Release 1.15 of Xaira is now available from our SourceForge Site at
http://xaira.sf.net
Click on the "Released files" link and then choose whether you want to
download the platform independent source files, or the Windows binary
release. The former should be used for Unix or MacOSX: see further BUILD
NOTES below.
This is the first release to be made available only via SourceForge: all
future releases will be freely downloadable from there. This is also the
version we will be distributing with the second edition of BNC Baby (on
which see http://www.natcorp.ox.ac.uk/babyinfo.html) when that is
released in a few weeks time.
CHANGES FROM PREVIOUS RELEASE
See the change log in CVS for a full list of changes. Most significant are:
(i) Collocations were broken in the 1.14 server: these now work again.
(ii) Substantial increase in context-sensitive (popup) help in the
Windows client. The F1 key now shows dialog help for each dialog box
rather than the popup. The source for all help (both dialog and popup)
is now maintained in one xml help file (helpfile.xml) for ease of
translation.
(iii) A new index format (24) has been introduced: all existing corpora
need to be reindexed.
(iv) The source for the Windows client is now maintained under CVS.
Indextools will follow. We plan to reorganise CVS so that the Windows
build is in a separate module. Building the Windows version from CVS is
not yet feasible because the project file makes assumptions about
library and header locations. We aim to fix this and would be very
interested to hear from any users willing to try some test builds:
Visual Studio is a requirement.
(v) The helpfile for the Windows Client has been repurposed as an online
reference manual: see http://www.oucs.ox.ac.uk/rts/xaira/Doc/refman.xml
NEW DESIGN DOCUMENT FOR CQL
Firming up the design of CQL, the underlying "corpus query language"
used by XAIRA, is the next priority. A design document for the new
version of CQL is now available at http://www.tonydodd.net/wiki and we
welcome input to it. The document also includes descriptions of searches
that users would like to make, together (occasionally) with the CQL that
would implement the search. Please have a look and add extra searches as
a challenge. There is also a note on how to use the Xairo API to solve
queries and statistical tasks that go beyond the scope of CQL.
BUILD NOTES FOR UNIX
To build the system you *must* also install both Xerces C++ 2.6 (source
from http://xml.apache.org/xerces-c/install.html) and ICU (International
Components for Unicode) version 3.2. (source from
http://www-306.ibm.com/software/globalization/icu/downloads.jsp). If you
plan to use the SOAP interface you will also need to install Apache Axis
C++ 1.4 (Source from http://ws.apache.org/axis/cpp/download.html)
Then unpack the Xaira tar ball into a directory; run the configure
script; type make; put the kettle on and have a cup of tea while
everything compiles. We have so far built the Unix system on various PCs
running Debian Linux and under MacOSX; we welcome feedback from those
trying to build it on other versions of Unix.
To check whether your build was successful, type 'make check'. This
indexes a small test corpus and evaluates a few queries. As testing
proceeds we aim to build the database of queries into a more substantial
regression testing tool.
Finally 'make install' moves all the files to their final positions (you
may need superuser privilege to do this of course). Most of the
traditional make targets e.g. make clean are also available.
BUG REPORTS
All bugs reported and fixed are now tracked on the SourceForge site.
Please use the bug reporting system there. Usability issues, enhancement
requests, and general discussion will continue to be hosted on the
(subscriber only) xara-preview mailing list (see http://www.xaira.org).
Subscribers to this mailing list will also receive project update notices.
25 July 2005
http://xaira.sf.net
Click on the "Released files" link and then choose whether you want to
download the platform independent source files, or the Windows binary
release. The former should be used for Unix or MacOSX: see further BUILD
NOTES below.
This is the first release to be made available only via SourceForge: all
future releases will be freely downloadable from there. This is also the
version we will be distributing with the second edition of BNC Baby (on
which see http://www.natcorp.ox.ac.uk/babyinfo.html) when that is
released in a few weeks time.
CHANGES FROM PREVIOUS RELEASE
See the change log in CVS for a full list of changes. Most significant are:
(i) Collocations were broken in the 1.14 server: these now work again.
(ii) Substantial increase in context-sensitive (popup) help in the
Windows client. The F1 key now shows dialog help for each dialog box
rather than the popup. The source for all help (both dialog and popup)
is now maintained in one xml help file (helpfile.xml) for ease of
translation.
(iii) A new index format (24) has been introduced: all existing corpora
need to be reindexed.
(iv) The source for the Windows client is now maintained under CVS.
Indextools will follow. We plan to reorganise CVS so that the Windows
build is in a separate module. Building the Windows version from CVS is
not yet feasible because the project file makes assumptions about
library and header locations. We aim to fix this and would be very
interested to hear from any users willing to try some test builds:
Visual Studio is a requirement.
(v) The helpfile for the Windows Client has been repurposed as an online
reference manual: see http://www.oucs.ox.ac.uk/rts/xaira/Doc/refman.xml
NEW DESIGN DOCUMENT FOR CQL
Firming up the design of CQL, the underlying "corpus query language"
used by XAIRA, is the next priority. A design document for the new
version of CQL is now available at http://www.tonydodd.net/wiki and we
welcome input to it. The document also includes descriptions of searches
that users would like to make, together (occasionally) with the CQL that
would implement the search. Please have a look and add extra searches as
a challenge. There is also a note on how to use the Xairo API to solve
queries and statistical tasks that go beyond the scope of CQL.
BUILD NOTES FOR UNIX
To build the system you *must* also install both Xerces C++ 2.6 (source
from http://xml.apache.org/xerces-c/install.html) and ICU (International
Components for Unicode) version 3.2. (source from
http://www-306.ibm.com/software/globalization/icu/downloads.jsp). If you
plan to use the SOAP interface you will also need to install Apache Axis
C++ 1.4 (Source from http://ws.apache.org/axis/cpp/download.html)
Then unpack the Xaira tar ball into a directory; run the configure
script; type make; put the kettle on and have a cup of tea while
everything compiles. We have so far built the Unix system on various PCs
running Debian Linux and under MacOSX; we welcome feedback from those
trying to build it on other versions of Unix.
To check whether your build was successful, type 'make check'. This
indexes a small test corpus and evaluates a few queries. As testing
proceeds we aim to build the database of queries into a more substantial
regression testing tool.
Finally 'make install' moves all the files to their final positions (you
may need superuser privilege to do this of course). Most of the
traditional make targets e.g. make clean are also available.
BUG REPORTS
All bugs reported and fixed are now tracked on the SourceForge site.
Please use the bug reporting system there. Usability issues, enhancement
requests, and general discussion will continue to be hosted on the
(subscriber only) xara-preview mailing list (see http://www.xaira.org).
Subscribers to this mailing list will also receive project update notices.
25 July 2005
