NAME
   Data::UUID::NCName - Make valid NCName tokens which are also UUIDs

VERSION
   Version 0.06

SYNOPSIS
       use Data::UUID::NCName qw(:all);

       my $uuid  = '1ff916f3-6ed7-443a-bef5-f4c85f18cd10';
       my $ncn   = to_ncname($uuid, version => 1);
       my $ncn32 = to_ncname($uuid, version => 1, radix => 32);

       # $ncn is now "EH_kW827XQ6719MhfGM0QL".
       # $ncn32 is "ed74rn43o25b255puzbprrtiql" and case-insensitive.

       # from Test::More, this will output 'ok':
       is(from_ncname($ncn, version => 1),
           $uuid, 'Decoding result matches original');

DESCRIPTION
   The purpose of this module is to devise an alternative representation of
   the UUID <http://tools.ietf.org/html/rfc4122> which conforms to the
   constraints of various other identifiers such as NCName, and create an
   isomorphic <http://en.wikipedia.org/wiki/Isomorphism> mapping between
   them.

FORMAT DEPRECATION NOTICE
   After careful consideration, I have decided to change the UUID-NCName
   format in a minor yet incompatible way. In particular, I have moved the
   quartet containing the "variant"
   <https://tools.ietf.org/html/rfc4122#section-4.1.1> to the very end of
   the identifier, whereas it previously was mixed into the middle
   somewhere.

   This can be considered an application of Postel's Law
   <https://en.wikipedia.org/wiki/Postel%27s_law>, based on the assumption
   that these identifiers will be generated through other methods, and
   potentially naïvely. Like the "version" field, the "variant" field has a
   limited acceptable range of values. If, for example, one were to attempt
   to generate a conforming identifier by simply generating a random Base32
   or Base64 string, it will be difficult to ensure that the "variant"
   field will indeed conform when the identifier is converted to a standard
   UUID. By moving the "variant" field out to the end of the identifier,
   everything between the "version" and "variant" bookends can be generated
   randomly without any further consideration, like so:

       our @B64_ALPHA = ('A'..'Z', 'a'..'z', 0..9, qw(- _));

       sub make_cheapo_b64_uuid_ncname () {
           my @vals = map { int rand 64 } (1..20); # generate content
           push @vals, 8 + int rand 4;             # last digit is special
           'E' . join '', map { $B64_ALPHA[$_] } @vals; # 'E' for UUID V4
       }

       # voilà:
       my $cheap = make_cheapo_b64_uuid_ncname;
       # EPrakcT1o2arqWSOuIMGSK or something

       # as expected, we can decode it (version 1, naturally)
       my $uu = Data::UUID::NCName::from_ncname($cheap, version => 1);
       # 3eb6a471-3d68-4d9a-aaea-5923ae20c192 - UUID is valid

   Furthermore, since the default behaviour is to align the bits of the
   last byte to the size of the encoding symbol, and since the "variant"
   bits are masked, a compliant RFC4122 UUID will *always* end with "I",
   "J", "K", or "L", in *both* Base32 (case-insensitive) and Base64
   variants.

   Since I have already released this module prior to this format change, I
   have added a "version" parameter to both "to_ncname" and "from_ncname".
   The version currently defaults to 1, the new one, but will issue a
   warning if not explicitly set. Later I will finally remove the warning
   with 1 as the default. This should ensure that any code written during
   the transition produces the correct results.

       Unless you have to support identifiers generated from version 0.04
       or older, you should be running these functions with "version => 1".

RATIONALE & METHOD
   The UUID is a generic identifier which is large enough to be globally
   unique. This makes it useful as a canonical name for data objects in
   distributed systems, especially those that cross administrative
   jurisdictions, such as the World-Wide Web. The representation
   <http://tools.ietf.org/html/rfc4122#section-3>, however, of the UUID,
   precludes it from being used in many places where it would be useful to
   do so.

   In particular, there are grammars for many types of identifiers which
   must not begin with a digit. Others are case-insensitive, or prohibited
   from containing hyphens (present in both the standard notation and
   Base64URL), or indeed anything outside of "^[A-Za-z_][0-9A-Za-z_]*$".

   The hexadecimal notation of the UUID has a 5/8 chance of beginning with
   a digit, Base64 has a 5/32 chance, and Base32 has a 3/16 chance. As
   such, the identifier must be modified in such a way as to guarantee
   beginning with an alphabetic letter (or underscore "_", but some
   grammars even prohibit that, so we omit it as well).

   While it is conceivable to simply add a padding character, there are a
   few considerations which make it more appealing to derive the initial
   character from the content of the UUID itself:

   *   UUIDs are large (128-bit) identifiers as it is, and it is
       undesirable to add meaningless syntax to them if we can avoid doing
       so.

   *   128 bits is an inconvenient number for aligning to both Base32 (130)
       and Base64 (132), though 120 divides cleanly into 5, 6 and 8.

   *   The 13th quartet, or higher four bits of the
       "time_hi_and_version_field" of the UUID is constant, as it indicates
       the UUID's version. If we encode this value using the scheme common
       to both Base64 and Base32, we get values between "A" and "P", with
       the valid subset between "B" and "F".

   Therefore: extract the UUID's version quartet, shift all subsequent data
   4 bits to the left, zero-pad to the octet, encode with either
   *base64url* or *base32*, truncate, and finally prepend the encoded
   version character. Voilà, one token-safe UUID.

APPLICATIONS
   XML IDs
       The "ID" production appears to have been constricted, inadvertently
       or otherwise, from Name <http://www.w3.org/TR/xml11/#NT-Name> in
       both the XML 1.0 and 1.1 specifications, to NCName
       <http://www.w3.org/TR/xml-names/#NT-NCName> by XML Schema Part 2
       <http://www.w3.org/TR/xmlschema-2/#ID>. This removes the colon
       character ":" from the grammar. The net effect is that

           <foo id="urn:uuid:b07caf81-baae-449d-8a2e-48c0f5fa5538"/>

       while being a *well-formed* ID *and* valid under DTD validation, is
       *not* valid per XML Schema Part 2 or anything that uses it (e.g.
       Relax NG).

   RDF blank node identifiers
       Blank node identifiers in RDF are intended for serialization, to act
       as a handle so that multiple RDF statements can refer to the same
       blank node. The RDF abstract syntax specifies
       <http://www.w3.org/TR/rdf-concepts/#section-URI-Vocabulary> that the
       validity constraints of blank node identifiers be delegated to the
       concrete syntax specifications. The RDF/XML syntax specification
       <http://www.w3.org/TR/rdf-syntax-grammar/#rdf-id> lists the blank
       node identifier as NCName. However, according to the Turtle spec
       <http://www.w3.org/TR/turtle/#BNodes>, this is a valid blank node
       identifier:

           _:42df00ec-30a2-431f-be9e-e3a612b325db

       despite an older version
       <http://www.w3.org/TeamSubmission/turtle/#nodeID> listing a
       production equivalent to the more conservative NCName. NTriples
       syntax is even more constrained
       <http://www.w3.org/TR/rdf-testcases/#ntriples>, given as
       "^[A-Za-z][0-9A-Za-z]*$".

   Generated symbols

           There are only two hard things in computer science: cache
           invalidation and naming things [and off-by-one errors].

           -- Phil Karlton [extension of unknown origin]

       Suppose you wanted to create a literate programming
       <http://en.wikipedia.org/wiki/Literate_programming> system (I do).
       One of your (my) stipulations is that the symbols get defined in the
       *prose*, rather than the *code*. However, you (I) still want to be
       able to validate the code's syntax, and potentially even run the
       code, without having to commit to naming anything. You are (I am)
       also interested in creating a global map of classes, datatypes and
       code fragments, which can be operated on and tested in isolation,
       ported to other languages, or transplanted into the more
       conventional packages of programs, libraries and frameworks. The
       Base32 UUID NCName representation should be adequate for placeholder
       symbols in just about any programming language, save for those which
       do not permit identifiers as long as 26 characters (which are
       extremely scarce).

EXPORT
   No subroutines are exported by default. Be sure to include at least one
   of the following in your "use" statement:

   :all
       Import all functions.

   :decode
       Import decode-only functions.

   :encode
       Import encode-only functions.

   :32 Import base32-only functions.

   :58 Import base58-only functions.

   :64 Import base64-only functions.

SUBROUTINES
 to_ncname $UUID [, $RADIX ] [, %PARAMS ]
   Turn $UUID into an NCName. The UUID can be in the canonical (hyphenated)
   hexadecimal form, non-hyphenated hexadecimal, Base64 (regular and
   base64url), or binary. The function returns a legal NCName equivalent to
   the UUID, in either Base32, Base58, or Base64 (url), given a specified
   $RADIX of 32, 58, or 64. If the radix is omitted, Base64 is assumed.

   The following keyword parameters are also accepted, and override the
   positional parameters where applicable:

   radix 32|58|64
       Either 32 or 64 to explicitly specify Base32, Base58, or Base64
       output. Defaults to 64.

   version 0|1
       Version 0 will generate the original version of NCName identifiers,
       prior to the changes noted above. Version 1 is the new version,
       which is *not* backwards-compatible. The default, for a transitional
       period, is to generate version 0, but complain about it. Set the
       version explicitly (to 1, or to 0 if you need backwards
       compatibility) to eliminate the warning messages.

   align $FALSY|$TRUTHY
       Align the last 4 bits to the Base32/Base64 symbol size. You almost
       certainly want this, so the default is *true*. (Does not apply to
       Base58.)

 from_ncname $NCNAME [, $FORMAT [, $RADIX] ] [, %PARAMS ]
   Turn an appropriate $NCNAME back into a UUID, where *appropriate*,
   unless overridden by $RADIX, is defined beginning with one initial
   alphabetic letter (A to Z, case-insensitive) followed by either:

   25 Base32 characters, or
   21 Base64URL characters.

   The function will return "undef" immediately if it cannot match either
   of these patterns. Input past the 21-character mark (for Base64) or
   25-character mark (for Base32) is ignored.

   This function returns a UUID of type $FORMAT, which if left undefined,
   must be one of the following:

   str The canonical UUID format, like so:
       "33fcc995-5d10-477e-a9b4-c9cc405bbf04". This is the default.

   hex The same thing, minus the hyphens.

   b64 Base64.

   bin A binary string.

   This function also takes the new keyword-style parameters:

   format
       As above.

   radix
       As above.

   version
       Sets the identifier version. Defaults to version 0 with a warning.
       See the note about setting an explicit "version" parameter in
       "to_ncname".

   align
       Assume the last few bits are aligned to the symbol, as in
       "to_ncname".

 to_ncname_64 $UUID [, %PARAMS ]
   Shorthand for Base64 NCNames.

 from_ncname_64 $NCNAME [, $FORMAT | %PARAMS ]
   Ditto.

 to_ncname_58 $UUID [, %PARAMS ]
   Shorthand for Base58 NCNames.

 from_ncname_58 $NCNAME [, $FORMAT | %PARAMS ]
   Ditto.

 to_ncname_32 $UUID [, %PARAMS ]
   Shorthand for Base32 NCNames.

 from_ncname_32 $NCNAME [, $FORMAT | %PARAMS ]
   Ditto.

AUTHOR
   Dorian Taylor, "<dorian at cpan.org>"

BUGS
   Please report bugs/issues/etc in GitHub
   <https://github.com/doriantaylor/p5-data-uuid-ncname/issues>.

   *   MetaCPAN

       <https://metacpan.org/release/Data-UUID-NCName>

   *   GitHub repository (bugs also go here)

       <https://github.com/doriantaylor/p5-data-uuid-ncname>

   *   AnnoCPAN: Annotated CPAN documentation

       <http://annocpan.org/dist/Data-UUID-NCName>

   *   CPAN Ratings

       <http://cpanratings.perl.org/d/Data-UUID-NCName>

SEE ALSO
   *   UUID::Tiny

   *   Data::UUID

   *   OSSP::uuid

   *   RFC 4122 <http://tools.ietf.org/html/rfc4122>

   *   RFC 4648 <http://tools.ietf.org/html/rfc4648>

   *   Namespaces in XML <http://www.w3.org/TR/xml-names/#NT-NCName>
       (NCName)

   *   W3C XML Schema Definition Language (XSD) 1.1 Part 2: Datatypes
       <http://www.w3.org/TR/xmlschema11-2/#ID> (ID)

   *   RDF/XML Syntax Specification (Revised)
       <http://www.w3.org/TR/rdf-syntax-grammar/#rdf-id>

   *   Turtle <http://www.w3.org/TR/turtle/#BNodes>

   This module lives under the "Data::" namespace for the purpose of
   namespace hygiene. The main module *does not* depend on Data::UUID,
   howevever the script uuid-ncname *does* depend on UUID::Tiny to generate
   UUIDs.

LICENSE AND COPYRIGHT
   Copyright 2012-2018 Dorian Taylor.

   Licensed under the Apache License, Version 2.0 (the "License"); you may
   not use this file except in compliance with the License. You may obtain
   a copy of the License at <http://www.apache.org/licenses/LICENSE-2.0> .

   Unless required by applicable law or agreed to in writing, software
   distributed under the License is distributed on an "AS IS" BASIS,
   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
   See the License for the specific language governing permissions and
   limitations under the License.

You are viewing proxied material from ftp.icm.edu.pl. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.