Re: Run-time checking of fonts for Sinhala support

From: Doug Ewell (
Date: Mon Sep 07 2009 - 12:08:15 CDT

  • Next message: verdy_p: "Re: Visarga, ardhavisarga and anusvara -- combining marks or not?"

    "Harshula" <harshula at gmail dot com> wrote:

    >> Users obviously want text in their language to be displayed
    >> correctly.
    > I agree. How do we get there? There's probably more than one approach
    > that can be undertaken in parallel to achieve that goal.

    Standardization may or may not be the answer here. Users of Sinhala
    might simply reject a system that doesn't display Sinhala correctly out
    of the box. Vendors that ship a so-called "Sinhala-ready" system that
    includes a compliant font, but doesn't use it, might pay dearly in terms
    of public derision and lost sales.

    > Firstly, I actually said "should", not "MUST": "Operating systems
    > should, at a minimum, check that a Sinhala font meets these
    > requirements before using it."

    OK, if this is explicitly a recommendation and not a requirement, and if
    it doesn't say that the OS has to check *each time*, that might be
    reasonable. But see below about the intent of the standard as written.

    > 'Operating systems shall, to the extent of their capabilities, only
    > recognise Level 1 compliant fonts as Sinhala fonts.'

    It might be better to state explicitly that fonts compliant to Level 1
    *or above* are recognized, if that's what you mean, even though Levels 2
    and 3 are defined elsewhere as supersets of lower levels. I think
    Roozbeh may have mentioned this already.

    > The intent is that if the operating system already checks font
    > compliance then it must adhere to Level 1 compliance for the aspects
    > it can/does check. The other option is to change the "shall" to a
    > "should". i.e. change it to a recommendation.

    I'm actually not seeing any place where the standard requires the OS to
    check font compliance, only a stipulation that "One Sinhala font" [or
    presumably more] "shall be provided with the computer system."
    Providing is not checking. I understand that you are proposing to amend
    the standard, but you have to start with what the standard says;
    otherwise your statement about checking font compliance at all is itself
    part of your proposal.

    It seems to me, upon reading the standard a bit more carefully, that the
    lack of normative detail about having the OS check fonts at runtime
    might be intentional and not a loophole. Compare this to the detailed
    descriptions of compliant fonts and the installation and uninstallation
    process. SLSI may have considered a runtime font-checking mechanism to
    be an unnecessarily expensive burden on the OS and explicitly decided
    not to address the matter. If they are doing their job, they will
    require you not only to provide your additional wording, but also to
    justify why the standard should regulate or even mention this behavior
    when it did not do so before.

    However, as a side note, I have to say I'm a bit skeptical of a
    standardization review process that includes, in the section "Visibility
    of Symbols," the requirement "The keyboard shall be comfortable to the
    user." (Where does an ergonomic requirement fit here? How is
    "comfortable" defined? What does this have to do with engraving symbols
    on keytops?)

    Doug Ewell  |  Thornton, Colorado, USA  |
    RFC 5645, 4645, UTN #14  |  ietf-languages @

    This archive was generated by hypermail 2.1.5 : Mon Sep 07 2009 - 12:11:11 CDT