[Egyptian] Repeating operators

[Stéphane polis]

Hi guys,

After the fruitful discussions we had today with the UTC, let me quickly support the proposal sent by Serge a few minutes ago (before going back to my ‘holiday’ week).

It is certainly most economic and ‘expandable’ system proposed so far for horizontal and vertical groups. We would have — most of the time — a very simple syntax of the type p*t:pt (supported by Bob), but we could use **, :: instead of the ‘group joiners’ (proposed by Bob), and ***, :::, etc. whenever needed for additional levels of embedding. It might be that not all the fonts or software solutions need to support more than three or four levels, but at least, we have a single and coherent system (explicitly advocated for by the UTC members during the discussion today, who urged us not to develop several HLPs): only two operators, and the syntax of repeating these operators would make the trick for several levels of embedding. 

In practical terms, it’s equivalent to Bob's proposals regarding joiners and group joiners, except that we know how to expand the syntax in a coherent way for additional levels of embedding (that are well attested) and we won’t need to add several types of ‘groups joiners’ later on. I think that an agreement could be made on this during the meeting tomorrow. 
At least, I would definitely support an agreement that goes in this direction.

The ‘INSERT’ operator(s) could function along the same lines: low level of precedence with respect to ‘:’ and ‘*’, but higher than ‘::’ and ‘**’, allowing for the insertion of groups in ‘corners’ (see the example in the document sent by Serge). However, this needs to be tested in order to make sure to keep everything unambiguous (as Mark-Jan pointed out to me) and we should certainly refrain from any hasty decision on this point. Let’s prepare a document that tackles the various cases discussed so far for the next deadline. Better good than fast.

Have a good evening,

Stéphane

------------------------------------------------------
Chercheur qualifié F.R.S.-FNRS

Université de Liège
Service d'Égyptologie
Département des sciences de l’Antiquité
Place du 20-Août,
B-4000 Liège

http://www.egypto.ulg.ac.be
------------------------------------------------------



> Le 4 août 2016 à 21:06, Serge Rosmorduc <rosmord at gmail.com> a écrit :
> 
> Hello,
> 
> Just to clarify what I said (near to a pint of cider a month ago, and during the discussion yesterday).
> 
> 
> <repeatOperatorProposal.pdf>
> 
> A few additional comments:
> 
> Regarding the INSERT-BOTTOM-LEFT control character, I would favor splitting it in two versions. 
> Basically, if we consider the graphical order, this control is somehow ambiguous, for it’s a mix of INSERT-BOTTOM and INSERT-LEFT.
> 
> With the current interpretation :
> 
> - seen as something like INSERT LEFT, A INSERT-BOTTOM-LEFT B suggests that A comes before B in the linguistic stream (case of <PastedGraphic-1.pdf>)
> 
> - seen as something like INSERT BOTTOM, A INSERT-BOTTOM-LEFT B, with B above A, suggests that B comes before A in the linguistic stream. (case of <PastedGraphic-2.pdf>).
> 
> As we have all remarked, the first case is often attested with birds, and somehow points out to the linguistic order A-B (but not always)
> 
> The second case, attested with <PastedGraphic-3.pdf>, is much more regular, and points out to the « main » sign being first.
> 
> On a practical level, due to the size of the gap, it’s also the one where the lack of a proper mechanism hurts most. I guess using <PastedGraphic-4.pdf> instead of <PastedGraphic-5.pdf> does not hurt scholars sensitivity as much as <PastedGraphic-6.pdf> instead of <PastedGraphic-7.pdf>.
> 
> So, I might favor the use of INSERT-BELOW in this case.
> 
> 
> As an aside, INSERT-TOP-RIGHT suffers from the same problem, but I feel it’s only problematic when rendering vertical texts (where it’s sometimes unclear if the 
> inserted text belongs to the end of the previous quadrant or to the current one).
> 
> Best regards,
> 
> Serge
> 
> 
> _______________________________________________
> Egyptian mailing list
> Egyptian at evertype.com
> http://evertype.com/mailman/listinfo/egyptian_evertype.com