OESF Portables Forum
Everything Else => General Support and Discussion => Zaurus General Forums => Archived Forums => Software => Topic started by: rafm on November 21, 2004, 01:48:45 pm
-
A new version of zbedic (0.9.2-3) is released and available for download at sourceforge:
(download (http://prdownloads.sourceforge.net/bedic/zbedic_0.9.2-3_arm.ipk?download))
Changes:
07.11.2004 0.9.2-1
* Added keyborad shorcuts for a list of dictionaries, history,
scrolling description, and selecting words from the wordlist
* Added option for changing size of the wordlist - gives more
space for description
* Added: zbedic switches to the right dictionary for items from
the history
* Added: A word is added to the history when it's viewed for 3
seconds.
* Documentation updated to relfect the changes
* Bug fix: A description sometimes didn't show up after entering
a word
07.11.2004 0.9.2-2
* Added: Layout mode (vertical or horizontal) for 7xx models
20.11.2004 0.9.2-3
* Fix: The input box is not emptied after changing a dictionary
* Added: The list of dictionaries is sorted
* Fix: Can be now installed on CF/SD
I have also written down some ideas for future releases and put them in a TODO file in the CVS. There is no much chance to realize them soon, but they may be worth discussing. There is a copy of that file:
* Navigation: back, forward
This should be integrated with the word history, not to make too
much duplicate work. The problem is that the history of words is a
simple list and browser-like navigation could be more graph-like.
For example: go back a few words in history, enter a new
word. Then the new word should be added at the beginning of the
history or just after a word recently recalled from the history?
* Pattern search
For example: prefix*postfix, i*ary
* Better rules for lexical precedence
Accented characters should be properly handled and have proper
precedence (after non-accented counterparts).
** Special treatment of "compound letters"
I am not sure what is the proper name for things like ss, au, eu, ou
in German, but it would be nice if those were put near corresponding
accented characters, so one could enter words using both methods of
spelling.
* Case-sensitive dictionaries
Some dictionaries, like German should be case-sensitive.
* Unlimited list of words
Instead of returning fixed number of words, the word list should show as
many words as fit in the window and it should be possible to browse it
backward/forward up to beginning and end of the dictionary. Of
course scrollbar is totally unnecessary in such case.
* Search for misspelled words
Search could return near matches for misspelled words.
* Pixmaps for dictionaries
A pixamaps for dictionaries shown in a list of dictionaries and in
history would be a nice GUI add-on.
* Customizable toolbar
It should be possible to enable / disable toolbar buttons in the
Settings. Possible new buttons: navigation back, forward.
* Reorganized settings dialog
There should be several panes, for example: Appearance,
Dictionaries, Toolbar, Advanced
* Additional tags for dictionary files
Meaning, sub-meaning, example, part-of-speech is often not enough to
describe entries. For example it is not clear how to enter
phrase/idiom or word category (e.g. math. biol., etc.)
-
when source archive will be available? OE has 0.9.1 now..
-
There is one particular thing I'd love to have someone figure out how to do.
The chinese-english english-chinese dictionary is great, except I have no clue as to how to pronounce the characters. For that, another dictionary has the answer (qpzidian which is in cedict format) http://killefiz.de/zaurus/showdetail.php?app=482 (http://killefiz.de/zaurus/showdetail.php?app=482)
Unfortunately, that dictionary program is very very slow, vs zbedic, which is much more responsive and faster.
The answer lies here:
http://www.dia.janis.or.jp/~syoshi/convert.htm (http://www.dia.janis.or.jp/~syoshi/convert.htm)
and here
http://home.k07.itscom.net/west/recommend/...ic/convert.html (http://home.k07.itscom.net/west/recommend/zbedic/convert.html)
(you MUST use http://www.systran.net/ (http://www.systran.net/) to translate it from Japanese to English).
I've followed all the steps, and zbedic can see the dictionary I convert, but I no see any information in the dictionary itself when I load it up. I'm at a loss as to what to do next, so I'm passing the information along, in hopes that someone else can figure out the answer here!
-
I've followed all the steps, and zedict can see the dictionary I convert, but I no see any information in the dictionary itself when I load it up. I'm at a loss as to what to do next, so I'm passing the information along, in hopes that someone else can figure out the answer here!
zedict = zbedic?
Please put the converted dictionary file (.dic.dz) somewhere on the web so I can download and check it.
-
when source archive will be available? OE has 0.9.1 now..
I would be nice if somebody could extend Makefile so it would be possible to build both OE and Sharp ROM releases. I cannot do it myself since I don't have OE and don't really want to invest time in learning how to compile for OE.
-
sorry, that was a typo on my part. I've posted 2 files for your review.
http://zaurus.mui.net/cedict-zb.dic (http://zaurus.mui.net/cedict-zb.dic) converted file
http://zaurus.mui.net/cedict-zb.txt (http://zaurus.mui.net/cedict-zb.txt) before using xerox on it
The original dictionaries can be found here:
http://www.mandarintools.com/cedict.html (http://www.mandarintools.com/cedict.html)
cedictgb.zip cedict dictionary in GB
cedictb5.zip cedict dictionary in Big5
cedictu8.zip cedict dictionary in UTF-8
(I've included the references for completeness)
As mentioned earlier, the advantages of translating these dictionaries is that it includes the pronounciation for the characters!
-
A new version of zbedic (0.9.2-3) is released and available for download at sourceforge:
rafm,
Many thanks for your attention on this, this is a real improvement over an already pretty good pieces of software.
- Andy
-
http://zaurus.mui.net/cedict-zb.dic (http://zaurus.mui.net/cedict-zb.dic) converted file
I've checked the file and there are some problems:
* You use DOS style end-of-line characters (0x0d, 0x0a) instead of UNIX ones (only 0x0a character). You probably run your script under Windows. All those DOS EOLs must be replaced with UNIX EOLs;
* The last byte in the file must be 0x00.
-
ok, I replaced the files with the corrected files. This time it displays things if nothing is typed, but it's all formatted goofy.
Any thoughts?
-
ok, I replaced the files with the corrected files. This time it displays things if nothing is typed, but it's all formatted goofy.
Again, could you put the file somewhere on the web?
...but it's all formatted goofy.
Could you be a bit more specific...
-
same locations.
http://zaurus.mui.net/cedict-zb.dic (http://zaurus.mui.net/cedict-zb.dic) after xerox
http://zaurus.mui.net/cedict-zb.txt (http://zaurus.mui.net/cedict-zb.txt) before xerox
-
http://zaurus.mui.net/cedict-zb.dic (http://zaurus.mui.net/cedict-zb.dic) after xerox
Are these new files? You have removed 0x0d characters from the hearder (properties), but not from the dictionary entries. And there is still 0x00 missing at the end of the file.
Please check CVS->libbedic->doc->bedic-format.txt at the project SF page:
http://sourceforge.net/projects/bedic/ (http://sourceforge.net/projects/bedic/)
-
ok, I think I followed your directions correctly this time. I think.
I removed the CR and left the LF in place, and added a NUL at the end of the file
This time it would lock up the system.
http://zaurus.mui.net/cedict-zb.dic (http://zaurus.mui.net/cedict-zb.dic) after xerox
http://zaurus.mui.net/cedict-zb.txt (http://zaurus.mui.net/cedict-zb.txt) before xerox
any thoughts?
Thanks!
-
when source archive will be available? OE has 0.9.1 now..
I would be nice if somebody could extend Makefile so it would be possible to build both OE and Sharp ROM releases. I cannot do it myself since I don't have OE and don't really want to invest time in learning how to compile for OE.
?
This is why we use qmake. We don't need fancy Makefiles, we just need a platform independent .pro file. Handcrafted Makefiles are only a source of trouble for distribution builders like us.
-
rafm:
The changes are nice, especially the new layout. One thing I noticed, though, is that this version still makes use of the Global::FindDocuments function, which is death for users with large media (I use a 4GB microdrive, loaded with files). Qtopia's file search functions are very, very slow.
As the LoadDictionaries function also uses a predefined path (as specified in the settings file), I don't think the global searching is really necessary; I compiled my own version, with the first half of the LoadDictionaries function commented out, and it works fine.
(On a side note, I don't know if you have access to the libbedic project, but I noticed that the files 'file.cpp' and 'file.h' are not in the CVS repository. I had to download the source tarball to get them.)
- ashikase
- anpachi, gifu, japan
-
ok, I think I followed your directions correctly this time.
Ken,
I can odly advise you to double check the format of your dictionary with the document at SF:
Browse CVS->libbedic->doc->bedic-format.txt at the project page:
http://sourceforge.net/projects/bedic/ (http://sourceforge.net/projects/bedic/)
Let me know if you have problem accessing this file.
The problems I can still see in your dictionary file (but there can also be something else):
* The last byte MUST be of value '0', not a string "\0"
* Entries MUST be separated with 0-byte, not EOL
* Each {s} tag MUST be closed with {/s} tag, and {ss} with {/ss}
-
This is why we use qmake. We don't need fancy Makefiles, we just need a platform independent .pro file. Handcrafted Makefiles are only a source of trouble for distribution builders like us.
qmake is nice and simplifies a few things. But I am not sure if it can also build .ipk packages (make ipk) and source distributions (make dist).
-
One thing I noticed, though, is that this version still makes use of the Global::FindDocuments function, which is death for users with large media (I use a 4GB microdrive, loaded with files). Qtopia's file search functions are very, very slow.
I am aware of this problem and I will try to address it in the next release. Briefly, my solution is:
* Remove search for dictionary files at startup
* Add a settings tab with the list of dictionaries found by Global::FindDocuments
** User can activate and deactivate listed dics
* At startup, load only activated dictionaries
This will speed up startup and eliminate a need for the paths to dictionary files. Global search will be run only when the user wants to add new dictionaries.
On a side note, I don't know if you have access to the libbedic project, but I noticed that the files 'file.cpp' and 'file.h' are not in the CVS repository. I had to download the source tarball to get them.
Thanks flagging that. I will add missing files to the CVS.
-
rafm:
The changes are nice, especially the new layout. One thing I noticed, though, is that this version still makes use of the Global::FindDocuments function, which is death for users with large media (I use a 4GB microdrive, loaded with files). Qtopia's file search functions are very, very slow.
As the LoadDictionaries function also uses a predefined path (as specified in the settings file), I don't think the global searching is really necessary; I compiled my own version, with the first half of the LoadDictionaries function commented out, and it works fine.
Could you post your compiled version?
I have a big CF and the same problem.
-
rehorrericha:
Here.
- ashikase
- anpachi, gifu, japan
-
rehorrericha:
Here.
- ashikase
- anpachi, gifu, japan
Hmmm...
But why is your compiled file one and a half Meg, as compared to slightly over 200 k for the unmodified zbedic file?
-
rehorrericha:
Yeah, I noticed it was rather big... it's probably linking to a library statically or something. I never bothered looking into it before, as I have plenty of space at the moment.
- ashikase
- anpachi, gifu, japan
-
zbedic 0.9.3 is available for testing at sf:
download (http://prdownloads.sourceforge.net/bedic/zbedic_0.9.3-0_arm.ipk?download)
This release may not be 100% stable since new features required some radical changes. Use at your own risk. Please report any problems (the best place is sf->bugs, but this forum is good as well).
ChangeLog:
27.11.2004 0.9.3 Rafal Mantiuk <rafm@users.sourceforge.org>
* Major change: improved loading of dictionary files
** Removed global search and path search at startup
** Added settings tab: search for dictionary files + switching dictionaries (in)active
** Added mime type for .dz files
** Loads only the dictionaries selected in the setings
(Should improve startup time for large SD/CF cards)
* Fix: Reorganized settings dialog (options categorized into 3 tabs)
* Added: Toolbar can be configured in the settings
* Added: "Switch dictionaries" button
* Added: Dictionary files can be openned from File Browser
* Added: Font preview in settings dialog
* Fix: Memory leak in libbedic
* Fix: Tool buttons are not hidden when larger font is selected on 5x00
* Change: Simplified state saving - only the word and dictionary is saved and restored
-
Great work! this program gets better and better...... has anyone got a list of keyboard shortcuts?
-
has anyone got a list of keyboard shortcuts?
How about checking the keyboard shortcuts in the program help
-
when i click help i get a blank page with a back button and a home button.... is there a help file to install?
-
when i click help i get a blank page with a back button and a home button.... is there a help file to install?
That is strange. The help files should install together with zbedic. Have you installed zbedic from an ipk archive? What language settings do you use? There are only English help files with zbedic.
-
I can odly advise you to double check the format of your dictionary with the document at SF:
Hi:
I'd like to make few dictionaries for myself from textfiles too.
Somehow though I'm not sure about the format exactly, though I have inspected the files.
All the available dictionaries seem to be compressed with SHCM, so I can't inspect their internal text layout.
Could you provide me with a simple non-compressed dictionaries of even just three lines of terms, so I could copy the structure exactly and produce my own?
Thank you.
-
I'd like to make few dictionaries for myself from textfiles too.
Somehow though I'm not sure about the format exactly, though I have inspected the files.
All the available dictionaries seem to be compressed with SHCM, so I can't inspect their internal text layout.
Could you provide me with a simple non-compressed dictionaries of even just three lines of terms, so I could copy the structure exactly and produce my own?
Thank you.
SHCM compression is not used any longer as it is less efficient than dictzip.
You can use dictunzip to decompress all *.dic.dz files. See: http://cvs.sourceforge.net/viewcvs.py/bedi...v=1.2&view=auto (http://cvs.sourceforge.net/viewcvs.py/bedic/libbedic/doc/bedic-format.txt?rev=1.2&view=auto) -> IV. Step by step guide for building new dictionaries
-
You can use dictunzip to decompress all *.dic.dz files. See: http://cvs.sourceforge.net/viewcvs.py/bedi...v=1.2&view=auto (http://cvs.sourceforge.net/viewcvs.py/bedic/libbedic/doc/bedic-format.txt?rev=1.2&view=auto) -> IV. Step by step guide for building new dictionaries
I looked but didn't find it at the dict.org site: do you know where I could download dictzip and dictunzip for either the Zaurus or for Windows?
-
I looked but didn't find it at the dict.org site: do you know where I could download dictzip and dictunzip for either the Zaurus or for Windows?
I used Debian package 'dictzip' from the testing distribution. You can try to compile it from the sources or use Linux machine. You will need this software anyway to prepare the final dictionary as non-compressed files are awfully big.