Messages

Just a short note on the zip classes: Some time ago I had a very quick look at the classpath zip classes (so that DfM could read the JAR-file via JSR 75). For me it seemed that it would take quite some effort to make these classes available for Java ME. Hope you manage to do it !

Gert

Besides, for the file browser we received a source code that we may use. See here: http://dictionarymid.gottfried-signs.ch/index.php?topic=38.0

Gert

Colleagues,

Steven provided a "Directory Chooser" that we can use for selecting the dictionaries that are accessed via JSR-75.

Maybe someone can integrate this "Directory Chooser" into DfM.

Best greetings,
Gert

Joshua,

great to know that you want to support the Development of the DictionaryForMIDs project !!

Yes we are using CVS for maintenance of the source code. You can use the anonymous CVS access to get the latest source. If lateron you completed some updates and you want to check in your updates to CVS, you need to get developer access to CVS.

And yes, we usually use Sun's WTK for compiling and building.

Best greetings,
Gert

Thanks for your feedback !

Well, I only have some basic knowledge on j2me programming, but hopefully I could be of some help in this project.

There are so many things where we need support on our project - most jobs are not even related to Java programming ! Just have a look at the "We need your support" section of our homepage.

Best greetings,
Gert

Hey Penny,

it is "for MIDs"; MIDs stands for Mobile Information Devices

Best greetings,
Gert

Just for me so that I understand correctly:

There is file 1: http://prdownloads.sourceforge.net/dictionarymid/DictionaryForMIDs_2.7.0_EngChi_CEDICT.zip?download

Did you try to set the logLevel to 3 for this one ? Did it work ? If yes, could you please report the results (i.e. the times of the log output) ?


And there is file 2: http://www.kugihan.de/dict/download/test_versions/cedict_bigger_files/DictionaryForMIDs_3.1.0_ZhEng_Cedict_BM.zip

File 2 does not work at with logLevel = 3, is this correct ? But it worked with the original setting of logLevel=0 ?

Thanks for clarifying !!
Gert

I just googled for that error: please read http://developerboards.cingular.com/cngddp/board/print?board.id=techsupport&message.id=1874&format=one

Is it true that those dictionaries where you have this error have a MIDlet-Name of 32 characters ?

Greetings,
Gert

The error on the 'empty' version is a correct one, because in the 'empty' version there is no file DictionaryForMIDs.properties (besides, will be different for 3.1.0).

The RecordStoreException however is a real bug (whether on your L7 or DfM is another question).

Can you please post the version numbers of the DfM application that you tested and state which ones do work on your L7 (such as the IDP dictionaries that you tested) and which ones do not work. Maybe the problem does show up only in newer versions of DfM ?

Another question: is the RecordStoreException the first error message that is shown ? And is it always this error message that is shown (or is sometimes the first error message a different one) ?

Thanks !
Gert

And the only thing that is different is the following line  ?

logLevel: 3

Gert

Besides, could you report the response time on this version:
http://www.kugihan.de/dict/download/test_versions/cedict_bigger_files/DictionaryForMIDs_3.1.0_ZhEng_Cedict_BM.zip

This one has fewer files with bigger sizes. Will be interesting to know if the response time is better !

Gert

In the ZIP archive there is the file DictionaryForMIDs_EngChi_CEDICT.jad. Open this file in any editor. On Windows, use wordpad or notepad. Then just change the last line from
logLevel: 0
to
logLevel: 3

Save the file, then re-install the JAD-file.

Please keep asking !
Gert

Thanks ! Let me know if you should have any trouble to edit the JAD-file.

Gert

Thank you for your feedback !

Just a few quick words from me:

1. (some characters not correctly displayed): That problem will be solved with the bitmap fonts.

2. (response time slow): Please do read this posting: http://dictionarymid.gottfried-signs.ch/index.php?topic=34.0

Regards,
Gert

Dear DictionaryForMIDs users,

we are receiving reports that on some devices the response time can be rather bad.

In general, DictionaryForMIDs has a highly optimized search engine, and response time may be less than 100ms, depending on the device and on the dictionary. However, it seems that for some devices on large dictionaries it takes a few seconds for a translation.
A first analysis showed that there are devices where the time for accessing the files in the JAR archive does increase significantly when there is a large number of files in the JAR archive. That means that on those devices not the size of archive is the problem, but only the number of files (i.e. making less bigger files speeds up the search). Also, those devices have long startup times when there are many files in the JAR archive. We need to check this further.

If you like to do a measurement on your cell phone or PDA, then please do the following:

1. Download the file http://prdownloads.sourceforge.net/dictionarymid/DictionaryForMIDs_2.7.0_EngChi_CEDICT.zip?download and install that application on your mobile/PDA

2. Start the application and activate under Settings the "Show statistics"

3. Check that the translation direction is "English to Chinese", then translate the word "homeless"

4. Post the "Response time [ms]" here in this forum

5. If the response time is bigger than 1000 ms, continue with step 6

6. Close the application then in the file DictionaryForMIDs_EngChi_CEDICT.jad, change the line
logLevel: 0
to
logLevel: 3

7. Re-install and start the application then translate again "homeless"

8. Please post the log output following to the message "start translation" ("Time for open file", "Time for read file", ...) here in the forum


A general note: the translation time does increase with the number of found results ("hits"). So for example when a translation has 5 results, then it is normal that it takes more time than for 1 result.

Thank you !
Gert

Update, 13.04.2007: In step 3, please do this translation twice ! The time for the second translation will probably be less than the first, because on the first translation the Java code is compiled to native code for the CPU (typically a CPU with the ARM architecture).