Changes to Chatter

* 1.0-alpha-189 (February 26, 2024)
** XML Schema updated to version 2.20.2
Add more CA elements:
   ['hurried start', "\x{2907}", "HurriedStart"],
   ['sudden stop', "\x{2906}", "SuddenStop"],
   ['hardening', "\x{2051}", "Hardening"],

** Chatter
Add more CA elements

* 1.0-alpha-188 (February 15, 2024)
** Chatter
rename %moru to %umor and %grau to %ugra

* 1.0-alpha-187 (February 15, 2024)
** Chatter
add %moru and %grau tiers for universal dependencies

* 1.0-alpha-186 (February 9, 2024)
** Chatter
Allow Unicode 2013 in %x... dependent tiers.
Add the date built to the version information in "About".

* 1.0-alpha-185 (December 15, 2023)
** XML Schema updated to version 2.20.1
in element "p", remove "blocking"

** Chatter
no longer support initial ^ for "blocking" because redundant
since ≠ introduced in 2016

* 1.0-alpha-184 (August 14, 2023)
** Chatter
Add "-time" to output nanoseconds elapsed.
Remove "heritage" from @Options (was removed from XML Schema in 2.14.0

* 1.0-alpha-183 (March 30, 2023)
** Chatter
*** Fix retrace checking to ignore being in group
<ingroup [/]> [<] after

is now legal and does not require retraced material to be
in the group

*** Change "&nonword" to new syntax "&~nonword"
*** Remove "notarget" from @Options
*** Remove [x N] repetition
*** Fix unnoticed bug in round trip involving form suffix such as "-s"
The suffix was being dropped by xml2chat.

** XML Schema updated to version 2.20.0
*** Change some word attributes:
"fragment" for "&" to "nonword" with new CHAT syntax "&~"
"incomplete" for "&+" to "fragment"

*** Remove "notarget" from @Options

*** Remove "r" element for [x N] repetition

* 1.0-alpha-182 (March 24, 2023)
** Chatter
Ensure that if there is retraced material, then it must be followed
by content.

* 1.0-alpha-181 (December 21, 2022)
** XML Schema updated to version 2.19.0
Change segment-repetition element into a variant of ca-delimiter
called "repeated-segment", in order to provide substructure.

** Chatter
Allow substructure in the parsing of repeated segments inside words.

* 1.0-alpha-180 (December 15, 2022)
** Chatter
Allow languages to be any three letter code.

* 1.0-alpha-179 (October 3, 2022)
** Chatter
change syntax for non-local URL for @Media, by requiring surrounding in
double quotes (because of comma ambiguity).

* 1.0-alpha-178 (September 7, 2022)
** XML Schema updated to version 2.18.0
Allow general URL for @media name.
** Chatter
Allow general URL for @media name.

* 1.0-alpha-177 (August 4, 2022)
** Turn off bullet time sequence checker for %wor
Bullet time sequence is only relevant for main tier.

* 1.0-alpha-176 (July 13, 2022)
** XML Schema updated to version 2.17.0
Add new element "wor" which acts as a dependent tier "%wor" parsed
like the main tier.
** CHAT to XML and XML to CHAT support new "wor"
** Internals
Update many dependencies.
Use source and target Java 11.

* 1.0-alpha-175 (October 9, 2020)
** Fix CHAT to XML
There was an error in the past 18 years on the handling of
Gem label matching checking, which resulted in incorrect
behavior only in the case of @Bg and @Eg labels that are
identical except for possible spaces in the labels.

* 1.0-alpha-174 (October 9, 2020)
** Fix XML to CHAT
*** Regression error
A global error was inadvertently introduced in the previous version
in the case of no existing @Types header.
*** Blocked segments
1.0-alpha-134 (August 25, 2016) introduced an error
losing information during XML to CHAT of blocked segments

* 1.0-alpha-173 (October 8, 2020)
** Update send2clan feature for double-clicking error messages
*** Works with CLANc on macOS Catalina
macOS Mojave seems to have a problem communicating with CLAN

** Fix XML to CHAT generation of @Types
Previous version (1.0-alpha-172) forgot to generate @Types from XML.

** Update legal 3-letter language codes

** (Not user-visible) Update and clean up Java dependencies
And remove use of deprecated APIs where feasible.

We are still building with no later than Java 11, because for the JAR
file to be double-clickable on Mac, we still have to build with Java
11.

* 1.0-alpha-172 (March 9, 2020)
** Chatter
Add @Types header for CHAT

** XML Schema updated to version 2.16.0
Remove unused mk element from w.
Add optional CHAT attributes DesignType, ActivityType, GroupType.
Add @Types header.

* 1.0-alpha-171 (January 31, 2020)
** Chatter
Enforce strict order for the pre-@Begin optional headers
Add @Window optional header

** XML Schema updated to version 2.15.0
for @Window

* 1.0-alpha-170 (September 28, 2019)
** Chatter
Fix regression bug in 169 that resulted in always re-processing files
in tree mode even when they have not changed.

* 1.0-alpha-169 (September 27, 2019)
** Chatter
To work around macOS silently creating "._filename" files on various
file systems, Chatter will ignore them when searching for files to
process.

* 1.0-alpha-168 (September 24, 2019)
** XML Schema updated to version 2.13.1
Add a new participant role Therapist

** Chatter
*** pho and sin groups
Remove feature added in 1.0-alpha-130: pho and sin groups are no
longer treated as implicit word groups.

*** Fix xml2chat roundtrip bugs dating back to the beginning
**** Word groups
Always output explicit word groups in case of a single pho or sin
group or pause.
**** Pho groups
Always output explicit main tier pho group in case no words are inside
it

*** Ignore "heritage" CHAT files in tree mode rather than signal errors
Have Chatter completely ignore "heritage" CHAT files rather than
try to parse them and report errors.

* 1.0-alpha-167 (September 17, 2019)
** XML Schema updated to version 2.14.0
Remove "heritage" option since the XML generated was nonsensical all
along. Remove corresponding "blob" element.

** Chatter
Disable parsing of "heritage" documents and exit without generating
output.

* 1.0-alpha-166 (September 13, 2019)
** XML Schema updated to version 2.13.0
Allow Date to be optional
** Chatter
*** Allow days in age to be 31

* 1.0-alpha-165 (August 26, 2019)
** Chatter
Allow more languages to be legal.

* 1.0-alpha-164 (August 9, 2019)
** Chatter
Allow Unicode 2013 in %mor line.

* 1.0-alpha-163 (July 31, 2019)
** Chatter
Fix bug introduced in version 162 where continuation lines in %mor are
not handled right when there are translations.

* 1.0-alpha-162 (July 29, 2019)
** Chatter
Allow hyphen in mor translation segments, e.g.

%mor:  co|danke=thank-you .

* 1.0-alpha-161 (June 20, 2019)
** XML Schema updated to version 2.12.0
*** Clean up documentation links
*** Inline superfluous Schema types
*** Remove quotation2
*** Remove other elements and attributes that have never been used in CHAT

** Chatter
*** Update according to XML Schema changes

* 1.0-alpha-160 (May 15, 2019)
** XML Schema updated to version 2.11.0
*** Remove @Interaction Type
** Chatter
*** Relax restrictions for CA
Reallow (word) as well as old "xx" constructs
*** Tighten restrctions for replacements
Cannot have omitted or untranscribed material as word replacement.

* 1.0-alpha-159 (May 13, 2019)
** XML Schema updated to version 2.10.0
*** Remove "@fs" from list of form markers

*** Remove @Exceptions
*** Remove @New Language
*** Require @Page to have numerical argument

*** Remove %lan dependent tier

*** Language codes must come from a fixed set defined by CLAN ISO-639.cut
*** Speaker ID must be only ASCII characters

** Bug fixes
*** Tighten up checking of main tier words
There must be at least some spoken content inside every word.
*** Fix failing round trip of XML to CHAT for words with suffix and mor
*** Disallow &frag [: replacement]

** Allowed characters
Update based on latest https://github.com/ghedlund/phon/raw/master/ipa/src/main/resources/ca/phon/ipa/features/features.xml

* 1.0-alpha-158 (April 28, 2019)
%mor has removed generation of quotation mark material bq|bq and eq|eq,
so Chatter has been modified accordingly to no longer expect that material.

* 1.0-alpha-157 (April 22, 2019)
** XML Schema updated to version 2.9.0
Remove "sign" from @Options, add "dummy"

* 1.0-alpha-156 (April 18, 2019)
** XML Schema updated to version 2.8.2
More participant role changes.

* 1.0-alpha-155 (April 16, 2019)
** XML Schema updated to version 2.8.1
Add new participant roles.

** Words may contain more characters
/

** Long feature labels may contain more characters
such as _ and /

** Fix long-standing omissions in participant role checking
@ID header must have a role.
Also, the role must match what was listed in @Participants.

* 1.0-alpha-154 (March 11, 2019)
** XML Schema updated to version 2.8.0
*** Remove "caps" from @Options
Relax former restrictions on characters in words, so that everything
behaves as though the old "caps" options were always turned on.
** Support inverted Spanish question and exclamation marks in dependent tiers
\u00bf and \u00a1

* 1.0-alpha-153 (January 25, 2019)
** XML Schema updated to version 2.7.2
*** @Options
Add "notarget".

** @Media header checking
Reverse the restriction from 1.0-alpha-149 that disallowed bullets in case
of "missing". Now, "missing" may have bullets again.

Fix missing restriction that "notrans" disallows bullets.

* 1.0-alpha-152 (December 20, 2018)
** Add -allowAnyMediaName command line option
This allows relaxing the constraint added in 1.0-alpha-148 that enforces
media name matching CHAT file name.
** Remove -validate, add -noValidate command line option
XML validation is now the default and no longer needs to be specified.
Use -noValidate to turn off XML validation. This is really only useful
for debugging Chatter.

* 1.0-alpha-151 (October 22, 2018)
** Java 8 allowed
Inadvertently version 150 was compiled to work only with Java 11.

* 1.0-alpha-150 (October 19, 2018)
** Schema update to version 2.7.1
Add new @Videos header.

* 1.0-alpha-149 (October 5, 2018)
** Major changes to @Media header checking
*** If "missing" or "unlinked" or "notrans", cannot have bullets in CHAT file
*** If there are no bullets in CHAT file, must have "unlinked" or "missing" or "notrans"

** Rebuilt using Java 11

** macOS Note: send-to-CLAN feature does not currently work on macOS Mojave
This is under investigation.

* 1.0-alpha-148 (September 7, 2018)
** Enforce that media name matches file name

* 1.0-alpha-147 (September 6, 2018)
** Fix regression bug introduced in 146 that deleted and rebuilt all targets constantly
** Fix bug in -outputDir command line option
An absolute path is handled correctly now.

* 1.0-alpha-146 (September 5, 2018)
** Improve performance by ignoring .git directories
** In tree mode, delete all extraneous output in the destination directory
For example, in the past, if a CHAT file name1.cha resulted in
generation of name1.xml and was renamed to name2.cha, then name1.xml
would persist in the next run of Chatter; it is now deleted. Chatter
now takes full control of a destination directory's complete contents.
** XML Schema version bumped to 2.7.0
A new @Media option, "notrans", is provided to complement the existing
"missing" and "unlinked" options.

* 1.0-alpha-145 (August 6, 2018)
** Fix error-reporting omission
For the command-line version of Chatter, if there are any errors,
display a summary message "There were N files with errors."

* 1.0-alpha-144 (June 26, 2018)
** Change to age field in @ID
Force leading zero in month and day, e.g., "5" must now be "05".
Years are unaffected.

* 1.0-alpha-143 (May 3, 2018)
** Fix roundtrip XML to CHAT
*** echolalia
This has been inadvertently broken since echolalia was introduced in
1.0-alpha-135.
*** filler, incomplete
Inadvertently broken since introduced in 1.0-alpha-136.
** XML Schema version bumped to 2.6.1
No substantial changes, just http to https where appropriate for links.

* 1.0-alpha-142 (March 23, 2018)
** Fix faulty JAR file
Chatter has not been able to run from XML to CHAT on Java 9 or Java 10
since 141 because of building using Java 8.

Rebuild using Java 10 targeting Java 8 and higher.

* 1.0-alpha-141 (March 15, 2018)
** XML Schema version bumped to 2.6.0
Add "mor exclude", CHAT is [e]

* 1.0-alpha-140 (August 30, 2017)
** Change list of languages supporting numbers
zho
cym
vie
tha
nan
yue
min
hak

* 1.0-alpha-139 (June 9, 2017)
** XML Schema version bumped to 2.5.0
*** Remove inlined dependent tiers [%act], [%sch], [%sdi], [%sxx]
*** Fix documentation links
** Chatter lexer tightened
Constructs such as [= explanation] now require that single space, e.g.,
[=explanation] is no longer accepted (it was in the past).

* 1.0-alpha-138 (June 7, 2017)
XML Schema version bumped to 2.4.0

** Remove @pm
** Remove num from @Options
** Remove @EndTurn

* 1.0-alpha-137 (May 23, 2017)
XML Schema version bumped to 2.3.0
** Change segment repetition from a CA delimiter to its own element with text attribute
** Add "CA-Unicode" option
** TODO XML Schema documentation
Some URLs are currently broken, work in progress because of reorganization of CHAT and MOR manuals.

* 1.0-alpha-136 (November 4, 2016)
XML Schema version bumped to 2.2.2
** Add "&-word" (filler)
similar to "&word", participates in pho
** Add "&+word" (incomplete)
similar to "&word", participates in pho

* 1.0-alpha-135 (September 30, 2016)
XML Schema version bumped to 2.2.1
** Add form marker @e for echolalia

* 1.0-alpha-134 (August 25, 2016)
XML Schema version bumped to 2.2.0
** Update Phon IPA
make sure not to include space!!

** Turn CA delimiter "blocked segments" 2260 into a CA element

* 1.0-alpha-133 (June 29, 2016)
Add two CA delimiters:

2260 not-equal-to sign to mark "blocked segments" as in ru≠b-b-b≠bber
21AB leftwards arrow with loop to mark "initial segment repetition" as in ↫b-b-b↫boy

* 1.0-alpha-132 (April 29, 2016)
Version 131 fix was wrong. Really fix the bug now.

* 1.0-alpha-131 (April 25, 2016)
Fix bug in handling pho and sin groups that are annotated.

* 1.0-alpha-130 (April 18, 2016)
Allow pho and sin main groups to behave as annotated groups without
needing superfluous group brackets <>.

* 1.0-alpha-129 (February 5, 2016)
** Phon IPA
More changes to Phon IPA.

** Mac app
Removed entirely because it was unreliable and hard to maintain.

* 1.0-alpha-128 (February 3, 2016)
** Mac app
Working now.

Also, note that Chatter now requires use of Java 8.

* 1.0-alpha-127 (February 2, 2016)
** Schema
Update to version 2.1.0, removing Id attribute of CHAT and changing otherSpokenEvent to take
string not word element.

Update Phon IPA characters according to latest https://github.com/ghedlund/phon/raw/master/ipa/src/main/resources/ca/phon/ipa/features/features.xml

* 1.0-alpha-126 (September 1, 2015)
** Schema
Update to version 2.0.5.

Update participant roles.

* 1.0-alpha-125 (March 24, 2015)
** Schema
Update to version 2.0.4.

Allow phonological brackets on main line without %pho or %mod.

* 1.0-alpha-124 (August 14, 2014)
Allow more Unicode characters in %com and other places.

* 1.0-alpha-123 (June 9, 2014)
** %pho and %mod additions
For Phon, allow "^" and "\u0335" characters in %pho and %mod and also
"(", ".", and ")".

This is an interim new feature until the XML Schema properly
represents the new elements.

To use pauses in %pho and %mod, use the pho group characters around
pauses in the main line:

*CHI:	cache ‹(.)› cache . 612716_615716
%pho:	ˈkæ (.) ˈkæ
%mod:	ˈkaʃ (.) ˈkaʃ

* 1.0-alpha-122 (April 1, 2014)
Added new optional @PID header.

* 1.0-alpha-121 (March 11, 2014)
Fixed regression bug resulting in failure to launch CLAN on double-click of
error message.

* 1.0-alpha-120 (December 13, 2013)
** mor
Added support for IPA characters in %mor words.

* 1.0-alpha-119 (December 5, 2012)
** Integration with CLAN
Because of problems in Mountain Lion, do not launch CLAN automatically
from Mac OS. The user will have to manually launch CLAN.

* 1.0-alpha-118 (September 4, 2012)
** CHAT
xxx@a forces inclusion of mor 1-1 content
where xxx does not

* 1.0-alpha-117 (August 6, 2012)
** XML Schema update to 2.0.2
"addition" form marker
** CHAT
Added new form marker @a for "addition".

* 1.0-alpha-116 (July 13, 2012)
** XML Schema update to 2.0.1
Added support for nested quotations, using Unicode 2018 2019.

Removed morTagMarker and morQuotation because normal character stems
are now used instead.

** Chatter
On %mor, removed special punctuation for comma, tag markers,
quotation.

Added check that quotation marks are matched.

* 1.0-alpha-115 (July 12, 2012)
** XML Schema update to 2.0.0
*** Comma
Comma has been moved out of separatorType to tagMarkerType, to have
MOR information attached.

<s type="comma"> in old documents need to be changed to <tagMarker type="comma">

*** Quotations
<quotation> has been flattened and split into <quotation type="begin">
and <quotation type="end"> for the sake of MOR.

The new CHAT syntax replaces ["] with “ and ” (Unicode 201C and 201D).

** Phon characters
Updated from features.xml

* 1.0-alpha-114 (June 15, 2012)
** XML Schema update to 1.9.9
moved word's untranscribed attributes to "untranscribed" rather than
"type"

word types such as "unintelligible" are now parsed by Chatter into the
XML as such rather than as ordinary words.

Prevent xxx, etc. from being considered for 1-1 with %mor line.

** CHAT
Remove xx and yy from the Schema and CHAT

* 1.0-alpha-113 (May 11, 2012)
** Regression bug in packaging
send2clan.dll for Windows is now included again in JAR file.

* 1.0-alpha-112 (March 1, 2012)
** XML Schema update to 1.9.8
Add "custom field" to ID line
** CHAT
Note that all old CHAT documents will need to have @ID line augmented
with an extra vertical bar in order to be valid.

* 1.0-alpha-111 (February 1, 2012)
** XML Schema updated to 1.9.7
Added "real" attribute to replacement element, to indicate whether the
annotated word was real.

So CHAT now has the [:: replacement words] annotation as distinct from
[: replacement words]

* 1.0-alpha-110 (January 30, 2012)
Fix bug in XML generation when both %mor with translation and %gra exist.

* 1.0-alpha-109 (January 5, 2012)
Fix regression bug in handling form markers that broke @n.

* 1.0-alpha-108 (January 3, 2012)
** XML Schema updated to 1.9.6
Change to schema to allow "-s" and other suffixes after a form marker.
Specifically for a@l-s

** %com and other dependent tiers now allow all IPA characters

* 1.0-alpha-107 (December 31, 2011)
** XML Schema updated to 1.9.5
CHAT breathy voice character changed

Changed from 03BE to 264B.
03BE is now treated as the Greek character it really is.

* 1.0-alpha-106 (December 20, 2011)
** XML Schema updated to 1.9.3
mor translation is now supported for preclitic and postclitic also.

* 1.0-alpha-105 (October 5, 2011)
** XML Schema updated to 1.9.3
*** separated-prefix
Support # for word to be treated as prefix for Hebrew:
new word attribute "separated-prefix".
*** legacy
Remove the "legacy" option, finally.

* 1.0-alpha-104 (September 30, 2011)
** CHAT form marker misparsed
Fix form marker lexing so that invalid form markers are detected
rather than misparsed.

* 1.0-alpha-103 (September 29, 2011)
** XML Schema updated to 1.9.2
*** New stress characters
02C8 for primary stress
02CC for secondary stress

*** Support for new dependent tier %sin
There are new grouping brackets 3014 and 3015 for the main and %sin
tiers.

* 1.0-alpha-102 (September 20, 2011)
** Fix mor 1-1 bug
"&frag [: replaced]" was not being handled properly with %mor line.

** Internal
Update to ANTLR 3.4

* 1.0-alpha-101 (July 12, 2011)
** XML Schema updated to 1.9.1
*** CA
Added "Breathy voice" CA character 03BE.
** Other CHAT changes
*** User-defined dependent tiers
Restricted total number of characters to 7 (including the initial "x").

** Internal updates
Updated to ANTLR 3.4b4, JAXB 2.2.4.

* 1.0-alpha-100 (June 23, 2011)
** XML Schema updated to 1.9.0
Added word attribute user-special-form.
** CHAT
Now allow word@z:code

* 1.0-alpha-99 (June 16, 2011)
** Phon compatibility
Updated pho allowed characters from features.xml from Phon.
** CHAT
Now allow + in speaker ID.

* 1.0-alpha-98 (May 23, 2011)
** XML Schema updated to 1.8.3
Added participant role CallTaker.
Added "anonymized" to Transcription.

* 1.0-alpha-97 (May 18, 2011)
** XML Schema updated to 1.8.2
Avoid network access (since sometimes unreliable) of xml.xsd

by specifying a catalog file src/main/resources/catalog.cat

mapping "http://www.w3.org/2001/xml.xsd" to "xml.xsd"

* 1.0-alpha-96 (May 2, 2011)
** CHAT
*** Languages
Added support for more Chinese languages: nan, min, yue, and zho*

** xml2chat
Check that the date att, if not default, has a matching @Date header.

** Scripts as filters
Allow Chatter to be run as a filter from standard input to standard
output when run from the command line.  For example:

$ java org.talkbank.chatter.App -inputFormat cha -outputFormat xml < input.cha > output.xml

* 1.0-alpha-95 (April 8, 2011)
** CHAT
Do not allow initial spaces on header or tier lines.
** Intel Leopard support
Rebuilt the CLAN-communicating native code against old 10.5 SDK
instead of 10.6 SDK.

* 1.0-alpha-94 (February 21, 2011)
** CHAT @Number header
Changed @Number header info to require arabic numerals for 1 through 5.

* 1.0-alpha-93 (February 13, 2011)
** pho error message fixed
pho error messages involving words with more than one letter are now
giving correct location information.

* 1.0-alpha-92 (January 14, 2011)
** XML Schema updated to 1.8.1 (changes to CA)
*** CA symbol 224B
renamed from "TCU continuation" to "technical break TCU continuation"
*** CA symbol 2248
renamed from "no break continuation" to "no break TCU continuation"
*** CA symbol 2261
renamed from "latching" to "uptake"

* 1.0-alpha-91 (December 22, 2010)
** XML Schema updated to 1.8.0 (major change)
<model> and <actual> now have element content inside <pw> elements instead of
being in <value>, and <pw> contains <ph> and <wk> and <ss>

Changed to jaxb:version="2.1" and added explicit "simple" binding mode
directive.

Made sure that <xs:annotation> has default namespace as XHTML.
<xs:annotation xmlns="http://www.w3.org/1999/xhtml">

** Fixed language marker regression bug
@s with no arguments was not being parsed correctly, thanks to
bug introduced in alpha-90.

* 1.0-alpha-90 (December 13, 2010)
** Fixed packaging
Packaging failed to pick up new JAXB-generated classes properly,
so xml2chat was failing.
** Upgraded to JAXB 2.2.3
** Error handling for @s in CHAT
More graceful.

* 1.0-alpha-89 (December 9, 2010)
** XML Schema updated to 1.7.1
Allow <pg> to have <actual> and <model> in any order.

* 1.0-alpha-88 (December 9, 2010)
A major change to the XML Schema for Phon.

** XML Schema updated to 1.7.0 (major change)
*** Phon
<pho> has been split to <model> and <actual>, and value is now element
rather than attribute.  Other Phon-relevant changes.

*** role
Participant role is now required, and restricted.

** Bullet checking
If @Media is "unlinked", then bullets are not allowed.
** Java version information
"About" box now includes Java version information.

** xml2chat
Phon <align> and <sb> are now allowed, but signal a warning that they
are not yet supported for conversion to CHAT.

** Implementation upgrades
*** Upgraded ANTLR version to 3.3

* 1.0-alpha-87 (October 29, 2010)
** CHAT
Added @T for thumbnail header.
** XML Schema updated to 1.6.2
Added thumbnail header.

* 1.0-alpha-86 (October 13, 2010)
** CHAT
Allow words such as "d'Eoin", where the rule is that a capital letter
is allowed after a letter-apostrophe combination.
** Internal development
Now using Maven 3.0 for builds.

* 1.0-alpha-85 (October 5, 2010)
Fixed a regression bug resulting in empty CHAT output.

* 1.0-alpha-84 (September 15, 2010)
This is a substantial release, with many fixes and changes.
XML will need to be regenerated.

** GUI application changes
*** Output validation
Changed default to perform output XML validation instead of omitting it.
There are plans for other output validation as well.
*** Input validation
When unmarshalling input XML, validation is now turned on.

** CHAT changes
*** %xpho, %xmod
Now allow the bracket symbols 2039 and 203A, as hacks to support
temporary change of %pho and %mod to %xpho and %xmod to turn off all
parsing.

** XML Schema updated to 1.6.1
*** xs:float replaced with xs:decimal
For start, end times and pause length.
*** Fixed key and keyref constraints on CHAT
Full namespaces were put in to properly create XPath addresses.
Utterance id was changed from "key" to "unique".
*** Nonempty strings
Various attributes have been changed from xs:string to nonemptyStringType.
For example, speaker role.
*** Changed Media type from xs:anyURI
Set up allowed media reference pattern.
*** Other type changes
Other miscellaneous type changes.


* 1.0-alpha-83 (August 30, 2010)
** Renamed chat2xml to chatter
This affects all distributed application and JAR names and scripts.

Also, internal Java package and class names were changed as well.

* 1.0-alpha-82 (August 17, 2010)
** Cleaned up allowable IPA characters
*** On %pho tier
We now use http://www.acssoft.de/world-keyboard.com/ipatables.html
as the definition of allowable IPA.
*** UNIBET on main line
Also uses the IPA tables.

* 1.0-alpha-81 (August 10, 2010)
** %pho/%xpho characters
*** Updated from Phon features.xml
*** More IPA Unicode allowed
Found an extra source of IPA tables to extra legal Unicode characters
from, including 203F.

* 1.0-alpha-80 (August 4, 2010)
** XML schema updated to 1.6.0
Added "bullets" option.
** CHAT features
Added "bullets" option to turn off bullet time sequence checking.
** TRN translator complete
For Santa Barbara corpus.
Error messages are prefixed with "0" for communication with CLAN editor.

* 1.0-alpha-79 (July 19, 2010)
** XML Schema updated to 1.5.9
<media> element has been disambiguated into <media> (after utterance
terminator, or in dependent tier) and <internal-media> (inside
utterance before terminator).
** Serializing to CHAT
*** Part of speech in word
Fixed roundtrip of word$pos.
*** Bullets
Internal bullets are now generated with continuation line after them
in CHAT.

* 1.0-alpha-78 (July 14, 2010)
** Language lexing
Allow zho-min.

* 1.0-alpha-77 (June 27, 2010)
** XML Schema updated to 1.5.8
*** Enforced three letter language codes
*** More Schema documentation added to link to chat.html

** Added more supported languages
Various languages zho-* are now recognized along with zho.
** CHAT
Updated allowed "@Interaction Type" and "@Number" options.

* 1.0-alpha-76 (May 20, 2010)
** XML Schema updated to 1.5.7
*** Changed more documentation
*** Removed form marker @m

** Removed infinite loop bug
An infinite loop bug resulted from the removal of "/", "//", and "///"
prosody support inside words in 1.0-alpha-75 because the "/" character
was not being recognized and was being retried rather than rejected.

* 1.0-alpha-75 (May 19, 2010)
** XML Schema updated to 1.5.6
Big cleanup of XML schema, removing unused material, consolidating others.

First step at cleaning up documentation, inserting references to the
CHAT manual, which is now online as HTML generated from the master
Word document and placed along side the PDF version.  The HTML version
of the manual is now available as
http://childes.psy.cmu.edu/manuals/chat.html.

The schema is now processed by xsddoc from Filigris
(http://www.filigris.com/products/docflex_xml/xsddoc/)
to generate HTML documentation now available at
http://talkbank.org/software/xsddoc/index.html

*** Restored @Page
The @Page header was removed in 1.0-alpha-37, but has been restored.

*** Removed some dependent tiers
%pht
%ssy
%spe
%ton
%rom

*** Removed %mor error markers

*** Removed 0* and 00 prefixes for words on main line

*** Removed various form markers
@ap
@pr
@sc

*** Removed some old prosodic markers
/
//
///

*** Many other miscellaneous removals

** CHAT language changes
Changed two-letter language codes to three-letter codes in entire
database, and no longer support the two-letter codes where the
three-letter codes exist (ISO 639).

*** Removed special lexical support
zh-min
zh-guoyu
zh-yue
vi

*** Changed from two letters for special lexical support
zho
tha
cym
deu
heb

* 1.0-alpha-74 (March 30, 2010)
** GUI progress bar
A bug in the progress bar has been fixed, in which "redo" resulted in
the progress bar not being reinitialized back to 0.
** Duplicate dependent tiers
Duplicate dependent tiers for a single utterance are now detected and
flagged as erroneous.

* 1.0-alpha-73 (March 25, 2010)
** %gra and %grt are now fully supported
Dependent tiers that were temporarily named %xgra and %xgrt are now
fully supported.

* 1.0-alpha-72 (March 16, 2010)
Radical changes to <mor> structure.
** CHAT language change
Allow part of speech at the end of a word on the main line for
optional part of speech (for use by mor).  The syntax is

$category or
$category:subcategory or
$category:subcategory:subcategory (multiple subcategories allowed)

** XML schema updated to 1.5.5
Completely removed <mwg>.  The main space-delimited group is now just
under <mor>.

Changed <pos> to remove <err> and <omitted>.  Moved <omitted> into
an attribute of <mor>, since it is supposed to apply at that level.

Preclitics and postclitics are now <mor-pre> and <mor-post> children
of <mor>.

Mor translations <menx> have been moved out of <mw> to <mor>.

<pos> may now also be present in <w> (see CHAT language change).

** Bug fix in %mor omission
In %mor, 0cat... had always been parsed incorrectly and silently,
generating incorrect XML, but now is parsed correctly.

** GUI improvement
The Open dialog now selects by default the directory last opened
(whether in the current session or saved from the last launch of the
application).

* 1.0-alpha-71 (March 9, 2010)
Radical changes in schema
** CHAT language change
In addition to formerly supported @s and @s:otherlang, we now also
allow

@s:lang1&lang2 for a word that could be interpreted as either in lang1 or lang2

@s:lang1+lang2 for a word whose stem is in lang1 and suffix is in lang2

In either case, also more than two languages can be specified.

** XML schema updated to 1.5.4
<w> no longer uses the xml:lang attribute for language, because we
have introduced the possibility of multiple languages per word.  A new
element <langs> has been created, which has a child of <simple>,
<multiple>, or <ambiguous>.

* 1.0-alpha-70 (March 4, 2010)
** XML schema updated to 1.5.3
Added mor replacement (@replacement in %mor/%trn line).
** CHAT
New feature: mor replacement, which occurs after any optional translations.

Example:

%mor:   part|kick-PROG@pushing .

* 1.0-alpha-69 (March 1, 2010)
** Empty %mor line
1.0-alpha-68 removed support for an empty %mor line and crashed upon
encountering it.  This has been fixed.

* 1.0-alpha-68 (March 1, 2010)
Major changes in schema affecting all transcripts with %mor.

** XML schema updated to 1.5.2
Major change: terminators <t> now can take <mor> children.  The reason
for this change is to start including the %mor terminator explicitly
inside <mwg> inside <mor>.

Introduced <mt> for "mor terminator".

Various other schema cleanups, including putting back CA "unmarked
ending" that was removed in 1.0-alpha-33.

** CHAT
*** CA "unmarked ending" supported again
*** %pho characters updated
Regenerated legal %pho characters based on phon's features.xml.


* 1.0-alpha-67 (February 9, 2010)
** Added %pho and %mod character
Unicode 2026 is now supported.

* 1.0-alpha-66 (February 5, 2010)
** Updated Phon features for %pho and %mod
Set of Phon features had not been regenerated since 5-22-08.
Now updated based on latest features.xml in Phon.
** Error message for length mismatches in %pho and %mod, and %mor and %trn
An omission was made in the fix from 1.0-alpha-65, and has been fixed.
An analogous set of fixes were made for %mor/%trn error messages.
** Underscore in %xpho, %xmod
Allow underscore "_" as a character in %xpho and %xmod.

* 1.0-alpha-65 (February 4, 2010)
** Mac app universal support fixes
Various cleanups were made so that on the Mac, 64-bit or 32-bit mode
is used as appropriate, and Java 5 or 6 (or higher).

LSArchitecturePriority is now used instead of the obsolete JVMArchs.
This fixes various problems in running chat2xml on old Intel Macs,
as well as running chat2xml as a jar.

Added -x86_64 to the Universal binary for send2clan.

** tmplib cleaned up
Launching CLAN from the GUI error message window resulted in a
"tmplib" directory being created and left behind in the directory
containing the chat2xml application.  The behavior has been changed to
use instead the temporary directory given by the java.io.tmpdir
property, if the user has not already set the java.library.tmpdir that
mx-native-loader wants.

** Windows bug fix
An important bug was fixed that caused chat2xml to fail to delete
empty XML files upoon error.  This bug also had the side effect of
causing the "Redo" button to be useless, forcing users to manually
remove the 0-sized XML files.

** Error message for %pho and %mod
Now that %mod is supported in the same way as %pho, an error on it is
noted as an error in %mod rather than in %pho.

** Memory of last directory processed
The last directory processed is now remembered across sessions, so
that a fresh launch of chat2xml will prefill the text field containing
the path of the last directory in the previous session.

Note: for some reason, the Mac OS app version of chat2xml is not yet
saving the preference.  This is being investigated.  The Mac OS "Open"
dialog box, however, has always saved the position of the last
directory opened.


* 1.0-alpha-64 (February 2, 2010)
Important: CHAT syntax now has changed, so all transcripts that used
%pho, %mod, or %npho must be rechecked and modified as appropriate.

** XML schema to version 1.5.1
Removed legacy %pho and %mod support.

For the first time, transcripts with both %pho (formerly %npho) and
%mod are correctly represented in the XML.

** CHAT
Removed legacy %pho and %mod dependent tiers.

%pho and %mod are now fully parsed, so legacy %pho and %mod must be
renamed to %xpho and %xmod, and %npho renamed to %pho.

** Java version
Removed the unnecessary restriction (introduced in 1.0-alpha-61) of
running on Java 6, so now Java 5 can again be used.

** Date formatting
Date formatting when going from CHAT to XML or vice versa was failing
unexpectedly in France, etc., because of locale settings.  chat2xml
now explicitly uses English regardless of system locale setting.

* 1.0-alpha-63 (January 29, 2010)
xml2chat has been completed!

** GUI
User interface now allows choice of input format (CHAT or XML) and
output format (CHAT or XML).  There are therefore four possible
input-to-output conversions.  Note that CHAT to CHAT or XML to XML are
roundtrips going through the ANTLR tree as an intermediate internal
format.

** Performance improvements
A memory leak in the use of StringTemplate for generating output has
been removed.

* 1.0-alpha-62 (January 25, 2010)
Radical changes to the XML schema.

** Schema updated to 1.5.0
*** Cleaned up form markers
Removed <f> element altogether.
Shifted <f> type attribute to "ftype" attribute in <w>.
Removed <fplural> hack, turned it into character content in <w>.
*** Text before form marker is now treated as a word
Previously, there were restrictions on the text before a form marker.
Now, the whole unit is considered a first-class <w>.

** CHAT changes
"a@l-s" is no longer supported.  The plural of "a@l" is simply "a-s".

** Progress bar or text
Fixed a bug in which int overflow occurred if the amount of data being
processed exceeded int range.  Now we use long instead.

* 1.0-alpha-61 (December 15, 2009)
** Schema updated to 1.4.9
Ongoing cleanup.
Fixed eventType definition; created action and happening elements.

Note: because of XML changes, old CHAT files will have to be
reconverted to XML.
** Allowed CUJK symbols, punctuation in %ort
** Upgrades to libraries
Upgraded to JAXB 2.2, etc.
** Java 1.6
Moved from Java 1.5 to 1.6.

* 1.0-alpha-60 (November 19, 2009)
** Schema updated to 1.4.8
*** Allow tag/vocative in %mor
** Allow tag/vocative in unparsed dependent tiers

* 1.0-alpha-59 (November 18, 2009)
** Schema updated to 1.4.7
Changed tagMarkerType to allow two kinds, tag and vocative.
Corrected legacy hack of treating ",," as separator.
%mor 1-1 is changed accordingly.
** CHAT extended with tag and vocative characters
201E is tag, 2021 is vocative
** Internals
Reverted back to using CommonTreeNodeStream, and used dupTree to
resolve old parser tree problems.

Cleaned up code to take advantage of upgrade from ANTLR 3.1.1 to 3.2.

Noticed that {terminatorIsOptional} was incorrectly not used as a
gated semantic predicate.

* 1.0-alpha-58 (November 5, 2009)
** Upgraded to ANTLR 3.2
From ANTLR 3.1.1 (ancient) to 3.2
antlr3-maven-plugin to 3.2
StringTemplate to 3.2.1
** Capital letters
*** Welsh
Allowed capital letter after '

* 1.0-alpha-57 (October 31, 2009)
** Capital letters checking for Hebrew
Now allow ~ and ' before a capital letter in the middle of a word.

* 1.0-alpha-56 (October 30, 2009)
** Enforce CHECK's restrictions on capital letters
Capital letters in words are now allowed only in the following cases.
*** initial letter in a word
*** after _ (treated as word boundary)
*** after - (treated as word boundary)
*** after + (treated as word boundary) if German (de)
*** if @Options includes "caps"
*** after a prefix (from last word boundary) listed in CLAN's depfile.cut
"Mac", "Mc", "De", "La", "Las", "Los", "San", "St", "D'", "O'"
*** word matches one of the patterns listed in @Exceptions
Patterns can use any combination of literals and globs "*" which stand
for any (optionally empty) sequence of literals, e.g., "anException"
or "*N*".
*** Fixed bug in handling words
Underscore was not being treated as part of the stripped-down version
of a word for END_WORD internal token.

** Miscellaneous
*** Allow Welsh (cy) to have words with numbers
A hack.
*** Allow Hebrew (he) to have a capital letter after a glottal (Unicode 0294)

* 1.0-alpha-55 (October 29, 2009)
** XML schema updated to 1.4.6
*** New @Media option: unlinked
"unlinked" will mean that a transcript has media associated with it
but has not yet been linked to it.
*** Option "caps" supported
"caps" was inadvertently not added to the schema in the last chat2xml
release.

** New @Participants role: Partner

* 1.0-alpha-54 (October 28, 2009)
** Internal ANTLR AST change
Compound words now have END_WORD with the separator included in the
text.
** Bullet time start checking fixed
Bullets on dependent tiers should not be involved in checking time
start sequence.
** Compound words with form markers
Form markers @c and @n and @sl now are allowed for compounds.
** Letters
@l can only be used with a single letter.
** New @Options: caps
"caps" is now used to indicate that capital letters may be allowed in
a word.  Without "caps" or "sign", capital letters are no longer
allowed.

Note: capital letter checking to be compatible with CHECK is currently
completely turned off, awaiting further testing and modification.

* 1.0-alpha-53 (October 25, 2009)
** Schema updated to 1.4.5
Removed support for annotations to pauses.

Removed %: dependent tier.

Removed +"" (quotation assigned to other speaker).

Removed form markers @cue, @inf, @ins.

** Changed misleading %mor/%trn/%npho 1-1 error messages
The error messages were changed to simply inform of a mismatch in
lengths of tiers.

** Bullet checking added
Bullets are now checked for self-validity (start time earlier than end
time).

Bullets are now checked for consistency in the transcript (start time
of a particular bullet must be later than the start time of the
previous bullet)

** Comma-space
Require space after comma in @Languages, @Participants, @Options, @Media

* 1.0-alpha-52 (October 17, 2009)
** Participant roles updated
Added some more roles, and in sync now with CLAN's depfiles.cut

* 1.0-alpha-51 (October 16, 2009)
** Transcript must end right after the first @End
Previously, chat2xml erroneously simply ignored all text after the
first @End
** Enforced CLAN's depfile.cut restriction on participant role
** Relaxed restriction on participant name
** Every participant must now have an @ID line

* 1.0-alpha-50 (September 24, 2009)
Corrected lexing of special characters that were being erroneously
treated as regular word letter characters:  "laugh in word" and
"yawn".

Old XML files that are bad will need to be regenerated, or hand-edited
to change the Unicode of "laugh in word" to

<ca-element type="laugh in word"/>

and "yawn" delimited material must be changed to

<ca-delimiter type="begin" label="yawn"/>stuff<ca-delimiter type="end" label="yawn"/>

* 1.0-alpha-49 (August 27, 2009)
Corrected internal error upon parsing a nonempty "education" field of
@ID.

* 1.0-alpha-48 (August 26, 2009)
** chat2xml
Had forgotten to remove parsing support for removed inlined dependent
tiers add, int, sit, gpx, spe.

* 1.0-alpha-47 (August 26, 2009)
** Radical changes in XML schema, to version 1.4.4
Changed schema so that <a> is only used for utterance level
annotations, not for group level; new <ga> element for group level
annotation (including scoped symbols).

*** Added groupAnnotationType, groupAnnotationTypeType.

*** Removed some inlined dependent tiers
add, int, sit, gpx, spe

** chat2xml syntax change in @Media
For consistency, a comma is now required to separate the fields, e.g.,

@Media: filename video missing

must now be

@Media: filename, video, missing

* 1.0-alpha-46 (August 1, 2009)
** @L1 of
Removed incorrect restriction of language specified (alpha 45 required
it to be listed in @Languages)

* 1.0-alpha-45 (July 31, 2009)
Schema and chat2xml changes.
XML schema updated to version 1.4.3.
** Schema 1.4.3
Added attributes "birthplace" and "first-language".
Removed obsolete phoKey key and keyref.
Changed speaker ids from ID and IDREF to just string.
** New CHAT headers
@Birthplace of
@L1 of
** Phon "+" (7-28-09)
Supported in %npho line, passed into pho "value" attribute for now.
** Speaker IDs may now start with numeral
For example, "123" is now legal.

* 1.0-alpha-44 (July 20, 2009)
A radical change was made to the parsing of and the XML representation
of phonological fragments ("&frag"), to treat the "frag" part as fully
equivalent to an ordinary word, and therefore having all the
associated substructure.

All old XML documents using phonological fragments need to be
converted accordingly.

** XML schema updated to version 1.4.2
*** <italic>
Has been put back into chat2xml and the schema, for the sake of Heritage.
*** <e type="fragment">
Phonological fragments are now <w type="fragment"> with special
characters now parsed into substructure and obeying the constraints of
the current language.
*** <mediaPic>
Now allowed in dependent tiers.

* 1.0-alpha-43 (June 16, 2009)
Retraces now work in all contexts properly with pho 1-1.

* 1.0-alpha-42 (June 15, 2009)
** CHAT changes for pho
Groups are now fully supported on the main line.

* 1.0-alpha-41 (June 15, 2009)
** CHAT changes for pho
Added pho group begin (Unicode 2039) and end (Unicode 203A) for main
line and the %npho line.

Removed use of # from the %npho line, replacing with the new group
begin/end instead.

Groups are not fully implemented yet in the main line, but will be
soon.

* 1.0-alpha-40 (June 12, 2009)
** Major changes to %npho parsing
*** %npho now uses space to separate words
*** %npho now uses # to separate groups
Groups are not yet supported in the CHAT main line, but will be
eventually.  Right now, each individual word on the main line is
treated as a group containing a single word.

** Changes to allowable %npho phones
Was allowing A-Z before, but now restricting to C, G, V, X, S

* 1.0-alpha-39 (June 11, 2009)
** Schema updated to 1.4.1
*** removed <bold>, <italic>
*** rid "line" option
*** Various cleanups, refactoring
** $ in %npho
We are now temporarily (?) allowing $ to appear in %npho.

* 1.0-alpha-38 (June 1, 2009)
Note: this version breaks compatibility with XML files that have
compound words in %mor.
** Schema updated to 1.4.0
*** Created morphemicCompoundWordType
*** Allowed prefixes for a compound word

* 1.0-alpha-37 (May 29, 2009)
This is a major release altering the XML schema, updating it to
version 1.3.9.  Old XML files will no longer be valid if they used
certain headers or attributes, or more than one language.

** Header change of @Coder
@Coder has been replaced with @Transcriber and is no longer supported.
** Removed "@Language of" header
** Added headers to schema
@Interaction Type
@Number
@Recording Quality
@Transcription
@Blank
** Removed unused headers from schema
@Media (we have been putting into attribute instead)
@Education
@Code
@Page
@Script
@Keywords
** Removed unused CHAT attributes
Coder
Duration
Situation
Warning
** Cleaned up various language attributes
Renamed languageType to languagesType and made it no longer a string,
but a list of xs:language, hence fully checked, and space-separated
rather than comma-separated

* 1.0-alpha-36 (May 11, 2009)
** XML output
Fixed a localization problem in generation of floats for media
bullets.  Floats are now generated always with "." and never with ",".

* 1.0-alpha-35 (May 5, 2009)
** Schema changes to 1.3.8
*** Cleanup
Made media unit required.

Removed unused enumerations.

Unified various internal element definitions of the same names, moving
them out and using refs.

*** Options attribute
Added "sign".

*** CA
Changed double-wavy latching (subword element) 2248 to 2261

Change the function of the current double-wavy 2248 to a terminator
("no break continuation")
like the already existing triple-wavy 224B.

Added corresponding double-wavy continuer (with +).

** Roundtrip
Roundtrip is still being completed.

Now using JAXB 2.1.11.

* 1.0-alpha-34 (April 16, 2009)
** Fixed CA mode handling of missing terminator
In CA mode, a missing terminator should be OK, even without a bullet
following it.
** Fixed generation of multiple options into XML
Now space-delimited as required by schema.

* 1.0-alpha-33 (April 16, 2009)
** Fixed regression bug introduced in alpha-25
Quotations ["] now work properly.
** CHAT, XML schema change
XML schema updated to version 1.3.7.

*** Options
Added "CA" back to options.  This enables terminators to be optional
in utterances.

Added "line" to options.  This is purely for CLAN display of bullets,
just as "multi" was, and has no semantic significance.

*** CA separators, terminators
Removed CA separators added in alpha 30.
Changed 5 arrow CA terminators into CA prosodies.
Removed continuing terminators introducted in alpha 31.
Removed unmarked ending.

*** TCU pause/bullet
Removed TCU pause and bullet; these concepts are now folded into the
utterance.  Old transcripts must be changed to put any terminator
after the TCU pause, if there is one.

* 1.0-alpha-32 (April 8, 2009)
** CHAT, XML schema change
XML schema updated to version 1.3.6.
Added otherSpokenEvent.  New CHAT syntax is

&*WHO:word

* 1.0-alpha-31 (April 5, 2009)
** CHAT, XML schema change
XML schema updated to version 1.3.5.
Added continuing terminators parallel to the six special CA terminators.

* 1.0-alpha-30 (April 3, 2009)
** CHAT, XML schema change
XML schema updated to version 1.3.4.
Added separators parallel to the five special CA terminators.

* 1.0-alpha-29 (March 31, 2009)
** Subword handling
Allow a single hyphen to be a subword.

* 1.0-alpha-28 (March 31, 2009)
** CHAT, XML schema change
Added TCU continuation linker.
XML schema updated to version 1.3.3.
** Subword handling
Now allow hyphens to begin subwords.

* 1.0-alpha-27 (March 26, 2009)
** Library upgrades
Upgraded to JFlex 1.4.3
** XML generation
More "@Options: heritage" fixes.

* 1.0-alpha-26 (March 23, 2009)
This is an important release fixing bugs in XML generation.
** XML generation
*** Multiple @Date headers were getting overwritten by the first
Now all the information is preserved properly.
*** Fixed generation of retraces (regression bug)
*** Heritage control characters were corrupted
Fixed, and also changed and improved Heritage generation in any case.

* 1.0-alpha-25 (March 16, 2009)
** Fixed %mor handling
A main line word with both an explanation and a retrace was
incorrectly treated as non-ignored.

* 1.0-alpha-24 (March 11, 2009)
This is a major release, because of an incompatible CHAT format
change.  chat2xml will no longer accept CHAT files in the old media
bullet format.

File names no longer go into each bullet, because now, only one media
file is allowed per CHAT file.

Now there is a single @Media header that lists the media file for the
transcript.

Any transcripts that reference more than one media file will have to
be split up, or the media files joined into one.

** Major CHAT format change:  media bullets
*** @Media header after @ID (and optional @Birth of and @Lang of) headers
The allowed syntax:

@Media: name audio
@Media: name video
@Media: name audio missing
@Media: name video missing
*** Audio/video bullets no longer have a file name or type attached to them

** Huge number of XML schema changes, to version 1.3.2
*** Dependent tiers
**** %cod is now "coding"
**** Dependent tiers %pse, %fan, %wor have been removed
**** Dependent tiers %x... now have "flavor" attribute with "x" removed
**** %mor, %trn
<s> can now include an optional <mor> because of "tag question"
separator.

*** Media
**** <media> element has been stripped of "href" and "type" attributes
**** New <mediaPic> element introduced
**** <CHAT> "Mediatypes" attribute introduced, for "video", "audio", "missing"
*** Miscellaneous cleanups
**** <CHAT> "Options" attribute is now space-delimited and checked
**** Other schema cleanups

* 1.0-alpha-23 (February 12, 2009)
** Fixed critical regression bug in XML generation
Scoped annotations were not being generated into the XML.
** Fixed ampersands in happenings, e.g., &=A&P
Was being incorrectly parsed as &=A &P
** Added check to enforce @ID for each participant

* 1.0-alpha-22 (January 19, 2009)
** fixed bugs resulting in internal errors during XML validation
** updated schema to 1.3.1 for <duration> cleanup
** check whether a participant exists when id encountered in utterance
** fixed regression bug
existing good XML files were being deleted in some cases if no regeneration was
necessary

* 1.0-alpha-21 (January 16, 2009)
** Upgraded to ANTLR 3.1.1
** schema is now at version 1.3.0
** introduced <fplural> allowing -s -es etc.
** XML validation can now be turned on with -validate on the command line
Still have not provided a "preferences" option for the GUI, however
** Added -outputFormat for the command line (can either be "cha" or "xml")
"cha" is still not ready to be used

* 1.0-alpha-20 (December 8, 2008)
** New duration annotation for groups
Using the same format as for pauses, e.g.,

*CHI:   <a b> [# 3.] .
*CHI:   <a b> [# 2:3.] .
*CHI:   <a b> [# 2:3.4] .
*CHI:   <a b> [# 3.4] .

* 1.0-alpha-19 (December 5, 2008)
** Precodes removed
Precodes have been removed from the schema; any use of precodes
should be changed to freecodes instead
** CA
TCU continuation has been fixed to become a CA terminator rather
than a CA element

* 1.0-alpha-18 (November 24, 2008)
** Platform-independent jar
chat2xml.jar is now provided, which works for Windows, Unix, etc.

Double-clicking on error messages works for Windows to pop up CLAN.
** %co1 through %co9
%co1 through %co9 have been removed
** XML
*** The CHAT attribute Mode has been removed
*** The Media attribute has been added
This corresponds to CHAT @Media header.

* Earlier versions (omitted)