Release of Xaira version 1.14

xiaoz

永远的超级管理员
Staff member
Release 1.14 Available

The major change in this version is the switch to index format 23, with these changes: (i) support for thread file in several pieces (absence of this support was limiting corpus size to 500M words);
(ii) support for large indexes, that is, indexes for corpora with more than 4G words. This involves widening a couple of index fields to 64 bits and
(surprising to say) has not yet been tested;
(iii) 20% reduction in uncompressed index size;
(iv) redesign to reduce working space required by indexer. We also fixed a bug that only affected corpora of 1M+ words but caused disastrous memory leaks after that many words had been processed. This version has been used to index a corpus of approaching 1G words.

Because of the change in index format all corpora must be reindexed.

See the Download section for the download link.


[本贴已被 作者 于 2005年06月13日 01时07分35秒 编辑过]
 
there is a problem with the index version with this release. do not download it yet until I have tested and make a new posting.
 
We have fixed the bug with index version. Now you can download Release 1.14 from SourceForge.

http://sourceforge.net/project/showfiles.php?group_id=130289&package_id=142832&release_id=334435

Note: Collocations are broken in this release and will remain so until the client is based on the new server. If you want to use the collocation functionality, please stick to Release 1.13. Or if you to try and test new features of Release 114, download it from the above link.

[本贴已被 作者 于 2005年06月14日 04时12分21秒 编辑过]
 
This release like all earlier releases, is fully functional. You can use the indexer tool to index your own corpora (in Unicode [utf8 or utf16] for non-Latin scripts such as Chinese). The corpus must be in XML. You can use the preprocessor in the indexer tool to add minimal XML markup to your data. If you like, you can also use the indexing wizard to help you. An indexed corpus can be explored using the client.

[本贴已被 作者 于 2005年06月14日 22时41分46秒 编辑过]
 
Got it, but you told me last year some time that the program was commercial. ???
 
Extra funding was granted to make Xaira a system independent indexing and retrieval tool package. So not only is the binary for Windows freely available but the source codes are also made available at SourceForge for compilation for use on Unix/Linix and Mac machines. Good news isn't it?
 
Back
顶部