Difference between revisions of "Archiphonemes"

From Apertium
Jump to navigation Jump to search
m
 
(4 intermediate revisions by 2 users not shown)
Line 7: Line 7:
 
* If the archiphoneme has a range of default surface forms (even if rarely subject to deletion), it should be written in upper case, e.g. <code>{A}</code>
 
* If the archiphoneme has a range of default surface forms (even if rarely subject to deletion), it should be written in upper case, e.g. <code>{A}</code>
 
* If the archiphoneme is always deleted, it ''may'' consist of more than one character, e.g. <code>{dup}</code>. This is, however, advised against.
 
* If the archiphoneme is always deleted, it ''may'' consist of more than one character, e.g. <code>{dup}</code>. This is, however, advised against.
  +
  +
==Common archiphonemes==
  +
  +
  +
==Frequently asked questions==
  +
  +
; Why use {C} and not ^C ?
  +
  +
<pre>
  +
<spectie> was thinking about {A} over ^A
  +
<Flammie> good
  +
<spectie> and worked out a nice argument for it (aside from pure aesthetics):
  +
<spectie> other programs (e.g. morphological segmenters) parsing the output with {A} don't need to know about multicharacter symbols
  +
<spectie> compare:
  +
<spectie> foo{A}z{A}l
  +
<spectie> foo^Az^Al
  +
<spectie> with the first you know where the symbol ends
  +
<spectie> in the second you do not know
  +
<spectie> it may be ^Az and ^Al or ^A z ^A l
  +
</pre>
   
 
[[Category:Terminology]]
 
[[Category:Terminology]]
  +
  +
[[Category:HFST]]
  +
[[Category:Writing dictionaries]]
  +
[[Category:Documentation in English]]

Latest revision as of 16:22, 26 September 2016

Guidelines[edit]

  • Archiphonemes should be a single character.
  • Archiphonemes in lexc should be encased in { and }.
  • Archiphonemes should be declared in the Multichar_Symbols section in the header of the file after the grammatical tags, with a comment giving their possible forms.
  • If the archiphoneme is subject to deletion, it should be written in lower case, e.g. {s}
  • If the archiphoneme has a range of default surface forms (even if rarely subject to deletion), it should be written in upper case, e.g. {A}
  • If the archiphoneme is always deleted, it may consist of more than one character, e.g. {dup}. This is, however, advised against.

Common archiphonemes[edit]

Frequently asked questions[edit]

Why use {C} and not ^C ?
<spectie> was thinking about {A} over ^A
<Flammie> good
<spectie> and worked out a nice argument for it (aside from pure aesthetics): 
<spectie> other programs (e.g. morphological segmenters) parsing the output with {A} don't need to know about multicharacter symbols 
<spectie> compare: 
<spectie> foo{A}z{A}l
<spectie> foo^Az^Al
<spectie> with the first you know where the symbol ends
<spectie> in the second you do not know
<spectie> it may be ^Az and ^Al or ^A z ^A l