A system identifier can either be a formal system identifier or a simple system identifier. A system identifier that is a formal system identifier consists of a sequence of one or more storage object specifications. The objects specified by the storage object specifications are concatenated to form the entity. A storage object specification consists of an SGML start-tag in the reference concrete syntax followed by character data content. The generic identifier of the start-tag is the name of a storage manager. The content is a storage object identifier which identifies the storage object in a manner dependent on the storage manager. The start-tag can also specify attributes giving additional information about the storage object. Numeric character references are recognized in storage object identifiers and attribute value literals in the start-tag. Record ends are ignored in the storage object identifier as with SGML. A system identifier will be interpreted as a formal system identifier if it starts with a < followed by a storage manager name, followed by either > or white-space; otherwise it will be interpreted as a simple system identifier. A storage object identifier extends until the end of the system identifier or until the first occurrence of < followed by a storage manager name, followed by either > or white-space.

The following storage managers are available:

osfile

The storage object identifier is a filename. If the filename is relative it is resolved using a base filename. Normally the base filename is the name of the file in which the storage object identifier was specified, but this can be changed using the base attribute. The filename will be searched for first in the directory of the base filename. If it is not found there, then it will be searched for in directories specified with the -D option in the order in which they were specified on the command line, and then in the list of directories specified by the environment variable SGML_SEARCH_PATH. The list is separated by colons under Unix and by semi-colons under MSDOS.

osfd

The storage object identifier is an integer specifying a file descriptor. Thus a system identifier of <osfd>0 will refer to the standard input.

url

The storage object identifier is a URL. Only the http scheme is currently supported and not on all systems.

neutral

The storage manager is the storage manager of storage object in which the system identifier was specified (the underlying storage manager). However if the underlying storage manager does not support named storage objects (ie it is osfd), then the storage manager will be osfile. The storage object identifier is treated as a relative, hierarchical name separated by slashes (/) and will be transformed as appropriate for the underlying storage manager.

The following attributes are supported:

records

This describes how records are delimited in the storage object:

cr

Records are terminated by a carriage return.

lf

Records are terminated by a line feed.

crlf

Records are terminated by a carriage return followed by a line feed.

find

Records are terminated by whichever of cr, lf or crlf is first encountered in the storage object.

asis

No recognition of records is performed.

The default is find except for NDATA entities for which the default is asis.

When records are recognized in a storage object, a record start is inserted at the beginning of each record, and a record end at the end of each record. If there is a partial record (a record that doesn't end with the record terminator) at the end of the entity, then a record start will be inserted before it but no record end will be inserted after it.

The attribute name and = can be omitted for this attribute.

zapeof

This specifies whether a Control-Z character that occurs as the final byte in the storage object should be stripped. The following values are allowed:

zapeof

A final Control-Z should be stripped.

nozapeof

A final Control-Z should not be stripped.

The default is zapeof except for NDATA entities, entities declared in storage objects with zapeof=nozapeof and storage objects with records=asis.

The attribute name and = can be omitted for this attribute.

bctf

The bctf (bit combination transformation format) attribute describes how the bit combinations of the storage object are transformed into the sequence of bytes that are contained in the object identified by the storage object identifier. This inverse of this transformation is performed when the entity manager reads the storage object. It has one of the following values:

identity

Each bit combination is represented by a single byte.

fixed-2

Each bit combination is represented by exactly 2 bytes, with the more significant byte first.

utf-8

Each bit combination is represented by a variable number of bytes according to UCS Transformation Format 8 defined in Annex P to be added by the first proposed drafted amendment (PDAM 1) to ISO/IEC 10646-1:1993.

euc-jp

Each bit combination is treated as a pair of bytes, most significant byte first, encoding a character using the Extended_UNIX_Code_Fixed_Width_for_Japanese Internet charset, and is transformed into the variable length sequence of octets that would encode that character using the Extended_UNIX_Code_Packed_Format_for_Japanese Internet charset.

sjis

Each bit combination is treated as a pair of bytes, most significant byte first, encoding a character using the Extended_UNIX_Code_Fixed_Width_for_Japanese Internet charset, and is transformed into the variable length sequence of bytes that would encode that character using the Shift_JIS Internet charset.

unicode

Each bit combination is represented by 2 bytes. The bytes representing the entire storage object may be preceded by a pair of bytes representing the byte order mark character (0xFEFF). The bytes representing each bit combination are in the system byte order, unless the byte order mark character is present, in which case the order of its bytes determines the byte order. When the storage object is read, any byte order mark character is discarded.

is8859-N

N can be any single digit other than 0. Each bit combination is interpreted as the number of a character in ISO/IEC 10646 and is represented by the single byte that would encode that character in ISO 8859-N. These values are not supported with the -b option.

Values other than identity are supported only with the multi-byte version of nsgmls.

tracking

This specifies whether line boundaries should be tracked for this object: a value of track specifies that they should; a value of notrack specifies that they should not. The default value is track. Keeping track of where line boundaries occur in a storage object requires approximately one byte of storage per line and it may be desirable to disable this for very large storage objects.

The attribute name and = can be omitted for this attribute.

base

When the storage object identifier specified in the content of the storage object specification is relative, this specifies the base storage object identifier relative to which that storage object identifier should be resolved. When not specified a storage object identifier is interpreted relative to the storage object in which it is specified, provided that this has the same storage manager. This applies both to system identifiers specified in SGML documents and to system identifiers specified in the catalog entry files.

smcrd

The value is a single character that will be recognized in storage object identifiers (both in the content of storage object specifications and in the value of base attributes) as a storage manager character reference delimiter when followed by a digit. A storage manager character reference is like an SGML numeric character reference except that the number is interpreted as a character number in the inherent character set of the storage manager rather than the document character set. The default is for no character to be recognized as a storage manager character reference delimiter. Numeric character references cannot be used to prevent recognition of storage manager character reference delimiters.

fold

This applies only to the neutral storage manager. It specifies whether the storage object identifier should be folded to the customary case of the underlying storage manager if storage object identifiers for the underlying storage manager are case sensitive. The following values are allowed:

fold

The storage object identifier will be folded.

nofold

The storage object identifier will not be folded.

The default value is fold. The attribute name and = can be omitted for this attribute.

For example, on Unix filenames are case-sensitive and the customary case is lower-case. So if the underlying storage manager were osfile and the system was a Unix system, then %<neutral>FOO.SGM would be equivalent to %<osfile>foo.sgm.

A simple system identifier is interpreted as a storage object identifier with a storage manager that depends on where the system identifier was specified: if it was specified in a storage object whose storage manager was url or if the system identifier looks like an absolute URL in a supported scheme, the storage manager will be url; otherwise the storage manager will be osfile. The storage manager attributes are defaulted as for a formal system identifier. Numeric character references are not recognized in simple system identifiers.

The entity manager generates an effective system identifier for every external entity using catalog entry files in the format defined by SGML Open Technical Resolution 9401:1994. The entity manager will give an error if it is unable to generate an effective system identifier for an external entity. Normally if the external identifier for an entity includes a system identifier then the entity manager will use that as the effective system identifier for the entity; this behaviour can be changed using OVERRIDE or SYSTEM entries in a catalog entry file.

A catalog entry file contains a sequence of entries in one of the following forms:

PUBLIC pubid sysid

This specifies that sysid should be used as the effective system identifier if the public identifier is pubid. Sysid is a system identifier as defined in ISO 8879 and pubid is a public identifier as defined in ISO 8879.

ENTITY name sysid

This specifies that sysid should be used as the effective system identifier if the entity is a general entity whose name is name.

ENTITY %name sysid

This specifies that sysid should be used as the effective system identifier if the entity is a parameter entity whose name is name. Note that there is no space between the % and the name.

DOCTYPE name sysid

This specifies that sysid should be used as the effective system identifier if the entity is an entity declared in a document type declaration whose document type name is name.

LINKTYPE name sysid

This specifies that sysid should be used as the effective system identifier if the entity is an entity declared in a link type declaration whose link type name is name.

NOTATION name sysid

This specifies that sysid should be used as the effective system identifier for a notation whose name is name. This is an extension to the SGML Open format. This is relevant only with the -n option.

OVERRIDE YES|NO

This sets the overriding mode for entries up to the next occurrence of OVERRIDE or the end of the catalog entry file. At the beginning of a catalog entry file the overriding mode will be NO. A PUBLIC, ENTITY, DOCTYPE, LINKTYPE or NOTATION entry with an overriding mode of YES will be used whether or not the external identifier has an explicit system identifier; those with an overriding mode of NO will be ignored if external identifier has an explicit system identifier. This is an extension to the SGML Open format.

SYSTEM sysid1 sysid2

This specifies that sysid2 should be used as the effective system identifier if the system identifier specified in the external identifier was sysid1. This is an extension to the SGML Open format.

SGMLDECL sysid

This specifies that if the document does not contain an SGML declaration, the SGML declaration in sysid should be implied.

DOCUMENT sysid

This specifies that the document entity is sysid. This entry is used only with the -C option.

CATALOG sysid

This specifies that sysid is the system identifier of an additional catalog entry file to be read after this one. Multiple CATALOG entries are allowed and will be read in order. This is an extension to the SGML Open format.

The delimiters can be omitted from the sysid provided it does not contain any white space. Comments are allowed between parameters delimited by -- as in SGML.

The environment variable SGML_CATALOG_FILES contains a list of catalog entry files. The list is separated by colons under Unix and by semi-colons under MSDOS. These will be searched after any catalog entry files specified using the -m option, and after the catalog entry file called catalog in the same place as the document entity. If this environment variable is not set, then a system dependent list of catalog entry files will be used. In fact catalog entry files are not restricted to being files: the name of a catalog entry file is interpreted as a system identifier.

A match in one catalog entry file will take precedence over any match in a later catalog entry file. A match in a catalog entry file for a SYSTEM entry will take precedence over a match in the same file for a PUBLIC, ENTITY, DOCTYPE, LINKTYPE or NOTATION entry. A match in a catalog entry file for a PUBLIC entry will take precedence over a match in the same file for an ENTITY, DOCTYPE, LINKTYPE or NOTATION entry.

The system declaration for nsgmls is as follows:

tab(&); c1 s1 s1 s1 s1 s1 s1 s1 s c s s s s s s s s l l s s s s s s s l l s s s s s s s l l s s s s s s s l l l s s s s s s c s s s s s s s s l l l l l l l l l l l l l l l l l l l l l l l l l l l l l s s s s s s s l l l s s s s s s l l l s s s s s s c s s s s s s s s l l l l l l l l l. SYSTEM "ISO 8879:1986" CHARSET BASESET&"ISO 646-1983//CHARSET &h'w'"'u'International Reference Version (IRV)//ESC 2/5 4/0" DESCSET&0012800 CAPACITY&PUBLIC&"ISO 8879:1986//CAPACITY Reference//EN" FEATURES MINIMIZE&DATATAG&NO&OMITTAG&YES&RANK&YES&SHORTTAG&YES LINK&SIMPLE&YES 65535&IMPLICIT&YES&EXPLICIT&YES 1 OTHER&CONCUR&NO&SUBDOC&YES 100&FORMAL&YES SCOPE&DOCUMENT SYNTAX&PUBLIC&"ISO 8879:1986//SYNTAX Reference//EN" SYNTAX&PUBLIC&"ISO 8879:1986//SYNTAX Core//EN" VALIDATE &GENERAL&YES&MODEL&YES&EXCLUDE&YES&CAPACITY&NO &NONSGML&YES&SGML&YES&FORMAL&YES c s s s s s s s s l l l l l l l l l. SDIF &PACK&NO&UNPACK&NO

The limit for the SUBDOC parameter is memory dependent.

Any legal concrete syntax may be used.

If the S declaration is omitted and there is no applicable SGMLDECL entry in a catalog, the following declaration will be implied: tab(&); c1 s1 s1 s1 s1 s1 s1 s1 s c s s s s s s s s l l s s s s s s s. <!SGML "ISO 8879:1986" CHARSET BASESET&"ISO 646-1983//CHARSET &h'w'"'u'International Reference Version (IRV)//ESC 2/5 4/0" DESCSET&0000090UNUSED &009002009 &0110020UNUSED &013001013 &0140180UNUSED &032095032 &1270010UNUSED l l l s s s s s s l l s s s s s s s. CAPACITY&PUBLIC&"ISO 8879:1986//CAPACITY Reference//EN" SCOPE&DOCUMENT SYNTAX SHUNCHAR&CONTROLS 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 &18 19 20 21 22 23 24 25 26 27 28 29 30 31 127 255 BASESET&"ISO 646-1983//CHARSET International Reference Version &h'w'"'u'(IRV)//ESC 2/5 4/0" DESCSET&0012800 l l l l s s s s s. FUNCTION&RE&&13 &RS&&10 &SPACE&&32 &TAB&SEPCHAR&9 NAMING&LCNMSTRT&"" &UCNMSTRT&"" &LCNMCHAR&"-." &UCNMCHAR&"-." &NAMECASE&GENERAL&YES &&ENTITY&NO DELIM&GENERAL&SGMLREF &SHORTREF&SGMLREF NAMES&SGMLREF QUANTITY&SGMLREF &ATTCNT&99999999 &ATTSPLEN&99999999 &DTEMPLEN&24000 &ENTLVL&99999999 &GRPCNT&99999999 &GRPGTCNT&99999999 &GRPLVL&99999999 &LITLEN&24000 &NAMELEN&99999999 &PILEN&24000 &TAGLEN&99999999 &TAGLVL&99999999 c s s s s s s s s l l l s s s s s s. FEATURES MINIMIZE&DATATAG&NO &OMITTAG&YES &RANK&YES &SHORTTAG&YES LINK&SIMPLE&YES01000 &IMPLICIT&YES &EXPLICIT&YES01 OTHER&CONCUR&NO &SUBDOC&YES099999999 &FORMAL&YES c s s s s s s s s. APPINFO NONE> with the exception that all characters that are neither significant not shunned will be assigned to DATACHAR.

A character in a base character set is described either by giving its number in a universal character set, or by specifying a minimum literal. The constraints on the choice of universal character set are that characters that are significant in the SGML reference concrete syntax must be in the universal character set and must have the same number in the universal character set as in ISO 646 and that each character in the character set must be represented by exactly one number; that character numbers in the range 0 to 31 and 127 to 159 are control characters (for the purpose of enforcing SHUNCHAR CONTROLS). It is recommended that ISO 10646 (Unicode) be used as the universal character set, except in environments where the normal document character sets are large character set which cannot be compactly described in terms of ISO 10646. The public identifier of a base character set can be associated with an entity that describes it by using a PUBLIC entry in the catalog entry file. The entity must be a fragment of an SGML declaration consisting of the portion of a character set description, following the DESCSET keyword, that is, it must be a sequence of character descriptions, where each character description specifies a described character number, the number of characters and either a character number in the universal character set, a minimum literal or the keyword UNUSED. Character numbers in the universal character set can be as big as 99999999.

In addition nsgmls has built in knowledge of a few character sets. These are identified using the designating sequence in the public identifier. The following designating sequences are recognized: tab(&); c c c c c c c c c ^ c c c c ^ l n n n l. Designating&ISO&Minimum&Number&Description Escape&Registration&Character&of& Sequence&Number&Number&Characters& _ ESC 2/5 4/0&-&0&128&full set of ISO 646 IRV ESC 2/8 4/0&2&0&128&G0 set of ISO 646 IRV ESC 2/8 4/2&6&0&128&G0 set of ASCII ESC 2/1 4/0&1&0&32&C0 set of ISO 646

The graphic character sets do not strictly include C0 and C1 control character sets. For convenience, nsgmls augments the graphic character sets with the appropriate control character sets.

It is not necessary for every character set used in the SGML declaration to be known to nsgmls provided that characters in the document character set that are significant both in the reference concrete syntax and in the described concrete syntax are described using known base character sets and that characters that are significant in the described concrete syntax are described using the same base character sets or the same minimum literals in both the document character set description and the syntax reference character set description.

The public identifier for a public concrete syntax can be associated with an entity that describes using a PUBLIC entry in the catalog entry file. The entity must be a fragment of an SGML declaration consisting of a concrete syntax description starting with the SHUNCHAR keyword as in an SGML declaration. The entity can also make use of the following extensions:

An added function can be expressed as a parameter literal instead of a name.

The replacement for a reference reserved name can be expressed as a parameter literal instead of a name.

The LCNMSTRT, UCNMSTRT, LCNMCHAR and UCNMCHAR keywords may each be followed by more than one parameter literal. A sequence of parameter literals has the same meaning as a single parameter literal whose content is the concatenation of the content of each of the literals in the sequence. This extension is useful because of the restriction on the length of a parameter literal in the SGML declaration to 240 characters.

The total number of characters specified for UCNMCHAR or UCNMSTRT may exceed the total number of characters specified for LCNMCHAR or LCNMSTRT respectively. Each character in UCNMCHAR or UCNMSTRT which does not have a corresponding character in the same position in LCNMCHAR or LCNMSTRT is simply assigned to UCNMCHAR or UCNMSTRT without making it the upper-case form of any character.

A parameter following any of LCNMSTRT, UCNMSTRT, LCNMCHAR and UCNMCHAR keywords may be followed by the name token ... and another parameter literal. This has the same meaning as the two parameter literals with a parameter literal in between containing in order each character whose number is greater than the number of the last character in the first parameter literal and less than the number of the first character in the second parameter literal. A parameter literal must contain at least one character for each ... to which it is adjacent.

A number may be used as a parameter following the LCNMSTRT, UCNMSTRT, LCNMCHAR and UCNMCHAR keywords or as a delimiter in the DELIM section with the same meaning as a parameter literal containing just a numeric character reference with that number.

The parameters following the LCNMSTRT, UCNMSTRT, LCNMCHAR and UCNMCHAR keywords may be omitted. This has the same meaning as specifying an empty parameter literal.

Within the specification of the short reference delimiters, a parameter literal containing exactly one character may be followed by the name token ... and another parameter literal containing exactly one character. This has the same meaning as a sequence of parameter literals one for each character number that is greater than or equal to the number of the character in the first parameter literal and less than or equal to the number of the character in the second parameter literal.

The public identifier for a public capacity set can be associated with an entity that describes using a PUBLIC entry in the catalog entry file. The entity must be a fragment of an SGML declaration consisting of a sequence of capacity names and numbers.

The output is a series of lines. Lines can be arbitrarily long. Each line consists of an initial command character and one or more arguments. Arguments are separated by a single space, but when a command takes a fixed number of arguments the last argument can contain spaces. There is no space between the command character and the first argument. Arguments can contain the following escape sequences.

\\

A \.

\n

A record end character.

\|

Internal SDATA entities are bracketed by these.

\nnn

The character whose code is nnn octal.

A record start character will be represented by \012. Most applications will need to ignore \012 and translate \n into newline.

\#n;

The character whose number is n in decimal. n can have any number of digits. This is used for characters that are not representable by the encoding translation used for output (as specified by the NSGML_CODE environment variable). This will only occur with the multibyte version of nsgmls.

The possible command characters and arguments are as follows:

(gi

The start of an element whose generic identifier is gi. Any attributes for this element will have been specified with A commands.

)gi

The end of an element whose generic identifier is gi.

-data

Data.

&name

A reference to an external data entity name; name will have been defined using an E command.

?pi

A processing instruction with data pi.

Aname val

The next element to start has an attribute name with value val which takes one of the following forms:

IMPLIED

The value of the attribute is implied.

CDATA data

The attribute is character data. This is used for attributes whose declared value is CDATA.

NOTATION nname

The attribute is a notation name; nname will have been defined using a N command. This is used for attributes whose declared value is NOTATION.

ENTITY name...

The attribute is a list of general entity names. Each entity name will have been defined using an I, E or S command. This is used for attributes whose declared value is ENTITY or ENTITIES.

TOKEN token...

The attribute is a list of tokens. This is used for attributes whose declared value is anything else.

ID token

The attribute is an ID value. This will be output only if the -oid option is specified. Otherwise TOKEN will be used for ID values.

Dename name val

This is the same as the A command, except that it specifies a data attribute for an external entity named ename. Any D commands will come after the E command that defines the entity to which they apply, but before any & or A commands that reference the entity.

atype name val

The next element to start has a link attribute with link type type, name name, and value val, which takes the same form as with the A command.

Nnname

nname. Define a notation. This command will be preceded by a p command if the notation was declared with a public identifier, and by a s command if the notation was declared with a system identifier. If the -n option was specified, this command will also be preceded by an f command giving the system identifier generated by the entity manager (unless it was unable to generate one). A notation will only be defined if it is to be referenced in an E command or in an A command for an attribute with a declared value of NOTATION.

Eename typ nname

Define an external data entity named ename with type typ (CDATA, NDATA or SDATA) and notation not. This command will be preceded by an f command giving the system identifier generated by the entity manager (unless it was unable to generate one), by a p command if a public identifier was declared for the entity, and by a s command if a system identifier was declared for the entity. not will have been defined using a N command. Data attributes may be specified for the entity using D commands. If the -oentity option is not specified, an external data entity will only be defined if it is to be referenced in a & command or in an A command for an attribute whose declared value is ENTITY or ENTITIES.

Iename typ text

Define an internal data entity named ename with type typ and entity text text. The typ will be CDATA or SDATA unless the -oentity option was specified, in which case it can also be PI or TEXT (for an S text entity). If the -oentity option is not specified, an internal data entity will only be defined if it is referenced in an A command for an attribute whose declared value is ENTITY or ENTITIES.

Sename

Define a subdocument entity named ename. This command will be preceded by an f command giving the system identifier generated by the entity manager (unless it was unable to generate one), by a p command if a public identifier was declared for the entity, and by a s command if a system identifier was declared for the entity. If the -oentity option is not specified, a subdocument entity will only be defined if it is referenced in a { command or in an A command for an attribute whose declared value is ENTITY or ENTITIES.

Tename

Define an external SGML text entity named ename. This command will be preceded by an f command giving the system identifier generated by the entity manager (unless it was unable to generate one), by a p command if a public identifier was declared for the entity, and by a s command if a system identifier was declared for the entity. This command will be output only if the -oentity option is specified.

ssysid

This command applies to the next E, S, T or N command and specifies the associated system identifier.

ppubid

This command applies to the next E, S, T or N command and specifies the associated public identifier.

fsysid

This command applies to the next E, S, T or, if the -n option was specified, N command and specifies the system identifier generated by the entity manager from the specified external identifier and other information about the entity or notation.

{ename

The start of the S subdocument entity ename; ename will have been defined using a S command.

}ename

The end of the S subdocument entity ename.

Llineno file

Llineno Set the current line number and filename. The file argument will be omitted if only the line number has changed. This will be output only if the -l option has been given.

#text

An APPINFO parameter of text was specified in the S declaration. This is not strictly part of the ESIS, but a structure-controlled application is permitted to act on it. No # command will be output if APPINFO NONE was specified. A # command will occur at most once, and may be preceded only by a single L command.

C

This command indicates that the document was a conforming S document. If this command is output, it will be the last command. An S document is not conforming if it references a subdocument entity that is not conforming.