The text dictionary consists of the following parts:

• head: contains the dictionary identifier;
• tags: elements of the body;
• restrictions: additional rules for constructing the document..

Here is an example of the file with several text dictionaries:

USDS DICTIONARY ID=888 v.1.0
{
    1: STRUCT I
    {
        1: UNSIGNED VARINT n;
        2: DOUBLE s;
        3: STRING<UTF-8> g;
        4: LONG t;
        5: BOOLEAN b;
    } RESTRICT {root=false;}

    2: STRUCT S
    {
        1: UNSIGNED VARINT n;
        2: INT m;
        3: LONG s;
        4: LONG e;
        5: ARRAY<I> v;
    }
} 

USDS DICTIONARY ID=888 v.1.1
{
    1: STRUCT I
    {
        1: UNSIGNED VARINT n;
        2: DOUBLE s;
        3: STRING<UTF-8> g;
        4: LONG t;
        5: BOOLEAN b;
        6: BOOLEAN с;
    } RESTRICT {root=false;}

    2: STRUCT S
    {
        1: UNSIGNED VARINT n;
        2: INT m;
        3: LONG s;
        4: LONG e;
        5: ARRAY<I> v;
    }
} 

Head

The head format:

The scheme includes:

• Dictionary identifier: integer, from 0 to 2147483647. The values from 1 to 65535 are reserved for public standards (HTML, SCT, etc.);
• Major version: integer, from 0 to 255;
• Minor version: integer, from 0 to 255;

The Dictionary identifier is used for integrity control when a dictionary and a body are storing separately. If software is compiled using the dictionary, only the head and body are necessary in USDS documents. If you are always storing a dictionary and body together (e.g. for a configuration file) then you can select zero as the dictionary identifier. It's recommended to use the dictionary versions anyway.

Tags

The tag format:

The scheme includes:

• Tag identifier:
• an integer, from 1 to 2147483647;
• unique in one dictionary version;
• it is necessary to number tags sequentially;
  
• in the binary body the type "unsigned varint" is used;
• Tag type: all USDS types represented below;
  
• Tag name: not less than 1 symbol; can contain Latin letters, numbers and the symbol '_'. The first symbol must be a non-number. The binary body doesn't contain the tag names;
  
• "RESTRICT" - addition rules for tags:
  
• "notRoot": identifier for non-root tags. For example, the tag <H1> cannot be a root tag in HTML.

Data Types

Simple tags

The format of the simple tags (from 1 to 25 in the table above):

Signed integer numbers are represented as "Two's complement" in the binary body. Real number corresponds to IEEE 754.

Big-endian numbers are parsed correctly on Little-endian platforms.

The data types "VARINT" and "UNSIGNED VARINT" correspond to Little Endian Base 128.

The following restrictions are enabled for these types, instead of "BOOLEAN":

• "values": valid values for the tag (ranges or list of values);
• "precision": it's available for integer and float types.
  

Structure

The scheme includes:

• Field identifier:
• an integer, from 1 to 2147483647;
• unique in 1 structure;
• it is necessary to number fields sequentially;
  
• in the binary body the type "unsigned varint" is used;
• Field type:
  
• any USDS type instead of "struct";
• other tag name with any type;
• this tag name (recursion is supported): in this case the field must be optional;
• Field name: minimum of 1 symbol; can contain Latin letters, numbers and the symbol '_'. The first symbol must be a non-number. The binary body doesn't contain the field names;
• The next restrictions are enabled:
• all restrictions of the field's type;
• "default": the default value for the field, only for the simple types (1-25) and strings. If a field's value is equal to the default value, it won't be included in the binary body;
• "optional": for any type. Parsers will return NULL if the field was not included in the binary body.

Array

The format of the array:

The scheme includes:

• Element's type:
• the simple types (1-25);
• "STRING";
• other tag name with any type;
• list of the tag names with any type: in this case the array can include all these tags (polymorph array);
• the next restrictions are available:
  
• all restrictions of the element's type;
• "fixSize": fixed number of the elements in the array. The value starts at 1:
  
• "minSize" - the minimum size of the array. The value starts at 0. It's incompatible with "fixSize";
• "maxSize" - the maximum size of the array. The value starts at 1. It's incompatible with "fixSize";
• "minOccure" - for the polymorph array only: minimum number of the elements in array. The value starts at 0;
• "maxOccure" - for the polymorph array only: maximum number of the elements in array. The value starts at 1.

Strings

The format of the string:

The scheme includes:

• Text encode: one of the supported text encodes; the full list is represented below;
• Restrictions:
• "fixSize": fixed number of the symbols in the string. The value starts at 1:
  
• "minSize" - the minimum number of the symbols in the string. The value starts at 0. It's incompatible with "fixSize";
• "maxSize" - the maximum number of the symbols in the string. The value starts at 1. It's incompatible with "fixSize";
• "values": the regular expression for the string.

Lists

The format of the list:

The scheme includes:

• Key type:
• simple types (1-25);
• "STRING";
• other tag name with any type;
• list of the tag names with any type: in this case the key can be any of these tags (polymorph map);;
• Value type:
• simple types (1-25);
• "STRING";
• other tag name with any type;
• list of the tag names with any type: in this case the value can be any of these tags (polymorph map);
• the following restrictions are available:
  
• all restrictions of the element's type;
• "minSize" - the minimum size of the map. The value starts at 0;
• "maxSize" - the maximum size of the map. The value starts at 1;
• "minOccure" - for the polymorph map only: minimum number of the elements in list. The value starts at 0;
• "maxOccure" - for the polymorph map only: maximum number of the elements in list. The value starts at 1.

Polymorph Tags

1: STRUCT H1 { ... }
2: STRUCT H2 { ... }
3: POLYMORPH<H1, H2> BodySubTags;
4: STRUCT Body
{
    1: LIST<BodySubTags> subTags;
}

Other formats use type "ANY" for polymorphism, but this method reduces the accuracy of the schema description.

The format of the polymorph:

The scheme includes:

• Tag names - the list of the tags, which are part of the polymorph;
• only one restriction is supported:
• notRoot: identifier for non-root tags.

Dictionary restrictions

• "minOccure" - minimum number of the elements in the USDS Documents. The value starts at 0;
• "maxOccure" - maximum number of the elements in the USDS Documents. The value starts at 1.