TC 304
European Localization Requirements
Exigences européennes de la localisation
Europäische Lokalisationsanforderungen

Minutes of a workshop:

The Euro in IT standards, II

CEN/TC304 N770
Place: Dublin, Ireland
Time: Thursday 16 October 1997 14:00-19:30

Þorvarður Kári Ólafsson opened the workshop and welcomed the participants.
Michael Everson was selected as secretary.

Participants in the meeting:
DE Marc Wilhelm Küster
DK Keld Simonsen
FI Erkki Kolehmainen
FI Klaas Ruppel
GB Graham Dixon
GB Chris Makemson
GR Evangelos Melagrakis
IE Michael Everson
IE Mícheál MacConmara
IS Þorgeir Sigurðsson
IS Þorvarður Kári Ólafsson
NO Trond Trosterud
PL Elzbieta Broma-Wrzesieñ
SE Wolf Arfvidson
SE Karl-Ivar Larsson
SE Wera Lundström
SE Monica Ståhl
SE Göran Tellström
US Mike Ksar
CEC Ken Thompson
CEC Lazaros Tossounidis

The participants in the workshop were representing their expert views, not necessarily the views of their National Bodies or those of their employers.

0.1 Adoption of the agenda and identification of documents (ED6, ED12)

Agenda adopted and documents were assigned. The numbers in this document relate to the Recommendations arising from the Crete workshop.

1. Financial transfers and electronic commerce

1.1 Registration of currency code

Ken Thompson: I can confirm that the registration in ISO 4217 was done a year ago. The country code has not been registered in ISO 3166.
Michael Everson: The ISO 3166 Registration Authority says that entities like the EU, EFTA, NATO, etc. are outside their scope, but have asked ISO for clarification.
Mike Ksar: is there a document saying that EUR has been registered? Who did it? When did it happen:? We cannot take it for granted that it is done. We need to see a document.
Chris Makemson: We need to see a copy of the newsletter.
Mike Ksar: BSI is the RA for ISO 4217.

  • Recommendation 1.1: TC304 secretariat to obtain two documents (the application and the approval) from BSI and circulate it to TC304 and to SC2.

    2. Output: International registration of the EURO SIGN glyph (ED5, ED13)

    2.1 Registration of EURO SIGN with AFII

    Ken Thompson: The Commission has sent a letter to AFII requesting the forms and advice on how to register the glyph. We were waiting to see what the Dublin Workshop would do. The Commission would like TC304's help in preparing the application, but will retain responsibility for the registration.
    Mike Ksar: Asmus Freytag, president of AFII probably hasn't got it because it went to the wrong address (AFII has gone through a lot of changes recently). The Commission should send a new application to him at the AFII Post Office Box, cc: Mike Ksar. The registration will happen through the Unicode Consortium, because they are offering administrative support.
    Ken Thompson: The letter sent or to be resent is a formal letter; we have to have a formal situation. There are two actions: one is formal, one is practical. See also discussion under agenda item 2.4.

  • Recommendation 2.1.1: The workshop recommends that TC304 nominate Michael Everson and Karl-Ivar Larsson to assist DGIII/B to register a single generic glyph in the AFII registry, using the input from ED13, ED14, and ED15.

    2.3. Design of OCR glyphs

  • Recommendation 2.3.1 Karl-Ivar Larsson to submit a paper on OCR-B design. ("How do we achieve what is in Recommendation 5.4?")

    2.4. Design of EURO SIGN glyphs in ordinary fonts DT14

    Ken Thompson: What is there has been tabled internally; we should try to cover some lookalike forms, including a preferred form and permitted variants.
    Lazarus Tossounidis: In big organizations like the Commission, there is a main goal to achieve, and responsibilities are assigned. The EURO SIGN was created, then several departments had to do different things to put it forward. When it comes to IT, we have noted that the font issue needs work. The background is, when this logotype was first communicated, there were some metrics associated to it. The typographic community didn't take well to this. I was obliged to provide to the CEC fonts that supported it, so we had to go back to Monotype to get this for our implementation. It is not appropriate for everything to use the logotype. As far as I am concerned, an expert, or a company, should make a proposal for a preferred approach in a given font. This task hasn't been approved, and hasn't been done. It just hasn't been done. I don't know what will happen in the next few months. A professional description of these matters should be provided.
    Karl-Ivar Larsson: In one specific instance, OCR-B should have the same status as the original glyph, its shape must be absolutely defined as to outline, according to SISO 9541.
    Chris Makemson: we are looking at more than one registration, a specification of variations, a registration in relation to OCR.
    Mícheál MacConmara: There seems to be little consultation outside of the Commission. It may have been thought that it was just a marketing symbol. People didn't realize the large implications of the creation of this symbol. The implication is that the statement has been made is that for it to be a real currency it needs a symbol. There is no proof of a user requirement for this, other than the Commission's assertion that it is required. If you register a logo you can say what it
    should look like. A letter of the alphabet or a generic symbol like a dollar sign can't be specified. Style is involved. Typographers must be given freedom.
    Lazarus Tossounidis: It was a logotype, but now variation is being considered. I would support the idea that the Commission is not cut off from the people, and won't make decisions like that, thought the press often suggests this. It is a symbol like everything else. There is a strong political will behind this. Nothing has been enforced, if you don't want to use the symbol.
    Mícheál MacConmara: There is an appearance that some of the requirements for the glyph makes it harder to implement. US and Japanese companies won't put it on the keyboard.
    Klaas Ruppel: ED13 is a useful paper: the only thing that AFII can register is the logo; the glyphs will be made by companies.
    Karl-Ivar Larsson: There is only one standard for glyphs: ISO 1073: OCR-A and OCR-B. There are situations where you need a specific font, but these are very specialized instances.
    Graham:Under ISO 10036 you can register glyph collections and individual glyphs. What you record is a representative version of that shape.
    Mike Ksar: I think that the AFII registry you can register a collection. Variations of the font can be there. Anyone in the world can register the same symbol. The Commission's registration is identified as that, Microsoft can do the same and say that this is their version. A glyph is a well-defined shape, not a shape from which other styles are derived or can be derived. That's why the character can have multiple glyphs. Registration can be done by anyone who wishes to. EU registration does not protect the sign in any way from other variants. They want to register a logo for their publications, but not as a currency sign as such.
    Michael Everson: US and Asians may not put the symbol on keyboards, but it is in character sets and access to the symbol in keyboard driver software will be provided. What we have is that for good or for ill the Commission has created a new cultural element which will very doubtless be used. The real interest in different variants is cultural. The Commission wants to popularize the currency and the symbol. The Commission is concerned about æsthetics. Æsthetics should be promoted. We don't need to register lots of glyphs in the AFII registry where people won't be using them. What we need is a best-practice advice to typographers. My company makes fonts and as you can see from ED13 I have
    taken some interest in defining the kinds of variations which are permitted, with regard to weights and proportions and so forth. I would be happy to take an action to liaise with Lazaros with regard to producing such set of a guidelines, whether alone or in conjunction with other typographers. Such an output document should be published on the web.
    Lazarus Tossounidis: In San José I gave a talk about the use of the currency in IT standardization, it is part of the Universal Character Set. If there is a deadline we have to know about it. There is no guideline as to whether the sign precedes or follows the amount.
    Marc Küster: DIN is skeptical about the need for a currency symbol at all, the need to use one is hard to convey to Germans, especially with regard to the need to implements such an expensive thing. The costs seem to bear no relation to the gain one would have to use such a new way of writing a currency sign. If it is indeed necessary, there must be no stipulation as to the design of the shape itself, which is purely a matter of personal taste.
    Erkki Kolehmainen: Michael Everson said already what I was going to say, It is our duty to review the cultural element. As standardizers we have to make it possible to enter, process, and output. And we are in no position to discuss whether there is any merit whatsoever to the value of the cultural element. It is outside our scope.
    Karl-Ivar Larsson: OCR has to be defined by a mathematical outline description, whatever happens with other fonts.
    Trond Trosterud: We cannot limit the number of allographs (a term which is a synonym to glyph) in any way.
    Mike Ksar: It is a cultural element. We have to standardize it for use. From an AFII registry point of view, I would recommend the registration of only one glyph.
    Keld Simonsen: In Denmark we are also concerned about the usability of the sign. We are looking at looking at not using it. I agree with Erkki.
    Lazaros Tossounidis: We received some input from the typographic community. There is a way of saying x varies between this minimal value and this maximal value. When you shrink it, the proportions don't work, the metrics must have values between minimum and maximum. This is simple and typographers understand it.
    Mike Ksar: for AFII registration we should send it to AFII, and ask that Commission accept feedback. Ask to register one generic glyph.
    Evangelos Melagrakis: Conclusion should be that tomorrow TC304 should assign to some of its members the task to
    assist the Commission in the registration. Secondly, there is AFII, then there is ISO 1037 for OCR-B. That OCR-B is essential.
    Wolf Arfvidson: If we talk about the sign as a symbol, there is an ISO registry for symbols, that is not in IT. There are hundreds of fonts registered with AFII.
    Mike Ksar: You must refer to ISO 10036, see clause 6.1.
    Mícheál MacConmara:Recommended output for agreed fallback.
    Chris Makemson: Project 9.2 needs to have a fallback representation for the EURO SIGN.

  • Recommendation 2.4.1: Workshop recommends to the Informatics Department of the European Commission that a set of "best-practice" typographic guidelines as regards EURO SIGN glyph variations be drawn up and made available, on the Web and elsewhere, to provide guidance to typographers for producing fonts supporting the EURO SIGN.

    3. Input: Keyboard layouts (ED10, ED16)

    3.1 Adding the EURO SIGN to ISO/IEC 9995-3

    The Canadian NB has made this proposal and sent it out on e-mail. Recommendation: TC304 should monitor this activity. 1997-09-26 document from Alain LaBonté (ED16) proposes to replace one of the ways of typing the SOFT HYPHEN in ISO/IEC 9995-3.

    Mike Ksar: Does CEN TC304 want only to monitor this proposal or to take action with regard to it and to other proposals?
    Mícheál MacConmara: De facto keyboards should be a national keyboards. The commission should not be attempting to do this on behalf of the countries. Via the subsidiarity.
    Lazaros Tossounidis: This has been initiated by the Commission according to the principle of subsidiarity as a centralized action (when centralization is an advantage it should be employed), since there are few national standards for keyboards.
    See also discussion under next agenda item.

  • Recommendation 3.1.1: Erkki Kolehmainen to provide the discussion list address for JTC1/WG9.
  • Recommendation 3.1.2: National bodies to participate in the JTC1/WG9 discussion of the Canadian proposal before its meeting in November.

    3.2 Incremental and final solutions for implementation of the EURO SIGN on national keyboards

    Lazaros Tossounidis: There was an internal group preparing information from standardization bodies, to help give guidelines and make proposals. The usual procedures within the Commission were followed. At a political level from DGII, responsible for the Euro in all its aspects. I made this proposal after evaluating what the situation was, what the similarities with other currencies were, how the industry would react to such a proposal. We contacted industry and bodies like ECMA. We investigated 100 keyboards and typewriters of over the last 35 years. We used a number of guiding principles, including the prominence that the Commission wished the Euro to have on keyboards, with very strong advice from DGIII: visibility and ease of activation were considered to be essential. Character set issues were considered, fonts, etc. were considered. Generally, level 1 or level 2 characters have a prominent place on the keyboards, but level 3 characters don't. Levels 2 and 3 are considered equivalent in terms of ergonomics and visibility. Adding a new key is not reasonable today. The £ and $ should not be replaced. The numeric keypad shouldn't be used. Costs should be minimal for hardware and software vendors.
    Trying to place the Euro on level 1 involves moving characters. This is too complex. It was not considered to be good for generations of users used to level 1 characters. There are no real underlying principles apart from that things shouldn't infiltrate the alphanumeric area. The newest keyboard these days has 104 keys.
    1. So it was determined that on most keyboards it should be on alt- E and engraved there.
    2. Alternatively, it should be engraved and accessed by for example shift-alt-2. For instance, in the UK and Ireland, alt-E produces Éé and they don't want this changed. The Commission will ask in future that keyboards shipped for procurement will contain the EURO SIGN.
    In the Commission we are supporting Windows NT.
    We are asking for feedback on this.
    Göran Tellström: There exist other systems than NT. Moving Irish letters to introduce a symbol is very difficult. We must focus on 9995 revisions and national standards.
    Mícheál MacConmara: The DOLLAR SIGN has nothing to do with currency. It is a reserved symbol in many operating systems. Most European cultures would not have used the DOLLAR SIGN as a currency sign, but for other purposes.
    Michael Everson: The concern we have in Ireland is that we get UK and US keyboards, and if you they decide to engrave on the E key, we are trapped by the a market into a keyboard which is directly inimical to the way we type now. We prefer shift-alt-4, shift-alt-5; Canada has proposed shift-alt-2.
    Lazaros Tossounidis: All the points were taken into account. I say, solemnly, if I am today here it is for the minority languages. We have taken this into account and there are three proposals. There was a lot of time spend studying this. All German keyboards have @ on the Q key engraved on the keyboard.So we are following a precedent.
    Marc Küster: DIN is somewhat disturbed by this tendency. National keyboards are falling into the domain of national standards bodies within ISO/IEC 9995. In comparison to the real multilingual market, the Commission with its 50,000 keyboards is still a small market. The problem is that neither hardware nor software currently supports the @ sign being on the Q key. Only a minority of software actually supports this.
    Mike Ksar: Are you proposing a short-term solution using alt-E, or a long term solution? In the short-term solution is the Commission's proposal similar to the Microsoft proposal?
    Graham Dixon: I agree with Michael. Nobody expects a currency sign to be on an alphabetic key. The signs µ and @ are alphabetic characters. It seems more natural that the Euro be on a numeric key. I don't see a benefit from having a large number of keyboards having it in the same position.
    Göran Tellström: Microsoft has been mentioned several times. There was some information coming from Microsoft, that was the position in July. Changes were made to that list. What is on the end of page 23 and 24 is very close to the situation.
    Michael Everson: I would support it being on alt-2, 4, 5, but not on E. The Euro is a cultural element, but my keyboard is also a cultural element. Prominence may be important, but so are æsthetics. If you engrave it on the 4 or 5 key I will think that it is great. If you put it on my E key I will consider it an offence.
    Þorvarður Kári Ólafsson: The Commission should ask for written comments by a given time to Lazaros on the keyboard issue. We could ask the commission to consider the Canadian proposal.
    Marc Küster: There is the question of the cost of printing the EURO SIGN on the E key. The cost is also in the
    software, which is the point. My second point is that it should be reiterated that it is the national bodies' responsiblity to care about national keyboards, not the Commission's.
    Lazaros Tossounidis: The Commission's proposal is made to assist the national bodies.
    Mícheál MacConmara: If it becomes known that the sign is available on some keyboards, it will generate an artificial demand for having the keyboard changed. I would also like to agree with Michael, the e-fada (Éé) is very important to us.
    Action 3.2: National bodies to consider the Commission proposal in ED10 and the Microsoft proposal, and the Canadian proposal the implications of each and to comment on it to the TC by middle of December.
    Lazaros Tossounidis: With regard to legacy systems such as OS2 and Macintosh, if we do not place the Euro in legacy character sets, there is no point in talking about putting it on the keyboard.
    Klaas Ruppel: I would like shortly to say that I don't like signs brought into the alphabetic area of the keyboard. In this Canadian proposal. they proposed to have the Euro replace the soft hyphen, which "currently has little use if any". This is not true, at all, for Finnish and many, many languages.
    Göran Tellström: If you read ISO/IEC 9995 very deeply, there are two ways of creating soft-hyphen. You can use level 3 or you can use group 2. What I don't like is the use of key E03 level 3. According to 9995 today, there were 2 ways to type the soft hyphen, one which is logical and one not. The Canadian proposal is to remove the "less logical" one.
    Michael Everson: I don't want to spend time on this Canadian proposal, the soft hyphen and claims made by Alain LaBonté are not true. The soft hyphen is not optional, and if users are used to accessing it in certain ways these should not be removed from them.
    Mike Ksar: I would like to follow Michael's proposal. TC304 should recommend to JTC1 not to pursue this proposed draft amendment.
    Þorvarður Kári Ólafsson: We can urge TC304 members to participate in the discussion of this in the JTC1/WG9.

  • Recommendation 3.2.1: National bodies to consider the Commission proposal in ED10 and the Microsoft proposal, and the Canadian proposal the implications of each and to comment on it to the TC by middle of December.

    3.3 Contact with ISO TC68 with regard to the EURO SIGN in banking terminals

    Þorgeir Sigurðsson: I contacted TC68 and they asked to circulate our press release in their newsletter.
    Mike Ksar: Do they have a standard? Do we know who the secretariat is?

  • Recommendation 3.3: Þorgeir Sigurðsson to press ISO TC68 to answer the question, do they need the EURO SIGN on banking terminals?

    4. Electronic code for information interchange (ED4)

    4.1 Addition of the EURO SIGN to ISO/IEC 10646

    This is in progress in SC2/WG2. Reference is ISO/IEC JTC1/SC2/WG2 Resolution document N1604, Resolution 33.23

  • Recommendation 4.1 still valid.

    4.2 Revision of MES and EES

    This is moving to a PT and editorial instructions do refer to adding the EURO SIGN to these repertoires.

  • Recommendation 4.2 still valid.

    4.3 Addition of the EURO SIGN in 8-bit coded character sets

    The "Latin 0" proposal is out for NWI and CD ballot. Microsoft has added the EURO SIGN to cp1252 in position x80.
    Keld Simonsen: We did also talk about the addition of the EURO SIGN to Latin 2, the new Romanian part of 8859, etc. Does anyone know about that?
    Evangelos Melagrakis: We have been engaged in an immense exercise to review all parts of 8859. I never been able to understand why it took so many years. In order to do something of benefit to a small country like mine, I am looking at the Greek part of 8859 in a more productive manner. We are looking at all the characters needed to write monotonikó Greek at least, We also decided to add the DRACHMA SIGN some empty spaces are being used; I am afraid of the other parts of 8859 being registered,we will find we have the same problems as in 8859-1. We are considering the inclusion of
    EURO SIGN in 8859 parts in general. It is a problem administratively mainly, before technical issues can even be considered.
    Karl-Ivar Larsson: The delays in revising 8859 have indeed been incomprehensible. We did talk about this in SC2.
    Evangelos Melagrakis: What will be the principles by which we will start the whole exercise? We have to consider these before we even consider thinking about adding anything to the existing pages.
    Mike Ksar: This item is a touchy subject to many, including myself. If TC304 wants to study this further, that is OK. My concern is that the impact of creating new coded character sets will have a severe impact on industry, changing software, software drivers, locales, existing systems, whether 7-bit or 8-bit. It will cost money. If the European proposal for keyboards is considered to be sane, I would welcome the Commissions's study of creating another part of 8859 just for the EURO. Messages on TC304 reflector that announce that Latin 0 would replace Latin 1. It is easier to look at it objectively. Replacing 8859-1 as the default Western European characters set will have an immense impact everywhere! I would discourage the creation of any other parts of 8859 just for the EURO SIGN. I would like to see that implementors start to use the agreed upon code n 10646 as the method of implementation of the Europe sign, using 10646 as the preferred standard for implementations. and move away from using 8-bit standards.
    Göran Tellström: If we are doing anything in 8-bits it had better be done now and quickly. If we concentrate on this we will have delays. Registration of Latin 0 is regarded as the best introduction of We have to focus on the future. An 8-bit solution for non-Window systems is needed in the near future, but only the Latin 1 countries. We shouldn't try to cover everything now.
    Lazaros Tossounidis: Currently in the Commission people will be using Word 97 using the EURO. The exchange format is Word 6. This will happen over the next 2-3 years. I see a clear need to put the EURO SIGN in legacy systems, but I want to say as a user, the industry cannot come logically and reasonably and say both black and white about the same issue. Having struggled for many years to support Unicode and 10646 I can say this. Industry always said "it was not commercial to force the industry to 10646 right away", and now the industry says that "it is not commercial to force us to support the EURO in legacy code pages". Industry wants to do what it wants in its own time.
    Þorgeir Sigurðsson: In Crete Latin 0 the CURRENCY SIGN was replaced, then that was changed to PLUS-MINUS, now it seems that the will is to change it back.
    Mike Ksar: I agree we need a quick solution for 8-bit or 7-bit. Latin 0 might be one solution, but not in an ISO standard. The quick solution which exists today is in cp1252, where there is room for the EURO SIGN. This is OK as long as we don't mess around with 8859. We need to give our attention to 10646. We don't want to create new parts like Latin 0, it diminishes the amount of energy we can spend on 10646.
    Monica Ståhl: There are other systems than Windows.
    Mícheál MacConmara: DECUS/EUROBIT has written to to the Commission -- the attitude of most vendors is that they would rather not have the sign at all; their public attitude is that it is unnecessary.
    Karl-Ivar Larsson: Note that the so-called "Windows Western Extended" coding scheme in the handout comparing Latin-0, Latin-1 is not a Microsoft document.

  • Recommendation 4.3 is still valid.

    4.4 Vendor plans on the Euro in coded character sets.

    See agenda item 4.3.

    4.5 Implementation of the EURO SIGN in 7-bit systems

    ED04 proposes that characters in ASCII be replaced in 7-bit and 8- bit code tables.
    Mike Ksar: We should not think about touching ASCII.
    Some discussion ensued, partly offline. It can be summarized thus:

    The solution proposed is untenable, and other solutions are already available to users in 7-bit environments.

    1.Use the three letter code EUR. This is the simplest solution.
    2.Use, in accordance with clause 7 of ISO 646:1981, use BACKSPACE or CARRIAGE RETURN to create the EURO SIGN out of C and =.
    3.Derive, in accordance with clause 6.4.3 of ISO 646:1981, a coded character set based on the coded character set described in ISO 646:1991. Such a derived character set would need to be registered in ISO 2375 for anyone to take any notice. Positions which can be used for this are: 40 5B 5C 5D 5E 60 7B 7C 7D 7E, or @ [ \ ] ^ ` { | } ~

  • Recommendation 4.5.1: In 7-bit environments, the 3-letter code EUR is recommended. Clause 7 or clause 6.4.3 of ISO 646:1981 can also be applied if needed.

    4.6 Implementation of the EURO SIGN in barcodes and printers

    Liaison letter to TC225 has they appointed Chris Swindon from AIM Europe to TC304. We don't need liaison with TC31, but with TC225.

  • Recommendation 4.6.1 TC304 secretariat to follow up on the EURO SIGN in barcodes with CEN/TC225.

  • Recommendation 4.6.2 Printers should be discussed at the next workshop.

    5. Other aspects

    Sorting, matching, simultaneous rendering, various platforms, writing of amounts, OCR-B, interoperability, barcoding, HTML. There was no time to address each of these specifically, but many of them were discussed during the other topics. Recommendation 5.1 is covered by 5.2-5.5 and 4.3.

  • Recommendatios 5.2-5.5 are still valid

    Go to a summary of the recommendations of this meeting.