RE: [indic] Re: Top Nukta... and double nuktas ... and more nuktas

From: Mike Meir (
Date: Wed Aug 06 2003 - 06:59:47 EDT

  • Next message: Peter Kirk: "Re: Display of Isolated Nonspacing Marks (was Re: Questions on ZWNBS...)"

    I have been trying to track down a definition of Canonical Combining
    Class 7: Nuktas (and of the other combining classes): can anyone point
    me in its direction? A clear definition of the Canonical Combining
    Clases, would presumably form the basis of an evaulation of the
    viability of a spacing-headline-height nukta as a separate Unicode code
    The document I posted previously, which I attach again for reference,
    lists printed documents in which various placements for (I hardly dare
    say) nuktas are used, including more than one use by more than one
    author, in both India and Bangladesh, of the double nukta on Ja. The
    document was prepared for Dr Anthony P. Stone, Project Leader,
    ISO/TC46/SC2/WG12 Transliteration of Indic scripts, by Abu Jar M Akkas.
    Judging by this document the dot is found, in the case of Perso-Arabic
    transcription below, to the right, or aligned with the headline. Only in
    the first case is it non-spacing. In one case, both below, and to the
    lower right are found in the same dictionary, which suggests fairly
    strongly that there is no real difference between those two positions,
    one a spacing, the other a non-spacing, form of the dot.
    While the details of the schemes vary slightly, they are united in the
    principle that the dot does the trick: in other words, the simplest
    representation is of a Bengali Character, with a dot. There are
    personal, practical and typographic preferences for where the dot should
    be, but these are not basic.
    Solaiman, I was not suggesting that the placement of the nukta should be
    controlled in any way, nor that it is not useful, placed at
    headline/matra height, nor that it has not been used in books, but
    merely that there doesn't seem to be much of a case for making a top
    nukta an additional letter in Unicode, when you can place the dot which
    is represented by the current code point anywhere you want in relation
    to grpahemes in fancy text by constructing a font with ligatures in that
    As it is, the Nukta is listed as having General Category Mn, which is a
    Mark, Non-Spacing. It has the Canonical Combining Class 7: Nuktas. The
    Top Nukta you have identified definitely has the appearance of being
    General Category Mc, Mark Spacing Combining. Nevertheless, the
    documentation also suggests that the combining classes are not to be
    taken literally as applied to fancy text, whcih is what your scan is: an
    example of real-world, fancy, text.
    Michael, when you say that a second nukta should be stacked on top of a
    first, do you mean, in principle, in in a plain text representation only
    - i.e., one in which, symptomatically, no conjunct forms at all would be
    found? That would seem fair enough.
    The only form of the double-nuktaed Ja that I have seen does have the
    nuktas side-by-side, and was prepared by Linotype. I presume this was
    not done without some research, taking it back to the Bose instance,
    probably. However, this refleects fancy text, obviously.
    Typographically, the priority with nuktas is to place them so that they
    remain distinguishable at small sizes when other elements are combined
    within the same grapheme. Stacking ( I presume this implies one above
    the other, both remaining visible) in this instance is a bit
    counter-productive, since it inevitably results either in an increase in
    line spacing, or the danger that a further stacked element will crash
    into an element of the line below, becoming illegible. This would apply
    in both plain and fancy text.


    From: Omi Azad []
    Sent: 05 August 2003 19:27
    To: Solaiman Karim
    Cc: Paul Nelson (TYPOGRAPHY); Kenneth Whistler;

    What will be the result man?

    Solaiman Karim wrote:

    hello all

       I don't know if I misunderstood or not but someone said it is useless

    add in unicode. Someone is saying somthing which he doesn't even know

    is it he is talkign about. Are you guys saying is just made up what I

    to you that it is not only Arbic it is also use in english to translate

    some other language such as French so and so. Please let me know if I

    misunderstood you guys and it seems to me that Bangla should be limited.


    ----- Original Message -----

    From: "Paul Nelson (TYPOGRAPHY)" <>

    To: "Kenneth Whistler" <> <>

    Cc: <> <>

    Sent: Monday, August 04, 2003 7:16 PM

    Subject: [indic] Re: Top Nukta... and double nuktas ... and more nuktas



    I guess I totally misunderstood what Omi was stating then.

    It seems there are no less than 8 different ways to transliterate this



    -----Original Message-----

    From: Kenneth Whistler [ <>

    Sent: Monday, August 04, 2003 4:14 PM

    To: Paul Nelson (TYPOGRAPHY)

    Cc: <> ;

    Subject: Re: [indic] Re: Top Nukta... and double nuktas ... and more


    Paul said:



    Your proposal fails the plain text case and can only be obtained with

    higher level markup to set this notion of a "spacing Nukta". Why not

    simply encode a BENGALI SPACING NUKTA character be encoded? That

    removes any requirement for a higher level markup and allows for

    disambiuous use in plain text.

    Paul Nelson


    I don't understand this contention. What Mike was reporting is that

    there are two schools of practice for placing these nuktas on Bengali

    letters required for transcribing Arabic


      a. place the nukta (generally) under the base letter, in

         a typographically pleasing location for that particular

         base letter's shape.

      b. place the nukta (roughly) on the head line, often

         implemented simply with a full stop as an ad hoc

         solution, even if placed oddly.

    That is what Abu Jar M Akkas reported as part of the detailed analysis

    of usage.

    Particularly if (a) is the preferred approach and (b) is considered an

    unpleasing typographical hack, then surely we are simply talking about

    the already existing Bengali nukta, which, after all, was encoded

    precisely for such usage in Bengali.

    Nobody is claiming that in *plain* text a distinction has to be

    maintained between a nukta which displays appropriately underneath a

    consonant and a nukta which displays up to the right on a head line. So

    I don't see any need to

    *encode* a new character distinction here.

    Plain text should be focussed on representing the text *content*, not on

    the fine details of its display. And in this case, it seems appropriate

    to me that style or font markup should be the means for indicating any

    such distinctions in placements of a nukta.




    Omi Azad
    Altruists International <>  

    This archive was generated by hypermail 2.1.5 : Wed Aug 06 2003 - 07:42:10 EDT