Noms de domaine internationalisés (IDN)
Votre nom de domaine peut contenir des caractères comportant n’importe quel script de l’une des langues officielle de l’UE. Ces caractères incluent, par exemple, le å suédois, le ü allemand, le ș roumain et des caractères bulgares (cyrilliques) ainsi que l’alphabet grec dans son ensemble.
Les noms de domaine qui contiennent ces caractères spéciaux dits, non-ASCII, sont appelés Noms de Domaine Internationalisés (IDN).
Les IDN sont particulièrement importants pour le .eu car l’Union européenne compte 28 Etats-membres et 24 langues officielles. Beaucoup de ces langues utilisent des alphabets contenant des caractères non définis par la norme ASCII.
Pour voir les caractères non-ASCII pouvant être utilisés dans votre nom de domaine, veuillez consulter la liste des caractères autorisés ci-dessous.
Il y a également des règles pour les noms de domaine que vous devez garder à l’esprit lorsque vous enregistrez un IDN.
Veuillez noter qu’avec l’introduction du .ею (alphabet cyrillique), l’alphabet du nom de domaine de second niveau doit correspondre à l’alphabet utilisé pour l’extension de premier niveau (.eu ou .ею). C’est-à-dire que si le nom de domaine est enregistré en alphabet latin, l’alphabet de l’extension sera .eu. En revanche, si le nom de domaine enregistré utilise l’alphabet cyrillique, l’extension devra être .ею. Un bureau d’enregistrement souhaitant enregistrer un nom de domaine uniquement numérique –incluant éventuellement un tiret- doit spécifier l’extension de premier niveau lors de l’enregistrement. Si rien n’est spécifié, l’extension par défaut sera .eu.
Les internautes peuvent aussi atteindre votre site internet ou votre boite mail en utilisant la chaîne ACE de votre IDN, si leurs navigateurs ou applications de messagerie ne supportent pas encore les noms de domaine internationalisés.
Caractères autorisés et groupe d’homoglyphes
Les noms de domaine classiques (non IDN) se composent de:
- Caractères de a à z
- Chiffres de 0 à 9
- Tiret (-)
- Ils utilisent toujours l'extension .eu
Les noms de domaine internationalisés (IDN) se composent de :
- Caractères spéciaux (Unicode) des alphabets cyrillique, grec ou latin. Cliquez ici pour voir la liste de tous les caractères autorisés.
- Chiffres de 0 à 9
- Tiret (-)
- Ils ne peuvent mélanger des alphabets différents. Tous les caractères du second niveau (qui se trouvent avant l’extension) doivent être du même alphabet. Les noms de domaine utilisant l’alphabet grec ou latin auront l’extension .eu, alors que les noms de domaine utilisant des caractères cyrilliques auront l’extension .ею. Les chiffres de 0 à 9 et le tiret peuvent être utilisés avec tous les alphabets (latin, grec, cyrillique).
Voici la liste de tous les caractères non-ASCII que vous pouvez utiliser dans votre nom de domaine .eu ainsi que les tables des groupes d’homoglyphes. Chaque caractère est listé avec son Unicode officiel.
Please note the following information is only available in English.
IDNA2008 and homoglyph bundling
Following the amendment of , on 6 May 2015 EURid introduced a revised mechanism for handling Internationalised Domain Names containing non-ASCII characters (shift from IDNA2003 to IDNA2008) as well as the so-called “homoglyph bundling”.
Implications of moving from IDNA2003 to IDNA2008
IDNA stands for Internationalising Domain Names in Applications. It is a mechanism for handling internationalised domain names containing non-ASCII characters. For instance, IDNA2003 mapped IDNs as follows: café as a normalised IDN is converted into an ACE-string, namely: xn--caf-dma. The same applies to кафене which is converted into xn--80akarr4b.
From the moment the EURid registration system supported the IDNA2008 protocol, the following updates entered into force:
A) The list of accepted characters is adjusted to those supported by the IDNA2008 protocol. The most recent version of the list of accepted characters can be consulted . More specifically:
- ß and ς are no longer mapped to equivalent letters, but can be used in the input as fully accepted characters.
- The lower case mapping still converts upper case characters into their lower case equivalent (A → a, B → b, etc.), however there is an exception to this rule: Σ → σ.
- Mapping of ẞ → ß.
- ŀ and ŉ will continue to mapped to separate characters: normal l followed by dot and apostrophe followed by normal n.
- Greek letters with iota below continue to be allowed on the input, and will be mapped to separate characters. For instance: ᾳ → αι.
Once a domain name is case folded, it is normalized. In Cyrillic no further normalisation of the domain name is done. For the Latin and Greek scripts, the normalisation tables are and contain the characters that are actually normalised (transformed into another character or series of characters). The actual registered domain name is the domain name that is the result of this normalisation step.
B) Domain names from two different scripts that are visually indistinguishable and therefore, might lead to confusion, are bundled via the so-called “homoglyph bundling” procedure.
C) The legacy registrations, meaning all registrations existing prior to 6 May 2015 that are no longer compliant with the new registration rules, either because they contain characters no longer supported or contain sequences of characters no longer allowed, continue to be registered but specific Legacy rules apply.
As long as the legacy registration continues to be registered, standard transactions such as updates, renewals, transfers and reactivations from quarantine continue to be possible. Should the legacy domain name be deleted, it will no longer be able to be registered.
Standard transactions such as updates, renewals, transfers and reactivations from quarantine continue to be possible. Should the legacy domain name be deleted, its status becomes “not allowed”.
Introduction to Homoglyph Bundling
Homoglyphs are characters which, due to similarities in size and shape, might appear identical at first glance. The homoglyphs below represent two unique characters belonging to two different scripts, or alphabets:
Cyrillic character a → Unicode number 0430
Latin character a → Unicode number 0061
With the introduction of the so-called “homoglyph bundling” procedure, domain names that might look confusingly similar are prevented from being registered.
Homoglyph bundling is when you register an IDN and the registration system automatically bundles all the homoglyphs of that name (if there are any). This means that several domain names are bundled at one time, and none of the other domain names in that bundle can be registered.
The Homoglyph Bundling rules can be summarised as follows:
A) Visually similar characters across different scripts are bundled.
- Latin e versus Cyrillic е
- Latin a versus Greek α (uppercase)
There are exceptions to this rule. Below you will be able to find a non-exhaustive list. More detailed information can be found :
- Latin ß and Latin ss,
- Latin ss and Greek β: these are characters from 2 different scripts, which are not visually similar,
- Greek ς and Greek σ,
- Greek α and Greek ἀ ἁ ἂ ἃ ἄ ἅ and Greek ᾀ ᾁ ᾂ ᾃ ᾄ ᾅ and
- Greek αi and Greek ἀi ἁi ἂi ἃi ἄi ἅi.
B) If one domain name in a homoglyph bundle exists, none of the other domain names in that bundle can be registered.
The word “exists” should be interpreted in the previous sentence as having either one of the following .eu domain name statuses: in use, registered (on hold, suspended, seized), withdrawn, quarantine. When querying a domain name that is in a bundle via the WHOIS, it will return the status “homoglyph blocked”.
Should one or more domain names happen to be part of a bundle but were registered before 6 May 2015, they will continue to be registered. Should they be deleted, they will not be available for new registration and become “homoglyph blocked” in the EURid WHOIS database.
As described earlier, as a consequence of the implementation of the IDNA2008 standard protocol that replaces the currently deployed IDNA2003 protocol, new characters are going to be supported when registering a .eu domain name while others are going to be phased out.
This section aims to explain both the changes from the supported character perspective and the legacy policy for characters or sequences of characters being phased out.
Managing the introduction of the ß (Latin small letter Sharp S, Unicode U+00DF) and the ς (Greek small letter ending Sigma, Unicode U+03C2):
The IDNA2008 protocol supports both the German Eszett (ß) and the Greek ending sigma (ς) on input as fully allowed characters. Due to the introduction of the homoglyph bundling mechanism, both characters are part of the homoglyph bundling algorithm, meaning that registered domain names containing characters “ss” or the Greek normal sigma (σ) prevent domain names with German Eszett (ß) or Greek ending sigma (ς) from being registered.
However, considering the limited support of the newly introduced characters by many web browsers, a registrant who has registered a domain name containing characters “ss” or the Greek normal sigma (σ), or vice versa - German Eszett (ß) or Greek ending sigma (ς) - can request EURid to activate the corresponding domain name written with the the German Eszett (ß) or Greek ending sigma (ς) - or vice versa with the characters “ss” or the Greek normal sigma (σ) - at any time. The two names must be assigned to the same registrant. They will coexist and both be invoiced to the registrar.
EURid will regularly check that the domain names are assigned to the same registrant and if not, will revoke the domain name activated by the latter registrant.
EURid will continue to investigate and assess the support of the IDNA2008 protocol through the most common client software (web browsers, email clients, …). When the aforementioned support is deemed sufficient by EURid and the Internet, as well as by the technical community, the domain names for which two “versions” coexist - those with characters “ss”/ German Eszett (ß) or the Greek normal sigma (σ)/Greek ending sigma (ς) – the registrar will be requested to choose which domain name they wish to keep activated. The other domain name will be withdrawn and be homoglyph bundle blocked by the other name.
Registrants of existing domain names with the aforementioned characters who wish to activate the corresponding domain name written with the equivalent characters have to contact their registrar and request activation of the equivalent domain name.The registrar must then send the activation request to EURid and also state that they received the demand to activate the equivalent domain name from the current registrant.
This policy supersedes the previously communicated policy that foresaw the following:
"To allow registrants of existing domain names which contain the characters “ss” or the Greek normal sigma (σ) to switch to the corresponding domain name written with the German Eszett (ß) or the Greek ending sigma (ς), EURid has designed the following policy: If a registrant has registered a .eu domain name with “ss” or Greek normal sigma (σ) before 6 May 2015, it will continue to exist and will remain registered. The registrant may keep the currently registered domain name, or may at any time request that the equivalent domain name with German Eszett (ß) or Greek ending sigma (ς) be activated. By requesting that the equivalent domain name is activated, the registrant and registrar accepts that one year later the domain name with “ss” or Greek Normal Sigma (σ) is revoked and homoglyph bundled.
EURid will directly activate the domain name in the registrar’ portfolio. This is considered a normal new registration and is charged as such. Furthermore, the new domain name is going to have its own registration and expiry dates independently from those of the original name. The currently registered domain name will enter into a one (1) year phase-out period. After the phase-out period the original domain name will be revoked and will be homoglyph blocked in the EURid WHOIS database, which will prevent it from being registered.
Please note that the option of requesting the activation of the domain name with the newly supported character has unlimited validity considering that in any case the currently registered domain name will prevent the equivalent domain name with German Eszett (ß) or Greek Final Sigma (ς) from being registered (therefore, having the homoglyph blocked status in the WHOIS database)."
Managing .eu domain names with hyphens in the second, third and fourth position, or with “ŀ” (L followed by middle dot but not followed by a subsequent L), or with "ı" (dotless i):
.eu domain names that have been registered
- with hyphens in the second, third and fourth position, or
- with “ŀ” (L followed by middle dot but not followed by a subsequent L), or
- with "ı" (dotless i)
are no longer supported. To allow registrants to seek proper solutions to find possible alternatives, they will remain operational for a term of one (1) year until 6 May 2016. After then, they will be revoked and will not be allowed for re-registration. During the one (1) year phase-out period (until 6 May 2016) registrants can still update, reactivate the domain name when in quarantine or transfer it. However if deleted, these domain names will not become available for a new registration.