Difference between revisions of "Roadmap"
Jump to navigation
Jump to search
| Line 6: | Line 6: | ||
** Multi-stage [[chunk]] transfer |
** Multi-stage [[chunk]] transfer |
||
*** aren't we already doing this? |
*** aren't we already doing this? |
||
** [[Ideas_for_Google_Summer_of_Code/Weighted_transfer_rules|Weighted transfer rules]] |
|||
** Parser (e.g. [[Ideas for Google Summer of Code/Robust recursive transfer]]) |
** Parser (e.g. [[Ideas for Google Summer of Code/Robust recursive transfer]]) |
||
* <code>apertium-transfer-training-tools</code> and <code>apertium-tagger-training-tools</code> to multi-stage transfer. |
* <code>apertium-transfer-training-tools</code> and <code>apertium-tagger-training-tools</code> to multi-stage transfer. |
||
Revision as of 18:47, 15 February 2015
This page gives a roadmap of features which we hope will be available in future Apertium releases.
Apertium 4.0
- Transfer
- Multi-stage chunk transfer
- aren't we already doing this?
- Weighted transfer rules
- Parser (e.g. Ideas for Google Summer of Code/Robust recursive transfer)
- Multi-stage chunk transfer
apertium-transfer-training-toolsandapertium-tagger-training-toolsto multi-stage transfer.- Discontiguous multiwords
- Multiwords with inflection on both parts, circumfix inflection (like Georgian verbs) without overanalysing
- ?
Apertium 3.5
- TMX support
- We have some TMX support, but stuff like User:Pankajksharma/Patcher would be nicer
- Orthographic normalisation
- Noun-compound resolvation
- We do compound analysis (though where longest-match doesn't give the right answer, we have to lexicalise)
- Migration of
apertium-transfer-training-toolsto unicode - Port to Mac OS/X
- Isn't this done long time ago?
- Port to Windows
- People have done this as well, though no one uses Windows so who knows if it still works.
