Round Table in the UK

"Internationalising the Dots" by Ari-Pekka Launne This article is based on a presentation I gave in London, in the “Working with Dots" Seminar in January 2008. This was a joint seminar of ITMA and ECTA. I tried to address the questions related to internationalising of the domain names and to give the audience “a foreign point of view”. Some of this all may seem to be more of linguistics than related to IP law, but I believe it cannot be helped. The examples I have are given in my native tongue Finnish and I use some existing cases as examples. The internationalising of domain names is not an issue that can be completed just by making snapping your fingers to make it happen. There are technical issues and there are political issues. I try to avoid going deep into these, but I’m afraid that some words have to be said about these too. At the end of my article I try to summarise some pros and contras of the internationalising. Domain names are normally combinations of ASCII characters. ASCII script replicates the 26 characters of the standard Latin alphabet and numerals 0 to 9. There are languages that can be written by using these characters without any need for special letters or accents. An IDN (Internationalised Domain Name) is an Internet domain name that contains non-ASCII characters, in other words, characters that do not appear in the standard Latin alphabet. These include, for instance, standard Latin characters that have diacritics or accents marks e.g. é, ç or ö as well as characters from non-Latin alphabets e.g. Arabic, Cyrillic or Chinese. When an IDN is used, there may appear technical difficulties that are caused by the keyboard settings or by browser settings – these both are typically related to the local language, according to which the settings are made. So, if you entered café.com into your web browser would you expect it to point to the same website as cafe.com? How about hame.fi and häme.fi? Or even vodka.ru and водка.ru?

© ECTA May, 2008

23

I’m personally not able to spell in other languages than Finnish, Swedish and English, but I believe that these examples give you an idea about the IDNs. The short answer to the question posed above is that the compared pairs are not interchangeable domain names and they won’t point to the same website unless they are set up to do so. This does cause problems for the users of the Internet. Eric Ramage has noted in his presentations that both problems are extremely complex, and typical of those that he has been faced with when he’s been sitting on Nominet’s IDN sub-committee and on Eurid’s advisory committee on IDN. Internationalising is not entirely a new thing. The Internet Corporation for Assigned Names and Numbers (ICANN) has set a committee for studying issues related to this. The committee gave its Final Report to the ICANN Board on June 27, 2002. There have been several versions of draft guidelines which the community has commented on and last version of which was published almost a year ago. An online evaluation is currently going on, trying to test the system using several different scripts. This can be found at IDNwiki (idn.icann.org). Some of the problems can be easily understood using this test site. One of the problems of domain names is the fact that good domain names are hard to find – they all seem to be reserved already. A good domain name is short (easy to remember), relates to the content of the web site, to the company name or to trade mark, even to the goods or services. If it needs to be written with ASCII codes, there is only limited number of domain names available (well millions, but less of those that make any sense to anyone anyway). Internationalising of the dots gives more possibilities for the applicants, as it allows the use of several new languages and scripts as well as special letters. It is obvious that lots of information in the Internet is given in languages which are foreign to most of people. So it should also be obvious that also the domain names could be in languages not known to all of us. When internationalised domain names are used, communication between parties in different language zones may create need for technical solutions. Your keyboard may not contain the necessary letters. Or your browser may not work properly with them. For this reason, Finnish Communication Regulatory Authority (FICORA) does not recommend that an internationalised domain name is registered as the primary domain name. And after all, maybe the domain names would not be so easy to remember or search either, at least for those not speaking the respective language.

24

Round Table in the UK

It is estimated that there are about 6000 languages in the world. Chinese is spoken by 1,343 million people and English by 580 million people. The smallest languages have few speakers – for example, Skolt sami is estimated to have only 400 speakers. And sadly, some languages have died out recently. When writing, we need to use the characters of the respective language. In languages based in Latin script there are letters which are common to all of those languages, and then some – let’s call them special – other letters, letters with accents. You are likely to know lots of them: they exist in French, Spanish, Hungarian and Polish and in all of the languages spoken in northern Europe. Finnish is one of these and from now on I use it as an example. As many of the official EU languages make use of non-ASCII alphabets and characters which cannot currently be used in .eu domain names, IDNs will eventually be introduced by the .eu registry. Currently, there are only investigations and studies going on, and no time frame has been set yet. Finland has allowed internationalized domain names under .fi since 2005. The accepted characters include letters with accents from Finnish, Swedish and Sami. Finnish (suomi, or suomen kieli) is - A language spoken by - the majority of the population in Finland (92% as of 2006) by ethnic Finns outside of Finland - one of the official languages of Finland (Swedish, Sami) - an official minority language in Sweden - a member of the Finno-Ugric language family - one that modifies and inflects the forms of nouns, adjectives, pronouns, numerals and verbs, depending on their roles in the sentence - one that has 29 letters in the alphabet: 8 vowels, 21 consonants Finnish is spoken by about 5.3 million people. It is much unlike any other language in Europe, and has a curious thing in the writing of it: all letters carry a specific sound and therefore the pronunciation of a word can be detected directly from the spelling. Not many exceptions exist, although some sounds may be difficult to a Finn to produce. We have eight vowels: A, E, I, O, U, Y, Ä and Ö. Swedish Å (equals O) is used as well. Then we have, of course, the consonants, adding up to 29 letters that we need to be able to write understandable Finnish. Let’s now take a closer look at how the internationalising works. © ECTA May, 2008

25

First some words about the Internet. One of the problems with this wonderful invention lies in the fact that it is all about electricity: 1 and 0, on and off. All IP addresses are numbers, and IP numbers are hard to remember or search for, at least when you have lots of them. The solution has been to give the numbers corresponding names – domain names. Each domain name means one or several related IP address. And further more, one IP address can be reached through several domain names, if the holder has configured the computer to understand that it is, for example, kolster.fi, kolster.com, kolster.eu and so on. At the same time, the necessity of using the ASCII code version of the domain name does not disappear anywhere. The internationalised domain names may look confusing when written out with ASCII. With the whois information from FICORA, you will find for example the following: hame.fi and häme.fi (xn--hme-qla.fi). When I was trying to use this domain name my browser did not find www.häme.fi, but using the domain name in parentheses took me to the web site of the holder M-Bros Oy and Mr. Kimmo Mirjamo. And respectively, www.hame.fi takes us to the web site of the authorities of the Häme district. In a way the Internet makes the life of a Finn even harder than it is in the dark long winter days or long and sunny summer nights. This is mainly because of the ASCII code that has been agreed to be the basic code for use in messages and communication in the net. In principle all computers understand these codes, and writing of, for example, English is always possible with them. Why do we find it hard? Well, as I explained, in Finnish those three extra letters that are not included in the standard ASCII code carry information which may change the entire meaning of the message. These are the characters with the dots… Our word for a skirt is HAME. You also noted above that we have a district in Finland called HÄME (located in the southern Finland). These words carry no related meanings or connotations. So, when we tell our friends about the nice scenery, we may confuse the audience if the dots are not in the right place… Sometimes the difference can be overcome by using tricks. For example, HÄME could be spelled HAEME. You sometimes see this in sports, when the names of players from other countries are displayed using transliterations to local ways of speech. But this is not the best solution, if it is a solution at all. As in the example above, after internationalising the domain name hame.fi can be used also in the form häme.fi corresponding to the ASCII version of it, xn--hme-qla.fi). 26

Round Table in the UK

Hame.fi is reserved for the authorities of the district of Häme 27 September 2000. Häme.fi has been reserved for a private company since the beginning of the internationalisation from the beginning of September 2005. In situations like this, some problems may arise. What is the acceptable price for purchasing a domain name? Are there unfair intentions in registering the latter? Can the holder of the latter domain name rely on his reservation? Does a domain name carry some intellectual property related rights, in respect of confusing similarity, for example? In Finland, there is a family with the name BÅTMAN. When the registration of domain names was opened for individuals, this family would have been able to register batman.fi as their domain name (if it wasn’t registered by the trademark holder in the first place). Now, with the IDN’s they would be able to have the domain name they would actually have wanted in the first place, båtman.fi. This all means that, when the internationalising of a TLD begins, there is an important question of a possible sunrise period. Should the holders of the un-internationalised domain names be allowed to register the internationalised version as well or not? If we say that yes, they do have the right to reserve the respective internationalised domain names, the amount of domain names available does not actually grow any bigger. In such case new domain names would be available only if the holders of the earlier domain names decided that they do not want new domain names. The relation of domain names and trade marks is a question that deserves a study of its own. As for my opinion, I would like to say that there may appear some difficulties if we try to bring the concepts of trade mark law to the domain name issues directly, just as much as there would be if we brought the domain name regulations directly to the world of trade marks. Would we be willing to accept the possibility of thousands of new bars for registration or use of a trade mark appearing from the Internet – I do doubt it. But still, trade marks cannot be completely put aside when we discuss internationalising the domain names. Another question is the problem of spoofing and phishing. People tend not to be so careful when reading or writing a domain name or web site address. Misspelling is already a problem with domain names, and more possibilities to use it for unfair advantage will appear with the internationalisation. We know that this is a growing industry….

© ECTA May, 2008

27

The comparison of two marks may become more complicated with internationalisation. We already have a practice of having transliterations of marks into foreign languages. Now we are bringing into this the meaning of foreign words, and perhaps even the comparison of scripts. “Two marks that look different because one is in Latin characters and the other is [not] are not identical. Deciding whether marks are similar requires a comparison of the visual, aural and conceptual content of the marks. A decision on likelihood of confusion, where marks are not identical must involve a global assessment.” - OHIM For example, what if we should compare these two symbols 侖 and 俞 of the CJK unified ideographs, found with Word using the ”insert symbols” function. If these were used as trademarks, could they be considered confusingly similar? These are likely to be seen as devices because Chinese is not an official language in Europe. But they have meanings. The left one is pronounced as ”lun” and it means ”logical sequence” and ”coherence”. The right one is pronounced as ”yu”. It is used as a surname in China. The word has the meaning”consent” and”accede to (a request)”. It seems obvious that we cannot make correct interpretations of the signs without knowledge of the Chinese language. Googling does not help much in this exercise, since you usually get more than one hit. The correct and exact meaning of a word may be very hard to guess. In such cases the comparison can only be done on the overall visual impression the signs give. Another example. Greek in some cases looks very much like Latin when using upper case. What do you think about these signs?

So, to conclude my article, I see that there are many pros and cons to the internationalising of domain names. There is broad recognition that IDNs are highly likely to help facilitate Internet use by the majority of the world's population whose native scripts are non-Latin. Naturally, not everyone in the world is used to spelling with Latin letters, which is only one script available among the others. But there are also a number of crucial technical and policy issues essential to 28

Round Table in the UK

ensuring that IDNs might be deployed in a manner that does not harm the stability of the Internet. These issues need to be discussed further. The biggest pros may be that users of a specific language can use all of their own language’s characters when creating domain names. This is likely to make the domain names more useful in the respective area or language group. At the same time new domain names become available. The biggest cons would then be criminal actions in the net, enabled or made easier by the internationalisation of domain names. No doubt the users could grow suspicious towards sites in foreign languages. Internationalisation poses a huge technical challenge for the Internet society. These must and can be overcome, but lots of work is needed, and there’s a long way to go before this necessary development of information technology will be complete.

Ari-Pekka Launne Kolster Oy Ab Iso Roobertinkatu 23 00120 Helsinki, Finland

© ECTA May, 2008

Tel. +358 9 618821 Fax. +358 9 602 244 E-mail: [email protected] E-mail : [email protected] Internet: www.kolster.fi

29

30