articles

An article is represented as a value of the type @type{Text-Stat} and a
string containing the article contents. An article will usually have one
or more recipients and may be a comment or footnote to other articles.
Each article is kept in the database until it is older than the
@field{nice} value of each of its recipients and it is not marked by any
user.

There is an array of @type{Misc-Info} included in the
@type{Text-Stat}.  This array contains information about recipients,
senders, comments and footnotes.

Each article is identified by a number, the global@footnote{The number
is not truly global; it is local to a specific LysKOM server.}
article number (the @type{Text-No}).  Global numbers are assigned in
ascending order to new articles, and are never reused.  If an article
has recipients it will also have a local number for each recipient
(the @type{Local-Text-No}).  Local numbers are used in some data
structures to provide more compact storage and to provide an ordering
of articles for a particular recipient.  Local numbers are assigned in
ascending order and are never reused for a particular recipient,
though different recipients will have articles with the same local
numbers.

Occasionally it is necessary to map between local and global numbers.
The server call @reqdlink{local-to-global} does this@linkhere{}.



@node Conferences
@section Conferences

Conferences hold articles. They are represented in the protocol as a
data type called @type{Conference}. Each conference has a
@dfn{creator}, the person who created the conference, and a
@dfn{supervisor}, a conference whose members can modify the conference.
If the supervisor is a person, the members of that person's mailbox
are supervisors, which in most cases is only that person. We have also
introduced a type called @type{UConference} (pronounced micro-conf-stat)
which holds a subset of the information contained in the full
@type{Conference} type. Use the @type{UConference} type whenever
possible since it places a much smaller load on the LysKOM server.

Each conference has a type, which is essentially a collection of
boolean flags. Currently the flags @conftype{rd-prot},
@conftype{letterbox}, @conftype{secret}, @conftype{original},
@conftype{allow-anonymous} and @conftype{forbid-secret} are defined.

@table @conftype
@item rd-prot
The conference is protected from reading by non-members. Persons become
members by having one of the existing members or supervisors add him or
her to the conference. This restriction is enforced by the server.
@item original
Conferences of this type are intended for original articles only.
Comments are to be redirected to the super-conference instead. This
restriction is currently not enforced by the server; clients must
implement this functionality.
@item letterbox
Conferences of this type are connected to persons. Letters to a person
are sent to the mailbox and the name of the mailbox is synchronized
with the person name. It is currently not possible to explicitly set or
clear this flag on a conference.
@item secret
Conferences of this type are secret. The server will not divulge any
information about the existence of the conference to persons who are not
members or supervisors of the conference. If a mailbox is made secret,
that person cannot log in using the person name, but must specify a
person number instead.
@item allow-anonymous
Conferences of this type accept anonymous articles. Other conferences
will reject anonymous articles.
@item forbid-secret
Conferences of this type do not allow secret members. If a conference is 
changed to this type, preexisting secret members remain secret.
@end table



@node Persons and Sessions
@section Persons and Sessions

Persons are represented in the protocol by values of the type
@type{Person}. Associated with persons are statistics, a set of personal
flags and a set of privileges (@pxref{Security}). Persons are also
associated with a conference that has the same number as the person and
the @conftype{letterbox} bit set.

Connections to the server are represented as values of the type
@type{Static-Session-Info}, @type{Session-Info-Ident} or
@type{Session-Info}. Sessions have session number that are unique for
each session in the lifetime of the server execution. A single user can
have several sessions running at once. The session is not released until
the network connection is closed; a user can log in and out repeatedly
in a single session.


@node The Misc-Info List
@section The Misc-Info List

The @type{Misc-Info} list contains tagged data. The fields are sent in
groups pertaining to a particular type of information: information about
recipient; carbon copy recipient; blank carbon copy recipient;
comment to; footnote to; comment in
and footnote in. The information groups may be sent in any order and
there may be any number of groups. Within each group the elements are
always sent in the order listed below.

@subsection Recipient

@table @misc
@item recpt
Starts a recipient group. It contains the conference number of a
recipient of the article.
@item loc-no
Always present within a recipient group. It contains the local text
number of the article in the conference specified by the preceding
@misc{recpt} field.
@item rec-time
If the recipient is a person, this element is added by the server when
the recipient marks the article as read. It contains the time when the
text was read.
@item sent-by
Present when the recipient was added by a person other than the author
(after the article was created.) It contains the person number of the
person who added the recipient.
@item sent-at
Present when the recipient was added after the article was created. It
contains the time when the recipient was added.
@end table


@subsection Carbon Copy (CC) Recipient

The carbon-copy recipient group is identical to the recipient group
above. The difference is how new comments to an article with a recipient
or carbon-copy recipient are treated. A comment to an article is sent to
all recipients, but not to carbon-copy recipients of the original
article. This difference is enforced by the clients.

@table @misc
@item cc-recpt
Starts a carbon-copy recipient group. It contains the conference number
of a carbon-copy recipient of the article.
@item loc-no
Always present in a CC recipient group. It contains the local text
number of the article in the conference specified by the most recent
@misc{cc-recpt} field.
@item rec-time
Present after the CC recipient has read the article. It contains the
time when the article was read. Since only persons can read articles
this will only be seen if the CC recipient is a person.
@item sent-by
Present when a CC recipient was added by a person other than the author
after the article had been created. It contains the person number of the
person who added the CC recipient.
@item sent-at
Present when a CC recipient was added after the article had been
created. It is the time when the CC recipient was added.
@end table


@subsection Blank Carbon Copy (BCC) Recipient

The blank carbon-copy recipient group is identical to the carbon-copy
recipient group above. The difference is the visibility of the
information. A carbon-copy recipient group is visible to anyone that is
allowed to fetch both the text status of the involved text and the
conference status of the involved conference.  (That is, as long as the
conference isn't secret everybody is allowed to see the carbon-copy
recipient group.)

A BCC recipient group is basically only visible to members and
supervisors of the recipient.  Persons that have the right to become a
member of the recipient can also see it, as can the author of the text
(unless the recipient is secret to him).  This is enforced by the
server.

This type of group was introduced in protocol version 10.  When
old-style calls such as @reqlink{get-text-stat-old}
are used this will be converted to a CC recipient group by the server
for the benefit of clients that don't understand this group.  (This
conversion will of course only be performed when the user is allowed
to se the blank carbon copy.)

@table @misc
@item bcc-recpt
Starts a blank carbon-copy recipient group. It contains the conference
number of a blank carbon-copy recipient of the article.
@item loc-no
Always present in a BCC recipient group. It contains the local text
number of the article in the conference specified by the most recent
@misc{bcc-recpt} field.
@item rec-time
Present after the BCC recipient has read the article. It contains the
time when the article was read. Since only persons can read articles
this will only be seen if the BCC recipient is a person.
@item sent-by
Present when a BCC recipient was added by a person other than the author
after the article had been created. It contains the person number of the
person who added the BCC recipient.
@item sent-at
Present when a BCC recipient was added after the article had been
created. It is the time when the BCC recipient was added.
@end table


@subsection Comment To

@table @misc
@item comm-to
Always present when the article is a comment to another article.
@item sent-by
Present when the article was added as a comment by a person other than
the author, after the article had been created. It contains the person
number of the person who added the article as a comment.
@item sent-at
Present when the article was added as a comment after the article had
been created. It contains the time when it was added as a comment.
@end table


@subsection Footnote To

@table @misc
@item footn-to
Always present when the article is a footnote to another article.
@item sent-at
Present when the article was added as a footnote after the article had
been created. It contains the time when it was added as a footnote.
@end table


@subsection Comment in

@table @misc
@item comm-in
Present when there are comments to this article. It contains the article
number which is a comment to this article.
@end table


@subsection Footnote in

@table @misc
@item footn-in
Present when there are footnotes to this article. It contains the
article number which is a footnote to this article.
@end table


@node The Aux-Item List
@section The Aux-Item List

The aux-item list is used as a generic extension mechanism in the LysKOM
server and in protocol A. 

@menu
* About Aux-Items::             
* Predefined Aux-Item Types::   
* Client-Specific Aux-Item Types::  
* Experimental Aux-Item Types::  
* Defining New Aux-Item Types::  
@end menu

@node About Aux-Items
@subsection About Aux-Items

Aux-items were introduced in protocol version 10 as a mechanism for
extending the conference, text and server information structures without
changing the protocol. Persons were excluded since nobody could figure
out a case where setting an aux-item on the mailbox wasn't as good as
setting it on the person (another reason was that I was fed up writing
aux-item code by the time they were working on texts and conferences.)

The exact structure of an aux item is specified elsewhere
(@pxref{LysKOM Data Types}). The important fields here are the
@field{aux-no}, @field{tag} and @field{data} fields.

The @field{aux-no} field is used to identify an item. The
@field{aux-no} together with a text or conference number uniquely
identifies a particular aux item.  Items are numbered from one and up
within each item list. Once assigned, the @field{aux-no} for an item
is never changed. New items are guaranteed to be assigned numbers that
have never been used before within a particular list.

The @field{tag} field identifies the type of aux item. It is used by
the server and by clients to figure out how to interpret the data
field, and by the server to decide if the item needs special
treatment.

The @field{data} field is a simple string. The meaning of the string
is determined by the @field{tag} field, but since it is a string,
clients that have no understanding of the contents can successfully
parse the item anyway (in contrast to items in the misc-info list.)



@node Predefined Aux-Item Types
@subsection Predefined Aux-Item Types

Predefined Aux-Item types are part of Protocol A, and clients should
support all of them. As with other parts of the protocol, changes to
these definitions will be made backwards-compatible, if possible.

Creation and deletion of items with a predefined type can cause
arbitrarily complex and wonderous behavior in the server. Furthermore,
the server may place constraints on the items with regard to content,
flags, who can create them, to what objects they can be attached and
so forth. The server may also silently enforce specific values for any
field of an item, regardless of what the client requests.

All items with tags in the range 1-9999 and 30000 and up are considered
predefined. If a client attempts to create an item with a tag in this
range, but the server has no idea what that tag means, the server will
return an error (illegal-aux-item.)


@node Client-Specific Aux-Item Types
@subsection Client-Specific Aux-Item Types

Client-specific items do not cause the server to perform any magic. All
the flags (except the delete flag) are left untouched, the data is not
validated in any way, and anyone can create any item. If you need more
server support than this, your item should be on the predefined list.

All tags in the range 10000-19999 are reserved for clients. Blocks of
100 numbers at a time can be assigned to specific clients. A client
should never create items with tags in a range assigned to another
client or in an unassigned range. Assigned ranges will never change.

Currently, the following ranges are assigned to clients:
@itemize @bullet
@item 10000-10099: The Elisp Client
@item 10100-10199: komimportmail
@end itemize

If you want a range of numbers, send e-mail to the LysKOM development
group.

@menu
* komimportmail Aux-Item Types::  
@end menu

@node komimportmail Aux-Item Types
@subsubsection komimportmail Aux-Item Types

During the design and implementation of the @command{komimportmail} mail
importer, some new aux-items were defined to solve specific problems.
These aux-items are documented here:

@table @samp

@item mx-mime-belongs-to [10100] (text)

Data is a decimal text number that this text is an attachment to.  Most
likely, the current text is also a comment (or perhaps a footnote) to
the text mentioned in the aux-item. A client can use this aux-item to
alter the display format of the text (stating that this is an
attachment, not a normal comment).


@item mx-mime-part-in [10101] (text)

Data is a decimal text number of a text that is an attachment to the
current one. In other words: this is the converse of mx-mime-belongs-to.
A client can use this aux-item to know which comments to mark as
attachments; the remaining comments are assumed to be normal.

@item mx-mime-misc [10102] (text)

Data is a string that contains all of the MIME headers for the current
text. It is set by the importer. The fields are concatenated with
"\n".

Clients are encouraged to provide a command to display this.

@item mx-envelope-sender [10103] (text)

Data is the envelope sender of an imported text. The mail server is
supposed to pass this information to the importer, for inclusion here.

@item mx-mime-file-name [10104] (text)

Data is the file name of an attachment. Most likely, the importer gets
this information from a @code{name} parameter on a @code{Content-Type}
MIME header line.

Clients are encouraged to use this file name as the default file name
when the user chooses to save the text.

@end table

@node Experimental Aux-Item Types
@subsection Experimental Aux-Item Types

Experimental numbers are free for all. Use 'em any way you want. All
numbers in the range 20000-29999 are for experimental use. 



@node Defining New Aux-Item Types
@subsection Defining New Aux-Item Types

If you want a new predefined item type, just document what it does, what
the data format looks like and what the server is to do with the item
and send this to the LysKOM development group. We'll assign a number to
your item and put the documentation in this document. 

If you're not sure what you want the data to look like yet, make a note
in your documentation that the data format might change. Once you have a
data format you're happy with, update the documentation so others may
use your item.

If you need serious magic in the server (more than can be specified with
the lyskomd configuration file), you'll probably have to write the code
yourself, or hope that the development group thinks your idea is so cool
we do the job for you.

The idea is not to reject any type of item, unless there's already an
item type that does the job just as well. Adding item types should be a
much less painful process than adding new calls.





@node Security
@section Security

Security in LysKOM is based on two components. Each person has a set
of privileges and each session has a security level. Rights in the
system require both the sufficient privileges and a sufficient
security level. The privileges currently available are @priv{wheel},
@priv{admin}, @priv{statistic}, @priv{create-conf}, @priv{create-pers}
and @priv{change-name}. Security levels range from 0 to 255.

@table @priv
@item wheel
@emph{Normally not assigned}
@table @asis
@item Level 0
Person may always log in, even when LysKOM is crowded.
@item Level 6
Person may set Priv-Bits for all persons.
@item Level 7
Person may set password for all persons.
@item Level 8
Person acts as supervisor for everything.
@item Level 10
Person can read all articles.
@end table

@item admin
@emph{Normally not assigned}
@table @asis
@item Level 1
Shut down the server@*
Set motd-of-kom@*
Read last-login
@item Level 2
Read status of secret conferences and persons@*
Read the protected parts of person and conference statuses@*
Read the entire text status, even when there are secret recipients
@item Level 3
Change everybody's names
@item Level 4
Add/remove members@*
Add/remove recipients to articles
@item Level 5
Set super-conference@*
Remove articles
@item Level 6
Set administrator
@end table

@item statistic
@emph{Normally not assigned}
@table @asis
@item Level 2
Read the statistics portions of persons, even if protected
@end table

@item create-conf
@emph{Normally assigned}
@table @asis
@item Level 0
Create conferences
@end table

@item create-pers
@emph{Normally assigned}
@table @asis
@item Level 0
Create persons
@end table

@end table


@node Membership and Reading
@section Membership and Reading

Persons' memberships in conferences are represented in the protocol as
arrays of @type{Membership}-typed values. This structure contains
information about how and when the membership was created and which
texts have been read in the conference.

There are two kinds of memberships. An active membership indicates that
the person is actively participating in the conference, wants to know if 
there are unread texts and wants to receive messages send to the
conference. A passive membership is similar to no membership at all. The 
person is still a member but will not receive messages sent to the
conference and will not be notified when there are new texts. From the
user's perspective, passive membership should be like no membership at
all, but the server still remembers what the user has read in the
conference while he or she was an active member. Since protocol version
10 a bit in the membership type field of the membership structure
indicates the type of membership. Previously the server did not support
passive memberships, but there was a convention that clients should
treat the priority level zero as a passive membership.

The membership record indicates which texts have been read through the
@field{last-text-read} and @field{read-texts} fields. All texts with local 
numbers up to @field{last-text-read} have been read. In addition, all
texts with local numbers contained in the @field{read-texts} array have
been read.

Finding out which articles a person has read in a particular conference
requires a few calls. Normally, a client will retrieve a batch of
perhaps 50 articles at a time. The outline of the process is as follows:

@enumerate
@item Fetch the membership to get the @field{last-text-read}
@item Use @reqlink{local-to-global} to translate a number of local
numbers to global numbers.
@item Remove the global numbers corresponding to local numbers contained 
in @field{read-texts} from the result.
@item Get and translate more texts as needed.
@end enumerate

The process is complicated because of the translation between local
and global text numbers. If the server does not implement the
@reqdlink{local-to-global} call@linkhere{}, it is possible to use the less
efficient but perfectly serviceable @reqdlink{get-map} call
instead@linkhere{}.


@node Fundamentals
@chapter Fundamentals of Protocol A

The data types in protocol A come in two flavors. The first (vanilla)
are the simple data types from which the LysKOM (chocolate) data types
are built. Simple data types include things like integers and strings
while complex data types include things such as conferences and people.

@menu
* Notation::                    
* Simple Data Types::           
* Client-Server Dialog::        
@end menu

@node Notation
@section Notation

This specification uses a BNF-like grammar to describe the protocol and
its data elements.

Data fields have been given names that start with a lower-case letter.

Fundamental data types have names in all-caps (such as @type{INT32} and
@type{ARRAY}).

Derived data types have names that start with an upper-case letter. (If
the type contains more than one word, all words start with an upper-case
letter, like this: @type{Text-Stat}.)  The operator @code{::=} defines
the name to its left.

Comments start with @code{!} (exclamation mark) and alternatives are
separated by a @code{|} (vertical bar.) A @code{;} (semicolon)
terminates statements in the grammar. In some specifications there are
literal strings. There is to be no whitespace before or after literal
strings unless there is whitespace in the literal itself.


@node Simple Data Types
@section Simple Data Types

@anchor{INT32}
@anchor{INT16}
@anchor{INT8}
@anchor{BOOL}
@subsection Integers

@tindex INT32
@tindex INT16
@tindex INT8
@tindex BOOL
@type{INT32}, @type{INT16}, @type{INT8} and @type{BOOL} are non-negative
integers which must fit in 32, 16, 8 and 1 bits, respectively. They are
transmitted to the server in ASCII-encoded decimal notation.


@anchor{HOLLERITH}
@subsection Strings

@tindex HOLLERITH
@type{HOLLERITH} denotes character strings of arbitrary length. They are
transmitted as @code{@var{n}H@var{text}} where @var{text} is the string and
@var{n} is the number of characters in @var{text} in decimal
notation. All byte values are allowed in the string itself, including
nulls.

Long live FORTRAN!



@anchor{BITSTRING}
@subsection Bit Strings

@tindex BITSTRING
@type{BITSTRING} is a string of bits, commonly used for a set of
boolean-valued flags. Bit strings are denoted as

@example
        BITSTRING ( name-1; name-2; name-3; @dots{} )
@end example

in this specification. They are transmitted as a sequence of ASCII ones
and zeroes in the order the fields are listed.

For instance, given the specification

@example
        shape-of-world ::= BITSTRING (
                is-flat;
                is-round;
                is-2d;
                is-3d;
        )
@end example

most peoples idea of @code{shape-of-world} would be sent as @code{0101}
(round and three-dimensional.)


@anchor{ENUMERATION}
@subsection Enumerations

@tindex ENUMERATION
@type{ENUMERATION} is an integer constant. It is transmitted as an INT32, 
but only fixed values are permitted. Clients should be prepared to
receive numbers outside the enumeration and either handle this
gracefully as an error or use a reasonable default value in place of an
invalid enumeration value.

An enumeration is specified as 

@example
        ENUMERATION ( 
                name-1=value-1;
                name-2=value-2;
                name-3=value-3;
                @dots{}
        )
@end example

This specification states that name-1 is represented by the integer
value-1, name-2 is represented by value-2 and name-3 is represented by
value-3. 

An enumeration can also be inherited from a SELECTION datatype:

@example
        Info-type ::= ENUMERATION-OF(Misc-Info)
@end example

This means that Info-type is an enumeration, that contains the same keys
and values as the SELECTION Misc-Info.

For example, in the following specification, the constant guwal will be
transmitted as the integer 2, ciokwe as the integer 3, and hopi as the
integer 5.

@example
      language ::= ENUMERATION ( hakka      = 1;
                                 guwal      = 2;
                                 ciokwe     = 3;
                                 yoruba     = 4;
                                 hopi       = 5;
                )
@end example




@anchor{ARRAY}
@subsection Arrays

@tindex ARRAY
@type{ARRAY} is a list of a fixed number of elements of a single type.
The specification for an array is @code{ARRAY @var{type}} where
@var{type} is the type of the array elements.

Arrays are transmitted as an @code{@var{n} @{ @var{element}
@var{element} @dots{} @}} where @var{n} is the number of elements and
each @var{element} is an element of the array. A special case is when
the array is empty, in which case the server transmits it as @code{0 *}. 
Note that the client must always transmit empty arrays as @code{0 @{
@}}.

In some calls the client can ask the server not to send the array
contents, only its length. In these cases the array is transmitted as
@code{@var{n} *} where @var{n} is the number of elements in the array.


@anchor{SELECTION}
@subsection Selection

@tindex SELECTION
@type{SELECTION} is tagged data. It consists of an INT32 selector
followed by a tail of an arbitrary type and is specified as

@example
        SELECTION (
                @var{n}=@var{name}        @var{tail} : @var{type};
                @var{n}=@var{name}        @var{tail} : @var{type};
                @dots{}
        )
@end example

where each @var{n} is the selector value, @var{name} the selector
name and @var{tail} the name of the selector tail and @var{type}
its type.  @var{name} and @var{tail} are often very similar, such
as @samp{sent-by} and @samp{sender}.

When transmitted, the selector is transmitted as an INT32 followed by
the tail belonging to that selector. For instance, given the
specification

@example
        description ::= SELECTION (
               1=name           the_name : HOLLERITH;
               2=age            years    : INT32;
        )
@end example

two legal messages of the type @code{description} are @samp{1 4HJohn} and
@samp{2 18}.


@anchor{RPC}
@subsection RPC

@tindex RPC
@type{RPC} is a notation used to specify calls to the server. An RPC
specification has the following form:

@example
        RPC (
                @var{call} [@var{n}] ( @var{request} ) -> ( @var{reply} ) ;
                @var{call} [@var{n}] ( @var{request} ) -> ( @var{reply} ) ;
        )
@end example

where each @var{call} is the name of a call, @var{n} is the call
number, @var{request} is a single data element sent as a request and
@var{reply} is a single data element sent in reply from the server.

RPC calls are transmitted as @var{n} @var{request} where @var{n} and
@var{request} have the same meaning as above. Note that in the
client-server dialog a reference number must also be supplied with each
request. This reference number is not part of the RPC itself, but is
required for communications; see @ref{Client-Server Dialog}.



@subsection Structure

Structures are collections of several data items. In the specification
they are written as

@example
        ( @var{name} : @var{type} ;
          @var{name} : @var{type} ;
          @dots{}
        )
@end example

where each @var{name} is the name of a data field and the
corresponding @var{type} is its type.