gsoc-alphabetical_order

Michael Bauer fios at akerbeltz.org
Tue Apr 4 12:32:12 UTC 2017


Sgrìobh gcompris-devel-request at kde.org na leanas 04/04/2017 aig 13:00:
> I do not get the last point. Alphabetical order is not a phonetic tool, we
> do not need to order phonems [sh] but graphems s or h. We should not
> present multigraph to my point of view?
You cannot necessarily separate the two. There are a whole lot of 
characters taking a single Unicode codepoint which are technically digraphs:
ij dž nj lj dz

There are many more di or trigraphs but because of Unicode politics, 
many of these are not coded at single code points even though the 
languages themselves treat them as separate from the single letters. In 
Welsh for example, dd and ll are not listed in dictionaries under d and 
l but in separate categories i.e. you have

da
dim
dos
dda
ddi
lawen
lowr
llafar

so if have an activity and treat l as l + l, you'll end up sorting them as
da
dda
ddi
dim
dos
lawen
llafar
lowr

and there will be complaints because that will be a learning activity 
with wrong alphabetic sort order.

In the long run, it will be better if we just accept that there are 
digraphs or trigraphs and that we account for them, rather than insist 
on the outdated model of no-multigraphs and annoy a lot of users.

Michael
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/gcompris-devel/attachments/20170404/9e55a283/attachment-0001.html>


More information about the GCompris-devel mailing list