[mpeg-OTspec] Re: [OpenType] RE: Proposal: deprecate ReqFeatureIndex

[Behdad Esfahbod]

Thanks John.  You summarized my stance very nicely.

Peter, I'm not suggesting that people invent arbitrary tags.  I'm saying that
the tag used does not matter as it is not every read for any legitimate
purpose.  As Khaled pointed out, Fontforge for example already uses ' RQD' for
that purpose, and I just confirmed that I see some fonts having that feature.

behdad

On 08/23/2012 12:50 PM, John Hudson wrote:
>  
> 
> Peter Constable wrote:
> 
>> So, are you suggesting that people invent arbitrary tags to use in that case?
> 
> That is how I understood Behdad's suggestion and Eric's comment on it,
> and also why Behdad suggested that this could be done in such a way that
> the arbitrary tag wouldn't 'pollute the feature tag space'. If I
> understood Behdad correctly, the ReqFeatureIndex doesn't seem to require
> that the indexed required feature be associated with the language system
> alongside the usual layout features: it simply needs to reference an
> enumerated feature in the font.
> 
> The general principle is that the feature referenced by ReqFeatureIndex
> doesn't need to be a particular registered feature, but only needs to be
> a collection of lookups that the font developer wishes to be mandatory
> for a particular language system. Whether that feature is associated
> with the language system via the feature tree or only via the
> ReqFeatureIndex is an implementation question.
> 
> If it turns out to be necessary for such a feature to have a tag that is
> associated with a language system in a typical way, i.e. if Behdad's
> method is unworkable, and if you are concerned about arbitrary tag
> names, then one could easily define a standard approach to naming such
> tags, e.g. replicating the language system tag as a feature tag.
> [Indeed, one could have devised the functionality of ReqFeatureIndex by
> such a tag naming methodology: if a feature tag exists that matches the
> language system tag name to which it is associated, treat it as a
> required feature.]
> 
> JH
> 
> -- 
> 
> Tiro Typeworks www.tiro.com
> Gulf Islands, BC tiro at tiro.com <mailto:tiro%40tiro.com>
> 
> The criminologist's definition of 'public order
> crimes' comes perilously close to the historian's
> description of 'working-class leisure-time activity.'
> - Sidney Harring, _Policing a Class Society_
> 
>