Here aer comments on the actual form submitted by Rebecca.

Comments on form: >>>> indicates comments by John Clews


The form is rather too long. It should ideally occupy only a single

There are also too many hurdles shown, and it could be made more
user-friendly. There are also potential problems of having to deal
with frivolous or malicious submissions (always possible with
linguistic or ethnic issues) which could be avoided if the form were
to be redesigned somewhat, while still maintaining the essence of
Rebecca's proposal.

The form also lacks any information on the status of the submitter,
and their relationship to the language concerned (do they use it?
have they any professional or academic role in relation to the
language for which a request is being submitted?)

There is also no place for specifying if a specific Latin script is used.

A suggested simplified form was included in the previous email
from me.

Specific comments:

> <Picture: ISO 639-2 Registration Authority - Library of Congress>
> ISO 639 Registration Form
> Request for addition or change

>>>>    Make that

        "Request for code for new language."

        The form should allow requests for additions only.
        Delete section 2. Rationale for deletion under
        section 2. below.

> Please review the criteria for registration of language codes for the
> assignment of new language codes (ISO 639-2 or ISO 639-1)
> 1. New Languages
> Name of language (English):
> (Required)
> Name of language (French):
>         Reference where found:
> Name of language (indigenous)
>         Reference where found:

>>>>    Keep this much simpler: just ask for

        Language name(s)

        It's the job of the RA, assisted by the JAC, to disentangle
        English names, French names, other names, not that of the
        submitter, who might get confused by some of the questions.

        In passing, "Name of language (English): (Required)" - when
        no other language names are required - makes this seem very
        anglo-centric in an international context.

> If in non-Roman script and supply a romanized form, indicate
> transliteration system used:

>>>>    If the submitter has come across the name in a book, or in
        various books or other documents, unless it is documented in
        those documents, there is no reason why a submitted should
        know this.

        Again, it's the job of the RA, assisted by the JAC, to disentangle
        various possible names.

> Evidence of sufficient number of documents to establish separate code
> per ISO 639-2 Annex A A.2.1 (request by one agency with 50 documents
> or five agencies with a total of 50 among them). Please cite name of
> institution(s) where documents are held and number at each. Example:
> Library of Congress (65)a: (Required)

>>>>    Here too, it's the job of the RA, assisted by the JAC, to
        look at aspects such as this in establishing what amount of
        documents may exist in such languages.

        Documents _about_ such languages are important too - perhaps
        more important in some cases, and there may well be a
        significant number of such documents.

        The same aims could be achieved by a different box to fill in:

        Relevant bibliographic references, if any
        (Documents about the language and in the language).
        Note: priority will be given to requests where this
        section is well documented,

>>>>    There have always been major problems with this
        "50 documents" approach

1.      How does either the submitter, or the RA, easily find out how
        many documents there are?

2.      Crucially, why 50 documents? I have recently done a
        statistical analysis of various figures from different
        national libraries, including the British Library and the
        Library of Congress, and it is clear that this "50 documents"
        rule has been applied retrospectively - there are many
        languages in both the libraries that I mention that now have
        language codes in ISO 639 and ISO 639-2, which would not have
        had codes allocated had this "50 documents" instruction been
        applied at the time ISO 639-2 was prepared.

        It might be an interesting exercise for Jean-Arthur Creff
        (and any others) to check how many documents he can find in
        Alsatian and in the other languages with important status in
        France which were documented in ISO 639 JAC N19. I would
        guess that it would be hard to find 50 documents for several
        of these languages, and indeed for several in ISO 639-2 that
        have codes already allocated.

        This is not to suggest that any of those languages are in
        some way "unworthy" - just to highlight the difficulties of
        the "50 documents" hurdle in practice.

3.      Speaking on behalf of CEN/TC304 (Information and
        Communications Technologies: European Localization
        Requirements) the European standards committee that I
        represented at the the ISO 639 Joint Advisory Committee
        meeting in February, there is a strong view that the amount
        of documents available is only one of several factors that
        should be taken into consideration, and NOT the predominant
        one, given that many languages which already have codes in
        ISO 639-2 do not themselves meet this criteria. There are
        various other factors which may affect the numnber of
        documents, e.g. official status or otherwise in a country; or
        the current economic conditions in the country/countries

4.      Fitness of purpose in ISO standards - not just for libraries

        As Haavard Hjulstad wrote on Monday 6 Mar 2000, via
        <[log in to unmask]>:

> > PLEASE remember: ISO 639 (both parts) are !!!NOT!!! solely for
> > library use: It should be sufficient to document that there is a
> > need within other application areas. Within terminology there clearly
> > is a need. Within software localization there clearly is a need.
> > Within SOME libraries I would expect that there is a need...

>>>>    It is the view in CEN/TC304 that this criterion may need
        amendment when ISO 639-2 is reviewed, and probably revised,
        at or before its next 5-yearly review.

>>>>    Given this situation, when CEN/TC304 sends in its next block
        of requests for codes for languages which have some
        importance within Europe, CEN/TC304 will not expect that
        evidence on a particular number of documents will be all that
        important. The status of some languages as having official
        status implies that almost certainly there are likely to be
        a significant number of documents in that language, without
        needing to provide evidence.

        The suggested dialogue box which says
        Relevant bibliographic references, if any
        (Documents about the language and in the language).
        Note: priority will be given to requests where this
        section is well documented,
        is a way to get over this problem, and to get around problems
        that CEN/TC304 raised earlier when it previously requested

> Additional information (may include number of speakers):

>>>>    Change to

        Additional information (e.g. Official status, and where;
        or _estimated_ number of speakers):

> Request addition in:
>  - ISO 639-2 only
>  - ISO 639-1 and ISO 639-2
> (Note that ISO 639-1 is a subset of ISO 639-2)
> Code suggestion (Note that there is no guarantee that these codes
> will be assigned):
> 2. Changes

>>>>    All of section 2 should go, for the reasons stated above.
        It will lead to too many requests for changes, which can only
        get a standard reply such as "no more changes are made" which
        will only lead to dissatisfied submitters, who may take up
        their dissatisfaction elsewhere.

        Raising expectations that cannot be fulfilled only leads to

        There is also the opportunity for malicious use, e.g. to
        suggest names which might go through which turn out to be
        insulting to that language community, which the the ISO 639
        Joint Advisory Committee fails to spot. We should not be a
        hostage to fortune.

        NB: if there are any problems with existing codes or existing
        names of languages, the people concerned will find ways to
        contact us, particularly if email contact is provided,
        whether via [log in to unmask], or some named person attached to
        the ISO 639 Joint Advisory Committee.

>>>>    Regarding the phrases below, the same comments apply as in
        section 1. Most should go or be simplified.

> Name of language (English): (Required)
> Name of language (French):
> ISO 639-2/B code: (Required)
> ISO 639-2/T code: (Required)
> Indicate specific change requested including justification and any
> references: (Required)
> Note: Codes that have already been used are generally not changed.
> Language names may be changed when well documented.

>>>>    Language names are not a normative part of the standard in
        any case, it is only the language codes that are normative.

        NB: because of that it may be useful to state on the form:

        "Use of any form of name does not imply any endorsement or
        otherwise by ISO of a particular form of name, just that it
        is in general use in some areas."

> ------------------------------------------------------------------------
> Return to:
> •ISO 639 Home Page •
>  ISO 639-1 Registration Authority---coming soon •
>  ISO 639-2 Registration Authority •
>  Library of Congress Standards •
>  Library of Congress Home Page

>>>>    There should also be a chance to seek approval to join the
        <[log in to unmask]> list. I regularly approve new additions to
        the <[log in to unmask]> list. Haavard Hjulstad is currently
        the one who approves addition of new people to
        <[log in to unmask]>.

> ------------------------------------------------------------------------
> <Picture>Library of Congress
> Comments: [log in to unmask] (03/08/2000)

>>>>    I assume that "Comments: [log in to unmask] (03/08/2000)"
        relates to web-design issues, rather than to the
        ISO 639 Joint Advisory Committee, and would not be included
        in the final request form.

Best regards

John Clews

John Clews, SESAME Computer Projects, 8 Avenue Rd, Harrogate, HG2 7PG
tel: +44 1423 888 432; fax: + 44 1423 889061;
Email: [log in to unmask]

Committee Chair of  ISO/TC46/SC2: Conversion of Written Languages;
Committee Member of ISO/IEC/JTC1/SC22/WG20: Internationalization;
Committee Member of CEN/TC304: Information and Communications
 Technologies: European Localization Requirements
Committee Member of TS/1: Terminology (UK national member body of
 ISO/TC37: Terminology)
Committee Member of the Foundation for Endangered Languages;
Committee Member of ISO/IEC/JTC1/SC2: Coded Character Sets