CHANGES.TXT                                                     17.07.1999
--------------------------------------------------------------------------
                             ArabTeX Version 3
                             =================
--------------------------------------------------------------------------
        NOTE: Version management:

        - We distinguish major versions "m.00", minor versions "m.nn-"
          with nn > 0, and patch versions "m.nnx" where "x" is a letter.

        - to update from an older major version: reinstall completely.

        - to update from an older minor version: replace the contents of
          "TEXINPUT" and, if so indicated, "REPORT". Replace the fonts
	  only if so indicated explicitly.

        - to update to a newer patch version: replace the ASCII file
          "apatch.sty" and, if you use Hebrew, "hepatch.sty". These files 
          are distributed as unpacked text files, and will be applied
	  automatially whenever starting ArabTeX.

        - Patch versions within the same minor version are cumulative, so
	  a newer patch will encompass all earlier versions. 

	- To back out from a patch version, replace "apatch.sty" by an
	  earlier version (within the same minor version only!) if still
	  locally available, or else edit out the current additions from 
	  "apatch.sty" and, if applicable, "hepatch.sty" (save a copy 
	  first!). Also please inform the author why you backed out.

	- There is no provision to return gracefully to earlier minor or
	  even major versions. We can make them available on request, but
	  you should have good reasons, and there will be no support.
        
==========================================================================
17.07.1999: Version 3.09

	Status:
        
        - This is a new minor version. All existing patches have been 
          merged into the respective macro files. To install it, load
	  and replace all macros in the subdirectory "TEXINPUT"; there
	  are some new files, and some have been extensively rewritten!
	  The functionality is supposed to be unchanged.

	- We recommend installing of the fonts "xnash14" and "xnash14bf" 
	  to obtain better quality of output. These fonts are now the 
	  default; to revert to the old font versions use the command 
	  "\oldarabfont".

	New features:

	- Many ArabTeX commands that used to be fragile in connection 
	  with LaTeX are now robust.

	- Within the LaTeX document classes "arabart", "arabrep", and 
	  "arabbook" nearly all LaTeX environments are allowed also 
	  within Arabic mode.

	Recommendations:

	- When typesetting Roman text with Arabic or Hebrew insertions, a
	  better result can usually be obtained by typesetting on a grid;
	  this can be accomplished by resetting the parameters \lineskip 
	  to 0pt, \lineskiplimit to -30pt, and redefining \baselinestretch
	  to 1.2 (or \baselineskip to 1.2\baselineskip, with Plain TeX).
	  You may want to experiment with the factor quoted as "1.2".

        Bugs:

	- Presently none are known.

	Installation:

	- Download the contents of "TEXINPUT", and in case of any font
	  problems also "MFINPUT".

	- In case of problems with old DVI files regenerate them; or else
	  download "arabsymb.mf" and regenerate the font "nash14".

==========================================================================
24.06.99: Version 3.08i

	New Features:

	- there is an "abstract" environment in the Arabic classes.
	- mathematical displays bracketed by $$ $$ may be nested within an
	  Arabic environment. Bracketing by $ $ was always there.
	- The LaTeX commands \( \), \[ \], and also the LaTeX environments 
	  "math", "displaymath", "equation", "eqnarray", and "eqnarray*" 
	  are supported within Arabic environments, also with the options 
	  "leqno" and "fleqn", if LaTeX is used.

	Upcoming Features:

	- a new edition of the User manual is in preparation.

	Installation:

	- download "apatch.sty", if you are interested.

--------------------------------------------------------------------------
10.06.99: Version 3.08h

	Fixes:

	- there were a few glitches in Pashto and Kashmiri encoding
	- there were a few glitches in Pashto transliteration
	- the command \settrans #1 is allowed within Arabic mode
	- the command \tracingmacros #1 is allowed within Arabic mode

	New Features:

	- the environments "RLtext", "arabtext", and "hebtext" are now 
	  equivalent 
	- the braces \< and > are admissible within Arabic mode but have
	  no effect
	- the command \addcontentsline also works in the Arabic classes
	  "arabrep", "arabart", "arabbook"
	- the title of the table of contents is context dependent in the
	  Arabic document classes

	Upcoming Features:

	- a new edition of the User manual is in preparation
	- there is yet no "abstract" environment in the Arabic classes

	Installation:

	- download "apatch.sty" 

--------------------------------------------------------------------------
16.05.99: Version 3.08g2

	Fixes:

	- In the transcription mode, capitalization (see version 3.08) can 
	  now also be used within a word, even several times.
	- In the transcription, the positioning of the marks for Hamza and 
	  Ayn was slightly improved.
	- in Urdu mode, Ta' marbuta works again.	
	- in BHS mode, Shin works again.
	- in the Arabic document classes, some spurious font switches were 
	  fixed.

	New Features:

	- There is a new language mode: Kashmiri. See "kashmiri.tex" for a
	  short description.

	Installation:

	- download "apatch.sty" (and "kashmiri.sty", "kashmiri.tex" and/or
	  "bhs.sty" if required)

--------------------------------------------------------------------------
10.05.99: Version 3.08f

	Fixes:

	- In the transcription mode, capitalization (see version 3.08) can 
	  now also be used in the English (Encyclopedia of Islam) style.

	Installation:

	- download "apatch.sty".

--------------------------------------------------------------------------
22.04.99: Version 3.08e

	Fixes:

	- We found some grave errors in the Urdu mode; it seems to work
	  again now. Observe that there are several forms of "heh":
	  - the two-eyed form for aspiration: code <h>
	  - the wavy form for other cases: code <,h>
	  - the final closed form: code <H>
	  For the Name of God, use <H> in Urdu mode, or switch to \setarab

	Installation:

	- download "apatch.sty".

--------------------------------------------------------------------------
10.04.99: Version 3.08d

	Fixes:

	- For use with EDMAC, the file "aedpatch.sty" has been updated to
	  fix some spurious spaces.
	- For use with Hebrew, the files "hclassic.mf" and "hcaption.mf"
	  have been slightly adjusted, and "hcbase_pc.mf" was renamed to
	  "hcbase.mf" to fix an incompatibility in some operating systems
	  preventing automatic font generation; the fonts themselves are 
	  unchanged.

	Installation:

	- download the indicated files, if you need them; otherwise you
	  may safely ignore this version.

	Note: the version number of ArabTeX (in the LOG) is still 3.08c !

--------------------------------------------------------------------------
25.03.99: Version 3.08c

	Fixes:

	- Most of the LaTeX renumbering commands introduced in the last
	  version (3.08b) do NOT work and have been blocked off. LaTeX 
	  does not allow font changes within numbering commands, since
	  these must be fully expandable within TeX's mouth.

	- "arabrep.cls" has been updated again.

	Old Features:

	- There is only one new numbering command, and it only can be
	  used in the context of the Arabic document classes, also as a 
	  \pagenumbering style, and for renaming e.g. \thesection:

		  \abj{ctr}     "abjad" notation for counters.

	- \pagenumbering{abj} may also be written \pagenumbering{abjad}.

	- in the Arabic document classes, put \tableofcontents within 
	  an Arabic environment.

	Installation:

	- download "apatch.sty", and also "arabrep.cls" when required. 
	  The latter file is not contained within the packed versions!

--------------------------------------------------------------------------
21.03.99: Version 3.08b1

	New Features:

	- For LaTeX there are additional numbering commands for counters
	  (cf. Section C.8.4 "Numbering" in the LaTeX manual:)

	  \arab{ctr}    "Arabic" Arabic numerals
	  \Arab{ctr}    "Arabic" Arabic numerals, bold
	  \abj{ctr}     "abjad" notation
	  \Abj{ctr}     "abjad" notation, bold
	  \Arabic{ctr}  "European" Arabic numerals, bold (extension!)

	- There are also the corresponding \pagenumbering styles, e.g.:
	  \pagenumbering{Abj} for bold "abjad" page numbers.

	Fixes:

	- Some of these commands were present in 3.07g but did not work.

	- isiri.sty had become incompatible and has been updated.

	- arabrep.cls had become incompatible and has been updated.

	Hints:

	- if your MFINPUT subdirectory of ArabTeX should contain a file 
	  "asymbols.mf", delete it. It is obsolete and may be in conflict 
	  with AMSTeX and/or AMSLaTeX.

	Installation:

	- download "apatch.sty", and "isiri.sty" or "arabrep.cls" when
 	  required. The latter two files are not contained within the 
	  packed versions!

--------------------------------------------------------------------------
13.02.99: Version 3.08a

	Fixes:

	- The transliteration of "Quran alif" was in error.

	Installation:

	- download "apatch.sty" if required.

--------------------------------------------------------------------------
01.12.98: Version 3.08

	Status:
        
        - This is a new minor version. All existing patches have been 
          merged into the respective macro files. To install it, load
	  and replace all macros in the subdirectory "TEXINPUT"; there
	  are some new files, and some have been extensively rewritten!
	  The functionality is supposed to be unchanged.

	- We recommend installing of the fonts "xnash14" and "xnash14bf" 
	  (and activating them) to obtain better quality of output.

	- ArabTeX is believed to be fully compatible with Babel; however,
	  it does not use its features. Please report incompatibilities.

	- BHS mode is now also compatible with "german.sty".

	New features:

	- There are two new experimental fonts "xnash14" and "xnash14bf" 
	  which overcome some limitations inherent in the previously
	  available versions "nash14" and "nash14bf". These fonts can be
	  activated (if installed) by the new command "\newarabfont" and 
	  can be deactivated again by "\oldarabfont"; these commands 
	  change the meaning of the commands "\setnash" and "\setnashbf" 
	  to switch to the new versions, if required, and back. 

	- In transcription mode, single characters may be capitalized by
	  prefixing the command \cap in the input; the Arabic writing is
	  not affected.
	  - with ' (hamza) and ` (`ayn) the next visible letter will be
	    capitalized.
	  - this feature also works within words, after prefixes and after
	    the article, even if written in the assimilated form.

	- In transcription mode, the representations of "hamza" and "ayin"
          have been changed slightly; this might influence line breaks.

	Recommendations:

	- When typesetting Roman text with Arabic or Hebrew insertions, a
	  better result can usually be obtained by typesetting on a grid;
	  this can be accomplished by resetting the parameters \lineskip 
	  to 0pt, \lineskiplimit to -30pt, and redefining \baselinestretch
	  to 1.2 (or \baselineskip to 1.2\baselineskip, with Plain TeX).
	  You may want to experiment with the factor quoted as "1.2".

        Bugs:

	- Presently none are known.

	Installation:

	- Download the contents of "TEXINPUT".

	- In case of problems with old DVI files regenerate them; or else
	  download "arabsymb.mf" and regenerate the font "nash14".

==========================================================================
21.10.98: Version 3.07k

	Fixes:

	- A very obscure internal error has been corrected.

	- In Hebrew mode, the positioning of dagesh and also of some vowel 
	  points has been improved.
	- In Hebrew mode, "meteg" (and silluq) now also works with cholem.
	- In Hebrew mode, there now is a small space before sof pasuq.

	Installation:

	- download "apatch.sty" and "hepatch.sty".

--------------------------------------------------------------------------
18.10.98: Version 3.07j

	Fixes:

	- In Hebrew mode, patach furtivum after cholem works now.
	- In Hebrew verbatim encodings (HED and ISO 8859-8), neither two
	  identical letters nor a medial letter and its final form will be
	  any more combined (erroneously) to a letter with dagesh.

	Open Problems:

	- On some servers, the font "nash14" seems to be broken. This does 
	  NOT concern current ArabTeX runs, but displaying older DVI files
	  might be affected in that some characters, e.g., <^gim>, may be
	  missing. 
	- This will be fixed in the next minor version 3.08; for the time
	  being there are two workarounds:
	  - Reprocess the ArabTeX input file; the display will be correct.
	  - If this is not possible or not wanted, download the ASCII file
	    "arabsymb.mf" and regenerate the font as required.
	- The font "nash14bf" might have the same problem; the workarounds
	  apply also.
	- Neither new ArabTeX runs nor the font "xnsh14" are affected.

	Hints:

	- With Hebrew and LaTeX, the LaTeX2e version should be 1996/12/01
	  or newer; older versions have been reported to produce obscure
	  errors. We can no more check this.

	Installation:

	- download "hepatch.sty", and "arabsymb.mf" if required.

--------------------------------------------------------------------------
30.09.98: Version 3.07i

	Fixes:

	- ArabTeX and CJK are now fully compatible. The loading sequence 
	  of ArabTeX and CJK is arbitrary.
	- defective notation of fatha, kasra, damma works again.

	Installation:

	- download "apatch.sty".

--------------------------------------------------------------------------
17.09.98: Version 3.07h

	Fixes:

	- The position of some more diacritic dots has been corrected.
	- several ligatures have been improved.

	New Features:

	- ArabTeX is now compatible with the package "fnpara.sty" IF this
	  package is loaded BEFORE "arabtex.sty" so that ArabTeX can find
	  out about its presence.
	- ArabTeX is now compatible with the package "CJK.sty" IF this
	  package is loaded BEFORE "arabtex.sty" so that ArabTeX can find
	  out about its presence.

	Special Features:

	- for users of "tustep.sty" (if any :-):
	  ArabTeX is compatible again.

	Known bugs:

	- There is a problem concerning double consonants in Hebrew, only
	  in the HED encoding.
		
	Installation:

	- download "apatch.sty".

--------------------------------------------------------------------------
28.08.98: Version 3.07g

	Fixes:

	- The position of some diacritic dots has been corrected.

	New Features:

	- For LaTeX there are additional styles for page numbers:
	  \pagenumbering{arab}  will produce *real* Arabic page numbers
	  \pagenumbering{Arab}  will produce bold Arabic page numbers
	  \pagenumbering{abjad} will produce "abjad" page numbers
		
	Installation:

	- download "apatch.sty".

--------------------------------------------------------------------------
06.08.98: Version 3.07f

	Fixes:

	- there was a small problem transliterating emphasized words.

	- separating words by || did not work.

	- the character for final "ny" in Malay was wrong.
	- the character for final "h"  in Malay was wrong.

	New Features:

	- Arabic and Persian words may be separated by \, (small space)

	Installation:

	- download "apatch.sty".

--------------------------------------------------------------------------
31.07.98: Version 3.07e

	Fixes:

	- there were spurious characters in Hebrew output.

	Installation:

	- download "apatch.sty" and "hepatch.sty".

--------------------------------------------------------------------------
28.07.98: Version 3.07d

	New Features:

	- there is a new directory "psfonts" that contains the Postscript 
	  Type 1 versions of the fonts "xnsh14" and "xnsh14bf". The files
	  are called "xnsh14.pfb", "xnsh14bf.pfb", and "arabtex.map". The
	  fonts use the original TFM files. They have been tested with
	  dvips 5.76, ghostscript 5.10, a LaserJet 4+, and Acrobat Reader.

	  These font files have been built and donated by Taco Hoekwater
	  (taco.hoekwater@wkap.nl). Three cheers on him!

	Old Features:

	- the Arabic commands \lq (opening quotes), \rq (closing quotes), 
	  and \,(small non-breaking space) have been present for some time; 
	  we forgot to document this fact.

	Installation:

	- download and install the contents of the directory "psfonts"
	  wherever appropriate on your system; this is system dependent.

	- "apatch.sty" was changed, but need not be downloaded except on
	  special advice.

--------------------------------------------------------------------------
27.07.98: Version 3.07c

	Fixes:

	- in Farsi mode ,a and ,e work again.

	Installation:

	- download "apatch.sty" and "hepatch.sty".

--------------------------------------------------------------------------
23.07.98: Version 3.07b

	Fixes:

	- The letter SIN in Hebrew mode was broken.
	- The Sindhi abbreviations look better now.

	Installation:

	- download "apatch.sty" and "hepatch.sty".

--------------------------------------------------------------------------
20.07.98: Version 3.07a

	Fixes:

	- After installing the version 3.07 on our local server we noticed
	  an incompatibility with LaTeX2e; also by some new strategies the
	  patching mechanism had been broken. To fix these effects we had
	  to modify several source modules; we also found, and repaired,
	  several minor bugs. This is in fact nearly a new minor version;
	  on installation see below.

	- To our surprise we found, and repaired, a Y2000 problem.

	Changes:

	- Switch to the new font "xnsh14" by the command "\newarabfont"; 
	  use "\oldarabfont" for switching back. The previous command 
	  "\newfont" collides with a command in LaTeX2e.

	Extensions:

	- The font "xnsh14" contains all character fragments necessary to
	  build the characters in the extended Arabic segment of Unicode.
	  There is as yet no reading module for Unicode or UTF-8 encoding;
	  the present version of "utf8.sty" is still inconsistent.

	- There are new commands "\MIN" and "\IN" for two frequently used
	  abbreviations in Sindhi.

	- The file "atrans.sty" is loaded automatically.

	- The file "etrans.sty" has been deleted. Use "\settrans{english}"

	- The file "nashbf.sty" has been deleted. Use "\setnashbf"

	Installation:

	- If you already have loaded version 3.07, unpack and replace the 
	  macro files in "tex307a.zip"; otherwise do a complete install as
	  indicated below for version 3.07 first.

--------------------------------------------------------------------------
16.07.98: Version 3.07

	Status:
        
        - This is a new minor version. All existing patches have been 
          merged into the respective macro files. To install it, load
	  and replace all macros in the subdirectory "TEXINPUT"; there
	  are several new files!

	- We have rewritten substantial parts of the package in order to
	  increase stability and also to ease maintenance and extensions;
	  the modularisation is now different, and some of the internal 
	  interfaces have been modified. This need not concern the user;
	  the functionality has not been changed, and apparently nothing 
	  was broken during the transition to the new structure.
	- The main reason for the changes was, apart from improving the
	  overall structure, to prepare for the handling of the full 
	  Arabic segment of Unicode. This goal has not yet been attained,
	  but it required substantial changes in the internal structures.
	- The positioning of diacritical marks has on the average been 
	  much improved. There are a few exceptions; affected users that 
	  need the maximum quality available could try their luck with 
	  the new fonts "xnash14" and "xnash14bf" (see below). Otherwise 
	  the user interface has not changed.

	Additions:

	- There are two new experimental fonts "xnash14" and "xnash14bf" 
	  which overcome some limitations inherent in the previously
	  available versions "nash14" and "nash14bf". These fonts can be
	  activated (if installed) by the new command "\newfont" and can
	  be deactivated again by "\oldfont"; these commands change the
	  meaning of the commands "\setnash" and "\setnashbf" to switch 
	  to the new versions, if required, and back. Normal users not 
	  requiring the new versions can safely ignore this feature; they 
	  need not now install these fonts.

        Bugs:

	- Presently none are known.

==========================================================================
24.03.98: Version 3.06h

	Fixes:
        
        - some ligatures had been broken by the Sindhi mode; these have 
          been fixed.
        - kashida before alif maqsura works again.

        Problems:

        - verses produced via "verses.sty" will ruin the word spacing if
          they occur within an Arabic paragraph inside a footnote, and
          possibly in other circumstances. There is a workaround: put the 
	  command "\setspace{4.3pt plus 0.001fil}" after the last verse.
        
--------------------------------------------------------------------------
11.03.98: Version 3.06g3

	Fixes:
        
        - bi-directional linebreaking has been improved; there were some
          spurious blank spaces.
        
        New features:

        - ArabTeX is now believed to be compatible with the BABEL system.
          We cannot test this locally; reports are welcome. We acknowledge
          useful hints by Olaf Kummer and Johannes Braams.

        - there is now less spurious information in the LOG file.

        - there are some preparations for processing the UTF-8 encoding
          and the extended Arabic character set provided in UNICODE. These
          features cannot yet be accessed from the user level. 

--------------------------------------------------------------------------
26.01.98: Version 3.06f

	Fixes:
        
        - the special handling of final "mim" has been restricted to the
          Sindhi mode.
        - the positioning of diacritics has been improved.

        - Hebrew and EDMAC users: the "\if@lodia effect" has been fixed.

        Note to users of EDMAC:

        - download the new version 3.06a of 22.01.98 of "aedpatch.sty".
          This fixes the format of footnote paragraphs. (This correction
          cannot technically be handled by a patch within "apatch.sty").

        Note to users of the Sindhi mode:

        - please download the fonts again to get all characters.

--------------------------------------------------------------------------
15.12.97: Version 3.06e

	Fixes:
        
        - "alif maqsura" and isolated "hamza" take explicit vowel marks
          and also "tanwin" if required.
        - final "mim" in the Urdu mode again has the long shape. In the
          Sindhi mode a short form is simulated by a kashida if needed.
        - in Urdu the encoding <,d> produces "dal" with a "toi" accent, 
          and initial <U> becomes "alif+damma" and "waw" as required.
        - some transcription errors in Urdu mode have been fixed.
        - some ligatures in Sindhi mode are improved.

        Problems:

        - medial "wavy hah" in Urdu has a mark below it (controversial).
        
--------------------------------------------------------------------------
27.11.97: Version 3.06d

	Fixes:

        - the input letter <A> denotes alif also in initial position and 
          after a prefix separated by a hyphen. It may take short vowels, 
          and also sukun (only if quoted). This feature may be useful when 
          copying corrupted texts.
        - the "lam" of the article carries "sukun" also if the following
          word starts with "`ayin", like with the other "moon letters".
        - the encodings <I> and <iy> are equivalent; also <Iy> and <iyy>,
          <U> and <uw>, <Uw> and <uww>.
        - <"> and <""> at the beginning of a word generate alif with wasla
          even after a prefix.
        - a prefix before hyphen no more needs a short vowel to work.
        - the invisible character <|> is legal in the initial or the final 
          position (but makes no sense there). It never takes a vowel mark.

        New features:

        - the inflection endings <-a>, <-i>, <-u>, <-aN>, <-iN>, <-uN> may 
          be separated by a hyphen, also if written <-aNA>. Otherwise the
          character after a hyphen will be considered to start a new word, 
          but may be joined to the prefix if possible (e.g., in <al-i-ismu>
          where <-i-> is a connection vowel producing "kasra", and <ismu> 
          starts with "alif" "wasla" combined with "lam" into "lam-alif"
          carrying both "wasla" and "kasra".)
        - if a ligature is separated by <|> or <--> (kashida), any vowel 
          mark present will migrate to the appropriate character glyph. The
          same happens with the implicit vowel mark that is produced by the
          long vowels in the vowelizing modes.
        - a comma <,> between digits will no more break the digit sequence 
          but will be recognized as a decimal separator. A hyphen <-> and a
          solidus </> will also not break a digit sequence. A "number" must
          still begin with a digit to be recognized.

        Problems:

        - the Urdu mode is still broken.
        - for the new language modes (see 3.06) there is yet no (correct)
          transcription.

        Hints:

        - users of the Sindhi mode should regenerate the fonts otherwise
          the "wide kaf" may be missing.

--------------------------------------------------------------------------
23.11.97: Version 3.06c

	Fixes:

        - final "alif maqsura" carries no sukun.
        - in connection with EDMAC, an Arabic lemma bearing a footnote is
          spread out by the same factor as the rest of the line, (e.g. in 
          verses) even if it contains spaces.
        - Arabic footnotes with EDMAC no more produce overfull boxes, and
          will no more let the page overflow.

        Known bugs:

        - prefixes separated by a hyphen must presently be given in the 
          fully vowelized input form, otherwise the writing is wrong.
        - The Urdu mode is broken; also its transcription is wrong.

--------------------------------------------------------------------------
25.09.97: Version 3.06b

	Fixes:

        - in Kurdish mode, kashida and breaking of ligatures work now

--------------------------------------------------------------------------
18.09.97: Version 3.06a

	Fixes:

        - there were sometimes spurious newlines if the transliteration 
          was generated via an Arabic insertion
        - ligatures can again be broken by inserting a bar |
        - the ligatures family lam-gim, ta-gim etc. works again
        - there were spurious spaces in bidirectional text under EDMAC

--------------------------------------------------------------------------
06.08.97: Version 3.06

	Status:
        
        - This is a new minor version. All existing patches have been 
          merged into the respective macro files.
        - This doc file has been trimmed down; history information up to
          version 3.00 has been moved to "changes2.txt".
        - Several file names of the mini-documentation have been changed 
          to lower case and carry the suffix ".txt", for the benefit of
          certain web-browsers. See also the file "arabtex.htm".

	Fixes:

        - some minor bugs have been fixed that had introduced spurious
          spaces or newlines when switching between Arabic and Hebrew.
        - "kashida" (--) between a consonant and a (long or short) vowel 
          will no more split the vowel indicator from the consonant glyph.

	Additions:

        - the new command "\setsindhi" will switch to a language mode for
          Sindhi in an extended Arabic script. For details, process and
          read the file "sindhi.tex".
        - the new command "\setuighur" will switch to a language mode for
          Uighuric in an extended Arabic script. For details, process and
          read the file "uighur.tex".
        - the new command "\setmalay" will switch to a language mode for
          Old Malay (Jawi) in an extended Arabic script. For details, 
          process and read the file "malay.tex".
        - These new language modes are strictly experimental and just a
          starting point. As the author speaks none of these languages,         
          feedback and comments by interested users are badly needed.

==========================================================================
19.06.97: Version 3.05h

	Fixes:

	- some ligatures with "kaf" have been improved.
        - "quoting" of "wasla" works in all vowel modes.
        - "kashida" works also before long vowels.
        - breaking of ligatures works correctly also with long vowels
          and in combination with "quoting".
                
--------------------------------------------------------------------------
12.06.97: Version 3.05g

	Fixes:
                
        - a bug fix for the new Sindhi module has been introduced (this
          mode is in alpha status and not yet publicly available; if you
          are interested in testing it: enquire)
                
--------------------------------------------------------------------------
10.06.97: Version 3.05f

	Fixes:
        
        - a spurious error message about a "missing character" has been
          eliminated.
        - "\settrans{english}" switches again to the transliteration
          conventions of the Library of Congress.
        - ArabTeX will no more overwrite a user macro "\x".

	Additions:

        - the new command "\arabfootnotes" makes LaTeX footnotes run from
          right to left, and their text is considered Arabic by default. 
          There is as yet no command to switch back to the default mode
          (left to right, Roman). With Plain TeX the new command is not
          yet available.

--------------------------------------------------------------------------
28.05.97: Version 3.05e

	Fixes:

        - <U> in the Urdu mode works again.
        - some errors in the transliteration have been corrected.
        - some ligatures have been corrected.

--------------------------------------------------------------------------
22.05.97: Version 3.05d

	Fixes:

        - \tabular structures within the Arabic document classes are now
          correctly aligned.
        - a few ligatures have been improved.
        - some internal routines have been updated preparing for some
          planned extensions. No visible effect.
        - a new test for version clashes has been introduced.

--------------------------------------------------------------------------
14.05.97: Version 3.05c

	Fixes:  The patching routine seems to work again.

--------------------------------------------------------------------------
14.05.97: Version 3.05b

        Problems:

        - LaTeX is too clever and interferes with the patching routine.
          Therefore the existing patches for Hebrew have been moved to
          the file "hepatch.sty". Arabic mode is presently not affected.

--------------------------------------------------------------------------
13.05.97: Version 3.05a

	Fixes:

        - in Hebrew mode when using HED encoding, Thet and Maqqef did
          not work correctly.
        - in EDMAC mode with Plain TeX the list of modules was missing.
        - some internal interfaces have been updated. This has no effect
          on the external operation but helps stability.

--------------------------------------------------------------------------
07.05.97: Version 3.05

	Status:

	- The internal structure has changed extensively; this need not
	  trouble the user.
        - Warning: NEVER MIX old and new files; they may be incompatible!

	Fixes:

	- A number of minor and some major bugs have been fixed.

	Modifications:

	- With EDMAC, the first argument of \text need not fit on the
          current line. All arguments are assumed to be Roman text; they
          may contain (or consist of) \<Arabic insertions>.

	Additions:

	- The ISIRI 3342 encoding is supported; load the reading module 
          by \input isiri.sty or \usepackage {isiri}. Activate it by the
	  command \setcode {isiri}.
        - Hebrew support has been expanded; read the file "hebrew.305".
          ArabTeX will try to use fonts locally available, irrespective
          of their encoding.
        - the file "verses.sty" supports typesetting poetry in two blocks. 
          For a description, see the file itself.
	- See also the comments on Version 3.04 below.

	Hints:

	- If you want to process the manual "arabdoc.tex" under LaTeX 2e
          and get error messages, delete the *.aux, *.toc, and *.lot file
          and try again. These files depend on the LaTeX version used.
        - see also version 3.04.

	Complaints:

	- There is STILL no Nasta`liq font.
        - the user manual "arabdoc.tex" is outdated.

	Known bugs:

	- None as yet.

	How to move from earlier minor versions to version 3.05:

        - replace the contents of "TEXINPUT" and "REPORT".
	- read and/or print "README.305"
	- if interested, process the user manual "arabdoc.tex"; see the
	  hints above.

==========================================================================
05.10.96: Version 3.04e

	Fixes:

	- removed many small bugs in Persian (thanks, Ivan Derszanski!)
	- initial <alif> coded <A> in verbatim mode may carry a vowel
	- <yah> without dots coded <Y> acts like a consonant except at the
	  end of a word
	- <aNy> denotes <fathatan> on <alif maqsura> erroneously written
	  with dots

	Modifications:

	- removed quoted hamza ("') from verbatim mode (was ambiguous!)

--------------------------------------------------------------------------
18.05.96: Version 3.04d
10.05.96: Version 3.04c (not distributed)

	Fixes:

	- in "arabart.cls" (twocolumn) the right column is output first 
	  to get correct numbering of sectional units
	- corrected the vertical position of "marginpar"
	- a space is assumed after a new command defined via \allowarab

--------------------------------------------------------------------------
08.04.96: Version 3.04b

	Fixes:

	- \footnote works again within an Arabic environment

	Known bugs:

	- \footnote is incompatible with "arabart.cls"

--------------------------------------------------------------------------
21.03.96: Version 3.04a

	Fixes:

	- \baselineskip works correctly with EDMAC
	- \emphasize puts the overbar at uniform height in \novocalize

	Necessary action:

	If required, download the ASCII file "apatch.sty" again. This
	is a patch for version 3.04!

	See also version 3.04 below.

--------------------------------------------------------------------------
19.03.96: Version 3.04

	Status:

	- ArabTeX now cooperates well with LaTeX 2e (June 1995 and later)
	  and also still with Plain TeX and LaTeX 2.09. The configuration 
	  is determined automatically. With LaTeX 2e ArabTeX and all its
	  suboptions are packages.
	- The internal structure has changed extensively; this need not
	  trouble the user.

	Fixes:

	- \baselineskup, \parskip, \baselinestretch now work correctly
	  in the Arabic modes.
	- The TeX special characters \#, \$, \%, \& can now be used (they
	  just had been forgotten).
	- A number of minor bugs have been fixed.

	Modifications:

	- Digits and punctuation have been raised to the writing line.
	  We feel this looks better.
	- The handling of numbers (= digit sequences) has been improved.
	  Within a number, the decimal separator is the comma.
	- Bidirectional line-breaking has again been much improved. The 
	  command \goodpar still works but is rarely required.
	- The option "abidir" is no more necessary.
	- The option "nashbf" is no more necessary.
	- The option "abjad"  is no more necessary.

	Additions:

	- User defined macros can be made known to ArabTeX by the
	  declaration \allowarab {\macro_name} (previous to their use
	  within an Arabic environment). The replacement text after
	  substituting the parameters must be legal ArabTeX input and
	  may contain further macro calls. The macro must be previously 
	  defined via \newcommand or \def.

	- The Arabic Windows encoding is supported; load the reading
	  module by \input arabwin.sty, \usepackage {arabwin} or by
	  adding the LaTeX 2.09 option "arabwin". Activate it by the
	  command \setcode {arabwin}.

	- There are some additional transliteration modes for Persian
	  and Urdu; in case of interest: inquire directly.

	- The package "raw.sty" defines the command \setraw to switch 
	  off all special processing for numbers, special characters
	  and parentheses. This may be helpful for OCR work. Switch
	  back by \unsetraw.

	- The (experimental) LaTeX 2e class file "arabart.cls" provides
	  an Arabic article format. Within Arabic environments a number
	  of LaTeX commands are enabled: sectioning, list environments,
	  page headers and footers, table of contents, footnotes, quote,
	  quotation, verse, center, twocolumn. All text arguments are
	  considered to be Arabic text in the chosen encoding. Some of
	  the LaTeX environments are missing, e.g., tabular and tabbing,
	  minipages, pictures. They can usually be simulated by leaving
	  the Arabic environment and reverting to the standard LaTeX
	  commands.

	  We are working on extensions to this format; suggestions and
	  bug reports are welcome. We do not intend to supply a version 
	  for LaTeX 2.09.

	- See also the comments on Version 3.03 below.
	- The notes in README.303 still apply.

	Hints:

	- Uniform baselines may be obtained by setting \lineskiplimit
	  to -\baselineskip and adjusting \baselinestretch to 1.4 (with
	  LaTeX) or \baselineskip to 1.4\baselineskip with Plain TeX.
	  Some experimenting is advised.
	- If you want to process the manual "arabdoc.tex" under LaTeX 2e
	  first move the file "arabimax.tex" to the REPORT subdirectory.
	  You will get some error messages on the first run. Forget them 
	  and process the file again; LaTeX 2e is to blame. Warning: the 
	  report is not quite up to date.
	- Put any local macro definitions into the LaTeX preamble or else 
	  precede them by \setnone; thus the < sign can be used in the 
	  normal meaning as a comparison operator. Activate short Arabic
	  insertions again by \setarab or any other of the commands for
	  language choice. 
	- PiCTeX users may use Arabic labels within diagrams if and only 
	  if these are included in \< and > (NOT < and >).
	- If you use MlTeX and the output looks broken, note the version
	  number of MlTeX and inform us.

	Complaints:

	- There is STILL no Nasta`liq font.

	Known bugs:

	- None as yet.

	How to move from earlier minor versions to version 3.04:

	- read and/or print "README.303"

==========================================================================
04.07.95: Version 3.03e

	Fixes:

	- Final yah before hamza works again

--------------------------------------------------------------------------
24.05.95: Version 3.03d

	Fixes:

	- After switching to other encodings switching back to the
	  standard encoding works again.
	- An empty \RL {} insertion will no more blow the system.

	Problems: 

	- see version 3.03c

--------------------------------------------------------------------------
02.05.95: Version 3.03c

	Fixes:

	- The METAFONT routines for the fonts "nash14" and "nash14bf" 
	  used a file "asymbols.mf" that conflicted with the AMS-Fonts.
	  This file has been renamed to "arabsymb.mf". To use the new
	  version, download "arabsymb.mf" and "nashbase.mf" from the
	  subdirectory ARABTEX/MFINPUT and delete "asymbols.mf". This
	  only concerns users who need to regenerate the fonts locally.

	Problems: 

	- The font "dclassic" used by the Hebrew mode of ArabTeX has on
	  the archive "noa.huji.ac.il" been replaced by a new version 
	  that is incompatible to the previous version while keeping the
	  same name. We are working on a solution to resolve the ensuing
	  compatibility problems. Users of the previous version are not 
	  affected.

	Acknowledgment:

	- The routines for building up punctuated Hebrew characters have
	  profited from macros written by Joel Hoffmann and contained in
	  his packages "dots.sty" and "hvowel1.sty". We apologize for not
	  pointing out this fact clearly.

	Necessary action:

	- Normally none, except if indicated above.

--------------------------------------------------------------------------
24.04.95: Version 3.03b

	Fixes:
	
	- The handling of single and double quotes, and of other special
	  characters, in the ASMO 449 and ISO 8859-6 encodings has been 
	  improved.
	- After a punctuation character inside a word (usually an error)
	  no space will be produced.
	- <tanwIn> on <'alif maqsUraT> works again.
	- In \setverb mode, a <harakaT> on <'alif> now works.
	- In \setverb mode, <A> may be used to produce <'alif>.
	- The ligature <lm|a> is produced instead of <l|ma>.
	- <ba'>, <ya'> etc. before <lAm> now have the short form.

	Known bugs in ArabTeX:

	- \begin{arabtext} inside a minipage will produce a spurious
	  empty line; the reason is unknown.

	Hints:

	- Numbers after quotes will not be recognized as such and thus 
	  will be reversed; as a workaround, leave a space around the 
	  digits, e.g., <`` 1234 ''> or use a \nospace command, e.g., 
	  <`` \nospace 1234 \nospace ''>.
	  Mind the spaces before the \nospace commands in the example!
	- With LaTeX2e, load ArabTeX and its suboptions by the command
	  \usepackage.

--------------------------------------------------------------------------
28.03.95: Version 3.03a

	Fixes:

	- A workaround for an error in some versions of ML-TeX has
	  been tentatively introduced. Non-ML-TeX users may safely
	  ignore this.

--------------------------------------------------------------------------
10.03.95: Version 3.03

	Fixes:

	- A number of minor bugs have been fixed, nothing dramatic.
	- The spacing of punctuation has been improved.
	
	Modifications:

	- Hebrew vowel points are provided. Hebrew mode now uses the font 
	  "dclassic". It can be downloaded e.g. from "noa.huji.ac.il". 

	Additions:

	- The Hebrew encoding used in the machine readable version
	  of BHS is supported. See README.303!
	- \obeylines and \obeyspaces also work for Arabic text. They
  	  however must be used in a group that contains the Arabic
	  text AND this group must be followed by at least one space.
	- An <Arabic insertion> inside a \verbatim environment will be
	  typeset in Arabic. Big TeX may be necessary.
	  An Arabic environment will be typeset verbatim.
	- \vskip, \hskip, \vspace, \hspace, \newpage, \vfill are legal
	  in Arabic mode.
	- A list of the loaded modules is put into the log file.

	Known bugs:

	- None as yet.

	How to move from earlier minor versions to version 3.03:

	- completely replace the contents of the subdirectory TEXINPUT
	- move the updated files "arabdoc?.tex" from the outermost level 
	  to the subdirectory "REPORT"
	- read and/or print "README.303"

==========================================================================
12.08.94: Version 3.02b

	Fixes:
	
	- \fullvocalize mode uses somewhat less TeX main memory than
	  before but might still overflow; using \newpage explicitly
	  could help
	- some ligatures have been improved
	- spacing has been improved in short insertions when the
	  transliteration is also generated
	- in Hebrew mode, HED encoding, double letters work again
	- explicit spreading works again correctly
	- <hamza> in rare cases had the wrong carrier
	- if an Arabic word starts with a vowel, "quoting" toggles 
	  <wasla> if no vowel is given in the input
	- bidirectional line-breaking has been much improved. The
	  command \goodpar is rarely necessary

	Hints:

	- etrans.sty is no more necessary; load atrans.sty 
	- \settrans{english} switches to the romanization conventions
	  of the Library of Congress

--------------------------------------------------------------------------
19.07.94: Version 3.02a

	Fixes:

	- in Arabic, alif maqsura after tanwin works again
	- in Hebrew, before an apostrophe the medial form is used
	- in Hebrew, hyphen/maqqef works correctly

	Hints:

	- with Plain TeX (and/or EDMAC) only: load the ASCII file
	  "apatch.sty" explicitly as the last file!

--------------------------------------------------------------------------
14.07.94: Version 3.02

	Fixes:

	- some minor bugs have been corrected
	- the User Manual has been slightly updated, but is otherwise
	  somewhat out of date.

	Modifications:

	- for storage reasons the transcription module is not loaded
	  automatically; add the LaTeX option "atrans" if required,
	  or say "\input atrans.sty".

	Additions:

	- EDMAC 3.00 and EDMAC 3.15 are both supported.
	- the LaTeX option "abidir" provides the new command \RL {text}
	  for right-to-left insertions; it is used like <text> which is
	  still provided, and which should be used wherever possible for 
	  reasons of efficiency.
	- the LaTeX option "hebtex" loads the modules for processing text
	  in Hebrew. Switch to Hebrew by \sethebrew, and back to Arabic
	  by \setarab. Hebrew text can be coded in the standard ArabTeX
  	  input notation, or in the encoding used by the editor HED; the
	  latter coding is activated by \setcode {hed}. The reading module
	  is supposed to also process texts in "newcode" (not tested!)
	- For using the Hebrew mode, the font "deadsea" must be available;
	  it can be downloaded e.g. from "noa.huji.ac.il". Vowels are
	  presently not provided.
	- For details see the ASCII file "readme.302".

	Hints:

	- only load the modules really required, or you might have to use 
	  a Big TeX!
	
	Known bugs:

	- None.

	How to move from earlier minor versions to version 3.02:

	- completely replace the contents of the subdirectory TEXINPUT
	- move the updated files "arabdoc?.tex" from the outermost level 
	  to the subdirectory "REPORT"
	- read and/or print "README.302"

==========================================================================
05.05.94: Version 3.01a

	Fixes:

	- \parskip now has effect also inside Arabic environments.

	Additions:

	- The EDMAC commands \pstart, \pend, \autopar are allowed
	  inside an Arabic environment, IF properly nested.

--------------------------------------------------------------------------
28.04.94: Version 3.01

	There has been extensive internal rewriting but the external 
	effects are minor.

	- \setarab is the language default again. It will no more break
	  the user's macro definitions IF these are put into the LaTeX
	  preamble, or before "\input arabtex.sty" when using Plain TeX.

	Fixes:

	- Spaces are normally ignored after allowed control sequences 
	  inside Arabic text, (if not produced by the command itself.)
	- The spacing between a final silent <hah> in Persian and the 
	  next word has been improved.

	- An error in "twoblocks.sty" has been corrected that produced 
	  a spurious vertical space.
	- An error in "asmo449.sty" has been corrected that produced 
	  nonsense from the combination "Gdr" (article with an explicit
	  <sukUn>) 
	- In "abjad.sty", the <gIm> for the numeral 3 has no dot.

	Additions:

	- The commands \smallskip, \medskip, \bigskip, \newpage,
	  \clearpage, \pagebreak, \hskip {glue}, \vskip {glue}, \today
	  now work inside Arabic text.
	- there is a new language mode: \setturk for Ottoman.
	  The input notation has been extended to cover both the standard
	  and the modern Turkish transliteration.
	  The documentation remains to be written :-(
	- there is a new language mode: \setkurdish for Kurdish.
	  The documentation remains to be written :-(
	- "" produces an explicit <sukUn> except in \fullvocalize mode
	  where it deletes a <sukUn> otherwise present. 
	- The new commands \setarabfont {\nash} or \setarabfont {\nashbf}
	  supersede \setnash and \setnashbf (which are still supported).

	Known bugs:

	- None.

	Hints:

	- The LaTeX style option "atrans" is redundant.
	- After changing the input coding, select the language mode and
	  the vowelization mode anew (this is a feature :-).

	How to move from earlier minor versions to version 3.01:

	- completely replace the contents of the directory TEXINPUT
	- observe that there is a new file "acmd.sty"

==========================================================================
07.12.93: Version 3.00b

	Fixes:

	- corrected handling of <tanween fatha> in \setverb mode.
	- improved error handling for illegal commands.
	- ligature <geem>-<meem> also in spreading mode.

	Additions:

	- \setspace{rubber_length} changes the default for spaces in 
	  Arabic text; 3pt plus .5fil is recommended for \spreadbox.

	Known bugs:

	- arguments of boxing commands may not contain insertions.
	- in ASMO449, the combination Gdr (article with an explicit
	  sukoon) will produce nonsense.

	Hints:
	
	- There is no language default. Do not forget \setarab etc.
	  or otherwise you might get strange results.
	- After changing the input coding, select the language and
	  the vowelization mode anew (this is a feature).

	  These are in the manual but easily overlooked.

--------------------------------------------------------------------------
26.11.93: Version 3.00a

	Fixes:

	- the <hamza> carrier is now <ya'> the before diphthong <ay>

	Additions:

	- more Qur'anic writing variants provided
	- \hfil is allowed in Arabic text

--------------------------------------------------------------------------
20.11.93: Major Version 3.00

	Status:

	- large portions of the code have been rearranged and partly
	  rewritten
	- The User Manual has been extensively updated and describes 
          all (supported) ArabTeX features currently implemented
	- Version 3.00 seems to be compatible with Version 2 files
	- NFSS and NFSS2 can be used but are not required
	- all known bugs have been removed
	- Version 3.00 still can be run (slowly) on a standard PC XT

	Additions:

	- texts coded in ASMO 449 or ISO 8859-6 can now be processed
	- several historic writing variants are additionally supported
	- details of Farsi support are now described in the User Manual
	- many new commands are valid inside Arabic text; a complete 
	  list appears in the log file of each ArabTeX run, and also
	  in the new User Manual
	- see also versions 2.00 - 2.08 below

==========================================================================

	How to move from version 2.0x, 0 <= x <= 8, to version 3.00:

	- completely replace the contents of the directory TEXINPUT
	- completely replace the contents of the directory REPORT
	- process AND READ the new User Manual "arabdoc.tex"
	- leave the fonts alone

        How to move from version 1 (or none at all) to version 3.00:

        - Do as above; install the fonts "nash14" and "nash14bf" first
          according to local TeX conventions.
        - If the font "nash10" is still there, delete it.

==========================================================================
        For information on older versions see the file "changes2.txt".
==========================================================================
    Prof. Klaus Lagally
    Institut fuer Informatik
    Universitaet Stuttgart
    Breitwiesenstrasse 20-22
    D-70565 Stuttgart
    GERMANY

    lagally@informatik.uni-stuttgart.de
    FAX: +49 - 711 - 7816 370
--------------------------------------------------------------------------
	Copyright (c) 1990 - 1999, Klaus Lagally
--------------------------------------------------------------------------