CBOR-XS/doc/stringref.pod

=head1 REGISTRATION INFORMATION

   Tag             256 (stringref-namespace)
   Data Item       multiple
   Semantics       mark value as having string references
   Reference       http://cbor.schmorp.de/stringref
   Contact         Marc A. Lehmann <cbor@schmorp.de>

   Tag             25 (stringref)
   Data Item       unsigned integer
   Semantics       reference the nth previously seen string
   Reference       http://cbor.schmorp.de/stringref
   Contact         Marc A. Lehmann <cbor@schmorp.de>

=head1 RATIONALE

These two tags can be used to efficiently share constant strings, which
are common in many data structures. They are an optimisation only,
similar to LZSS compresssion, for the string major types in CBOR.

Background: many data structures or protocols contain repeated string
constants (often called atoms or quarks). For example, this is a
real-world (JSON-formatted) data structure that contains many repeated
strings used as map keys:

  [
     {
       "name" : "Cocktail",
       "count" : 417,
       "rank" : 4
     },
     {
       "rank" : 4,
       "count" : 312,
       "name" : "Bath"
     },
     {
       "count" : 691,
       "name" : "Food",
       "rank" : 4
     }
  ]

This can be encoded nicely with CBOR, but each occurrence of "name",
"count" and "rank" will be encoded separately.  In highly repetitive data
structures (the above is taken from a game save file) these can take up
considerable space.

This scheme can be used to reduce this overhead with a simple scheme that
is easily implementable.

=head1 DESCRIPTION

Stringref consists of two tags, stringref-namespace (value C<256>),
which marks a value as containing string references, and stringref (value
C<25>), which references a string previously encoded in the value.

The stringref-namespace tag is used to define a namespace for the string
reference ids. stringref tags are only valid inside CBOR values marked
with stringref-namespace.

The purpose of stringref-namespace is three-fold:

=over 4

=item 1. It signals the decoder that it needs to keep a copy of every
string that is being decoded while decoding the tagged value (and
conversely, that there is no need for this extra overhead without this
tag).

=item 2. It can be used to confine string references to certain subtrees,
which might increase encoding efficiency by being able to use smaller
references.

=item 3. CBOR objects encoded with this scheme can be blindly embedded in
existing CBOR streams, as the indices are relative to the outer namespace
tag, and don't change when embedded.

=back

Within a value tagged with stringref-namespace, every string that is
encoded with a definite length and has a minimum length is assigned an
implicit index, starting from zero.

The minimum length required to get an index depends on the index value
that would be assigned to it, as stringref references should be shorter
than the string they reference, and there is no point in wasting indices
on strings that should not be referenced anyway.

The lengths are as follows:

   string index              minimum string length in octets
   0 .. 23                   3
   24 .. 255                 4
   256 .. 65535              5
   65536 .. 4294967295       7
   4294967296 ..             11

The minimum string length is simply the length of the stringref tag (2
octets), plus the minimum size the resulting index takes up, using major
type 0 encoding.

Each time a string is to be encoded by the encoder, it can choose to
encode it as a normal string, or attempt to encode it as a stringref.

If it attempts to encode a stringref, it needs to check whether the string
has been assigned an index already. If yes, it can emit a stringref tag
with the index.

In all other cases the encoder must check whether it has to assign an
index value to it by checking against the minimum length required for the
next index value to be assigned, and assigning the next index value if the
string has minimum length or is longer.

A typical encoder might want to look up every string (using the major
type and the octets) it encodes in some kind of internal map that stores
previously-assigned indices. If found, it will emit a stringref. If not,
it attempts to assign an index to the string and stores that in its map.

Decoders might choose not to look up previous indices for a string, for
example, because it only wants to compress strings used as map keys, or
because it intrinsically knows that the string is only used once. This is
fine, but each time the encoder emits a string, it MUST assign an index if
the minimum length is reached or exceeded, even if it doesn't store that
index for later use.

Byte strings and text strings share the same index namespace, but are
distinct, even if their contents are bit-identical. Indefinite length
strings and the chunks they consist of are never assigned an index.

A typical encoder would emit the stringref-namespace tag once to tag
the top-level value in a CBOR stream, but nothing keeps an encoder
from nesting stringref-namespace tags. Nesting might be done for
convenience, to improve coding efficiency, to embed values or for any
other reason. Similarly, an encoder can sort map keys for greater coding
efficiency, but is not required to.

A decoder follows the same process of assigning indices. A possible
implementation could be outlined like this:

When a decoder supports stringref, then, upon encountering a
stringref-namespace tag, it should initialise an array and start decoding
the tagged value. The array stores, for each index, the major type (byte
or text) and the string value. Implementations are free to use their
internal format, of course - languages that do not distinguish between
text and byte strings can just store their internal string type in the
array.

Since stringref-namespace tags can be nested, the decoder needs to save
and restore the outer array before starting and after ending the decoding
of the tagged value. In a recursive implementation this could simply look
like this:

   global stringref_array; // global per encoder
   local outer_stringref_array;

   outer_stringref_array = stringref_array;
   stringref_array = allocate_array ();

   value = decode_cbor_value ();

   deallocate_array (stringref_array);
   stringref_array = outer_stringref_array;

   return value;

When a stringref tag is encountered, it MUST contain an unsigned integer
(major type 0). The integer MUST be less than the number of entries in the
array. The decoder than looks up the string in the array, given the index,
and copies it to the result (string values should not be shared unless
the language normally does so - if strings are mutable, a copy should be
decoded).

When a definite-length text or byte string is encountered, it is decoded
normally. Afterwards, the decoder checks whether it needs to assign an
index to the string, by comparing the string length against the minimum
length required by the index. If the string is shorter, nothing happens,
otherwise the next free index is used and the string (and the information
whether it is a byte or text string) is recorded at the end of the array.

In languages with dynamic arrays, this could be accomplished by using
the array length as the next index to be assigned, and pushing the
string onto the end of the array when it is long enough.

=head2 IMPLEMENTATION NOTE

The semantics of stringref tags require the decoder to be aware and the
encoder to be under control of the sequence in which data items are
encoded into the CBOR stream. This means these tags cannot be implemented
on top of every generic CBOR encoder/decoder (which might reorder entries
in a map); they typically need to be integrated into their works.

=head2 DESIGN RATIONALE

The stringref tag was chosen to be short, without requiring standards
action. The namespace tag is rare, so doesn't benefit from a short
encoding as much.

Implicit tagging/counting was chosen to support stream encoders. Having
to tag strings first requires either multiple passes over the data (which
might not be available, ruling out some encoders) or tagging more strings
than needed (wasting space). Explicit tagging also isn't necessarily
better even under optimal conditions, as the explicit tags waste space.

Stream decoders are affected less by implicit tagging than encoders.

The namespace tag was introduced for two reasons: first to allow embedding
of CBOR strings into other CBOR strings, secondly for decoding efficiency
- the decoder only has to expect stringref tags inside namespaces and
therefore doesn't have to maintain extra state outside of them.

=head1 EXAMPLES

The array-of-maps from the rationale example would normally compress to a
CBOR text of 83 bytes. Using this extension where possible, this reduces
to 74 bytes:

   d9 0100                      # tag(256)
      83                        # array(3)
         a3                     # map(3)
            44                  # bytes(4)
               72616e6b         # "rank"
            04                  # unsigned(4)
            45                  # bytes(5)
               636f756e74       # "count"
            19 01a1             # unsigned(417)
            44                  # bytes(4)
               6e616d65         # "name"
            48                  # bytes(8)
               436f636b7461696c # "Cocktail"
         a3                     # map(3)
            d8 19               # tag(25)
               02               # unsigned(2)
            44                  # bytes(4)
               42617468         # "Bath"
            d8 19               # tag(25)
               01               # unsigned(1)
            19 0138             # unsigned(312)
            d8 19               # tag(25)
               00               # unsigned(0)
            04                  # unsigned(4)
         a3                     # map(3)
            d8 19               # tag(25)
               02               # unsigned(2)
            44                  # bytes(4)
               466f6f64         # "Food"
            d8 19               # tag(25)
               01               # unsigned(1)
            19 02b3             # unsigned(691)
            d8 19               # tag(25)
               00               # unsigned(0)
            04                  # unsigned(4)

The following JSON array illustrates the effect of the index on the
minimum string length:

   [  "1", "222", "333",   "4", "555", "666", "777", "888", "999",
    "aaa", "bbb", "ccc", "ddd", "eee", "fff", "ggg", "hhh", "iii",
    "jjj", "kkk", "lll", "mmm", "nnn", "ooo", "ppp", "qqq", "rrr",
    "333",
    "ssss",
    "qqq", "rrr", "ssss"]

The strings "1", "4" and "rrr" are too short to get an index assigned. All
others that are not encoded with a stringref do (this assumes that JSON
strings are encoded as CBOR byte strings):

   d9 0100           # tag(256)
      98 20          # array(32)
         41          # bytes(1)
            31       # "1"
         43          # bytes(3)
            323232   # "222"
         43          # bytes(3)
            333333   # "333"
         41          # bytes(1)
            34       # "4"
         43          # bytes(3)
            353535   # "555"
         43          # bytes(3)
            363636   # "666"
         43          # bytes(3)
            373737   # "777"
         43          # bytes(3)
            383838   # "888"
         43          # bytes(3)
            393939   # "999"
         43          # bytes(3)
            616161   # "aaa"
         43          # bytes(3)
            626262   # "bbb"
         43          # bytes(3)
            636363   # "ccc"
         43          # bytes(3)
            646464   # "ddd"
         43          # bytes(3)
            656565   # "eee"
         43          # bytes(3)
            666666   # "fff"
         43          # bytes(3)
            676767   # "ggg"
         43          # bytes(3)
            686868   # "hhh"
         43          # bytes(3)
            696969   # "iii"
         43          # bytes(3)
            6a6a6a   # "jjj"
         43          # bytes(3)
            6b6b6b   # "kkk"
         43          # bytes(3)
            6c6c6c   # "lll"
         43          # bytes(3)
            6d6d6d   # "mmm"
         43          # bytes(3)
            6e6e6e   # "nnn"
         43          # bytes(3)
            6f6f6f   # "ooo"
         43          # bytes(3)
            707070   # "ppp"
         43          # bytes(3)
            717171   # "qqq"
         43          # bytes(3)
            727272   # "rrr"
         d8 19       # tag(25)
            01       # unsigned(1)
         44          # bytes(4)
            73737373 # "ssss"
         d8 19       # tag(25)
            17       # unsigned(23)
         43          # bytes(3)
            727272   # "rrr"
         d8 19       # tag(25)
            18 18    # unsigned(24)

This example shows three stringref-namespace tags, two of which are nested
inside another:

   256(["aaa", 25(0), 256(["bbb", "aaa", 25(1)]), 256(["ccc", 25(0)]), 25(0)])

   d9 0100               # tag(256)
      85                 # array(5)
         63              # text(3)
            616161       # "aaa"
         d8 19           # tag(25)
            00           # unsigned(0)
         d9 0100         # tag(256)
            83           # array(3)
               63        # text(3)
                  626262 # "bbb"
               63        # text(3)
                  616161 # "aaa"
               d8 19     # tag(25)
                  01     # unsigned(1)
         d9 0100         # tag(256)
            82           # array(2)
               63        # text(3)
                  636363 # "ccc"
               d8 19     # tag(25)
                  00     # unsigned(0)
         d8 19           # tag(25)
            00           # unsigned(0)

The decoded data structure might look like this:

  ["aaa","aaa",["bbb","aaa","aaa"],["ccc","ccc"],"aaa"]

=head1 IMPLEMENTATIONS

This section lists known implementations of this extension (L<drop me a
mail|mailto:cbor@schmorp.de?Subject=CBOR-stringref> if you want to be
listed here).

=over 4

=item * [Perl] L<CBOR::XS|http://software.schmorp.de/pkg/CBOR-XS.html> (reference implementation)

=back

Revision:	1.6
Committed:	Mon Apr 30 11:24:17 2018 UTC (6 years, 1 month ago) by root
Branch:	MAIN
CVS Tags:	rel-1_71
Changes since 1.5:	+19 -0 lines
Log Message:	* empty log message *
#	User	Rev	Content
1	root	1.1	=head1 REGISTRATION INFORMATION
2
3	root	1.3	Tag 256 (stringref-namespace)
4	root	1.1	Data Item multiple
5			Semantics mark value as having string references
6			Reference http://cbor.schmorp.de/stringref
7			Contact Marc A. Lehmann <cbor@schmorp.de>
8
9	root	1.3	Tag 25 (stringref)
10	root	1.1	Data Item unsigned integer
11			Semantics reference the nth previously seen string
12			Reference http://cbor.schmorp.de/stringref
13			Contact Marc A. Lehmann <cbor@schmorp.de>
14
15			=head1 RATIONALE
16
17			These two tags can be used to efficiently share constant strings, which
18			are common in many data structures. They are an optimisation only,
19			similar to LZSS compresssion, for the string major types in CBOR.
20
21			Background: many data structures or protocols contain repeated string
22			constants (often called atoms or quarks). For example, this is a
23			real-world (JSON-formatted) data structure that contains many repeated
24			strings used as map keys:
25
26			[
27			{
28			"name" : "Cocktail",
29			"count" : 417,
30			"rank" : 4
31			},
32			{
33			"rank" : 4,
34			"count" : 312,
35			"name" : "Bath"
36			},
37			{
38			"count" : 691,
39			"name" : "Food",
40			"rank" : 4
41			}
42			]
43
44			This can be encoded nicely with CBOR, but each occurrence of "name",
45			"count" and "rank" will be encoded separately. In highly repetitive data
46			structures (the above is taken from a game save file) these can take up
47			considerable space.
48
49			This scheme can be used to reduce this overhead with a simple scheme that
50			is easily implementable.
51
52			=head1 DESCRIPTION
53
54	root	1.3	Stringref consists of two tags, stringref-namespace (value C<256>),
55	root	1.1	which marks a value as containing string references, and stringref (value
56	root	1.3	C<25>), which references a string previously encoded in the value.
57	root	1.1
58			The stringref-namespace tag is used to define a namespace for the string
59			reference ids. stringref tags are only valid inside CBOR values marked
60			with stringref-namespace.
61
62			The purpose of stringref-namespace is three-fold:
63
64			=over 4
65
66			=item 1. It signals the decoder that it needs to keep a copy of every
67			string that is being decoded while decoding the tagged value (and
68			conversely, that there is no need for this extra overhead without this
69			tag).
70
71			=item 2. It can be used to confine string references to certain subtrees,
72			which might increase encoding efficiency by being able to use smaller
73			references.
74
75			=item 3. CBOR objects encoded with this scheme can be blindly embedded in
76			existing CBOR streams, as the indices are relative to the outer namespace
77			tag, and don't change when embedded.
78
79			=back
80
81			Within a value tagged with stringref-namespace, every string that is
82			encoded with a definite length and has a minimum length is assigned an
83			implicit index, starting from zero.
84
85			The minimum length required to get an index depends on the index value
86			that would be assigned to it, as stringref references should be shorter
87			than the string they reference, and there is no point in wasting indices
88			on strings that should not be referenced anyway.
89
90			The lengths are as follows:
91
92			string index minimum string length in octets
93			0 .. 23 3
94			24 .. 255 4
95			256 .. 65535 5
96	root	1.2	65536 .. 4294967295 7
97			4294967296 .. 11
98	root	1.1
99			The minimum string length is simply the length of the stringref tag (2
100			octets), plus the minimum size the resulting index takes up, using major
101			type 0 encoding.
102
103			Each time a string is to be encoded by the encoder, it can choose to
104			encode it as a normal string, or attempt to encode it as a stringref.
105
106			If it attempts to encode a stringref, it needs to check whether the string
107			has been assigned an index already. If yes, it can emit a stringref tag
108			with the index.
109
110			In all other cases the encoder must check whether it has to assign an
111			index value to it by checking against the minimum length required for the
112			next index value to be assigned, and assigning the next index value if the
113			string has minimum length or is longer.
114
115			A typical encoder might want to look up every string (using the major
116			type and the octets) it encodes in some kind of internal map that stores
117			previously-assigned indices. If found, it will emit a stringref. If not,
118			it attempts to assign an index to the string and stores that in its map.
119
120			Decoders might choose not to look up previous indices for a string, for
121			example, because it only wants to compress strings used as map keys, or
122			because it intrinsically knows that the string is only used once. This is
123			fine, but each time the encoder emits a string, it MUST assign an index if
124			the minimum length is reached or exceeded, even if it doesn't store that
125			index for later use.
126
127			Byte strings and text strings share the same index namespace, but are
128			distinct, even if their contents are bit-identical. Indefinite length
129			strings and the chunks they consist of are never assigned an index.
130
131			A typical encoder would emit the stringref-namespace tag once to tag
132			the top-level value in a CBOR stream, but nothing keeps an encoder
133			from nesting stringref-namespace tags. Nesting might be done for
134			convenience, to improve coding efficiency, to embed values or for any
135			other reason. Similarly, an encoder can sort map keys for greater coding
136			efficiency, but is not required to.
137
138			A decoder follows the same process of assigning indices. A possible
139			implementation could be outlined like this:
140
141			When a decoder supports stringref, then, upon encountering a
142			stringref-namespace tag, it should initialise an array and start decoding
143			the tagged value. The array stores, for each index, the major type (byte
144			or text) and the string value. Implementations are free to use their
145			internal format, of course - languages that do not distinguish between
146			text and byte strings can just store their internal string type in the
147			array.
148
149			Since stringref-namespace tags can be nested, the decoder needs to save
150			and restore the outer array before starting and after ending the decoding
151			of the tagged value. In a recursive implementation this could simply look
152			like this:
153
154			global stringref_array; // global per encoder
155			local outer_stringref_array;
156
157			outer_stringref_array = stringref_array;
158			stringref_array = allocate_array ();
159
160			value = decode_cbor_value ();
161
162			deallocate_array (stringref_array);
163			stringref_array = outer_stringref_array;
164
165			return value;
166
167			When a stringref tag is encountered, it MUST contain an unsigned integer
168			(major type 0). The integer MUST be less than the number of entries in the
169			array. The decoder than looks up the string in the array, given the index,
170			and copies it to the result (string values should not be shared unless
171			the language normally does so - if strings are mutable, a copy should be
172			decoded).
173
174			When a definite-length text or byte string is encountered, it is decoded
175			normally. Afterwards, the decoder checks whether it needs to assign an
176			index to the string, by comparing the string length against the minimum
177			length required by the index. If the string is shorter, nothing happens,
178			otherwise the next free index is used and the string (and the information
179			whether it is a byte or text string) is recorded at the end of the array.
180
181			In languages with dynamic arrays, this could be accomplished by using
182			the array length as the next index to be assigned, and pushing the
183			string onto the end of the array when it is long enough.
184
185			=head2 IMPLEMENTATION NOTE
186
187	root	1.3	The semantics of stringref tags require the decoder to be aware and the
188			encoder to be under control of the sequence in which data items are
189			encoded into the CBOR stream. This means these tags cannot be implemented
190			on top of every generic CBOR encoder/decoder (which might reorder entries
191			in a map); they typically need to be integrated into their works.
192	root	1.1
193	root	1.6	=head2 DESIGN RATIONALE
194
195			The stringref tag was chosen to be short, without requiring standards
196			action. The namespace tag is rare, so doesn't benefit from a short
197			encoding as much.
198
199			Implicit tagging/counting was chosen to support stream encoders. Having
200			to tag strings first requires either multiple passes over the data (which
201			might not be available, ruling out some encoders) or tagging more strings
202			than needed (wasting space). Explicit tagging also isn't necessarily
203			better even under optimal conditions, as the explicit tags waste space.
204
205			Stream decoders are affected less by implicit tagging than encoders.
206
207			The namespace tag was introduced for two reasons: first to allow embedding
208			of CBOR strings into other CBOR strings, secondly for decoding efficiency
209			- the decoder only has to expect stringref tags inside namespaces and
210			therefore doesn't have to maintain extra state outside of them.
211
212	root	1.1	=head1 EXAMPLES
213
214	root	1.3	The array-of-maps from the rationale example would normally compress to a
215			CBOR text of 83 bytes. Using this extension where possible, this reduces
216			to 74 bytes:
217
218			d9 0100 # tag(256)
219			83 # array(3)
220			a3 # map(3)
221			44 # bytes(4)
222			72616e6b # "rank"
223			04 # unsigned(4)
224			45 # bytes(5)
225			636f756e74 # "count"
226			19 01a1 # unsigned(417)
227			44 # bytes(4)
228			6e616d65 # "name"
229			48 # bytes(8)
230			436f636b7461696c # "Cocktail"
231			a3 # map(3)
232			d8 19 # tag(25)
233			02 # unsigned(2)
234			44 # bytes(4)
235			42617468 # "Bath"
236			d8 19 # tag(25)
237			01 # unsigned(1)
238			19 0138 # unsigned(312)
239			d8 19 # tag(25)
240			00 # unsigned(0)
241			04 # unsigned(4)
242			a3 # map(3)
243			d8 19 # tag(25)
244			02 # unsigned(2)
245			44 # bytes(4)
246			466f6f64 # "Food"
247			d8 19 # tag(25)
248			01 # unsigned(1)
249			19 02b3 # unsigned(691)
250			d8 19 # tag(25)
251			00 # unsigned(0)
252			04 # unsigned(4)
253
254			The following JSON array illustrates the effect of the index on the
255			minimum string length:
256
257			[ "1", "222", "333", "4", "555", "666", "777", "888", "999",
258			"aaa", "bbb", "ccc", "ddd", "eee", "fff", "ggg", "hhh", "iii",
259			"jjj", "kkk", "lll", "mmm", "nnn", "ooo", "ppp", "qqq", "rrr",
260			"333",
261			"ssss",
262			"qqq", "rrr", "ssss"]
263
264	root	1.5	The strings "1", "4" and "rrr" are too short to get an index assigned. All
265			others that are not encoded with a stringref do (this assumes that JSON
266			strings are encoded as CBOR byte strings):
267	root	1.3
268			d9 0100 # tag(256)
269			98 20 # array(32)
270			41 # bytes(1)
271			31 # "1"
272			43 # bytes(3)
273			323232 # "222"
274			43 # bytes(3)
275			333333 # "333"
276			41 # bytes(1)
277			34 # "4"
278			43 # bytes(3)
279			353535 # "555"
280			43 # bytes(3)
281			363636 # "666"
282			43 # bytes(3)
283			373737 # "777"
284			43 # bytes(3)
285			383838 # "888"
286			43 # bytes(3)
287			393939 # "999"
288			43 # bytes(3)
289			616161 # "aaa"
290			43 # bytes(3)
291			626262 # "bbb"
292			43 # bytes(3)
293			636363 # "ccc"
294			43 # bytes(3)
295			646464 # "ddd"
296			43 # bytes(3)
297			656565 # "eee"
298			43 # bytes(3)
299			666666 # "fff"
300			43 # bytes(3)
301			676767 # "ggg"
302			43 # bytes(3)
303			686868 # "hhh"
304			43 # bytes(3)
305			696969 # "iii"
306			43 # bytes(3)
307			6a6a6a # "jjj"
308			43 # bytes(3)
309			6b6b6b # "kkk"
310			43 # bytes(3)
311			6c6c6c # "lll"
312			43 # bytes(3)
313			6d6d6d # "mmm"
314			43 # bytes(3)
315			6e6e6e # "nnn"
316			43 # bytes(3)
317			6f6f6f # "ooo"
318			43 # bytes(3)
319			707070 # "ppp"
320			43 # bytes(3)
321			717171 # "qqq"
322			43 # bytes(3)
323			727272 # "rrr"
324	root	1.5	d8 19 # tag(25)
325			01 # unsigned(1)
326	root	1.3	44 # bytes(4)
327			73737373 # "ssss"
328			d8 19 # tag(25)
329			17 # unsigned(23)
330			43 # bytes(3)
331			727272 # "rrr"
332			d8 19 # tag(25)
333			18 18 # unsigned(24)
334
335			This example shows three stringref-namespace tags, two of which are nested
336			inside another:
337
338			256(["aaa", 25(0), 256(["bbb", "aaa", 25(1)]), 256(["ccc", 25(0)]), 25(0)])
339
340			d9 0100 # tag(256)
341			85 # array(5)
342			63 # text(3)
343			616161 # "aaa"
344			d8 19 # tag(25)
345			00 # unsigned(0)
346			d9 0100 # tag(256)
347			83 # array(3)
348			63 # text(3)
349			626262 # "bbb"
350			63 # text(3)
351			616161 # "aaa"
352			d8 19 # tag(25)
353			01 # unsigned(1)
354			d9 0100 # tag(256)
355			82 # array(2)
356			63 # text(3)
357			636363 # "ccc"
358			d8 19 # tag(25)
359			00 # unsigned(0)
360			d8 19 # tag(25)
361			00 # unsigned(0)
362
363			The decoded data structure might look like this:
364
365			["aaa","aaa",["bbb","aaa","aaa"],["ccc","ccc"],"aaa"]
366
367	root	1.4	=head1 IMPLEMENTATIONS
368
369			This section lists known implementations of this extension (L<drop me a
370			mail\|mailto:cbor@schmorp.de?Subject=CBOR-stringref> if you want to be
371			listed here).
372
373			=over 4
374
375			=item * [Perl] L<CBOR::XS\|http://software.schmorp.de/pkg/CBOR-XS.html> (reference implementation)
376
377			=back
378	root	1.1