Introduction

For further discussion, see the DVX user group, at: http://tech.groups.yahoo.com/group/dejavu-l/

RESULTS OF ASSEMBLE

1. Still, the numbers bug. It occurs even in rows with a single word and single digit: With my database entries, 'Fig. 1' became 'Fig. 1-1', etc. It's caught by the Numbers Check but at this stage in the life of the program, it shouldn't happen at all.
2. Better handling of numbers in common phrases. For example, dates: March 9, 2011 ==> 9 de marzo de 2011. This form should match regardless of the actual number. Other MT and TM programs have had this for years.
3. Sometimes the current source sentence has a phrase that appears in both the MDB and TDB, perhaps several times in the latter. Some of the entries in the MDB end in punctuation, for example a colon. DVX is offering those entries during Assemble. They're not terrible suggestions, but they include the punctuation, and there are better candidates in the TDB that do NOT have that punctuation. I'm not sure of the best approach, but I was thinking: [BT*]
1. If there is no punctuation in the source phrase, exclude phrases in the MDB that include punctuation. Or, omit the punctuation when retrieving this term.
2. Do not consider MDB entries at all during Assemble.
4. Options to display text retrieved from the TDB, MDB, or Lexicon in different colors. Also a different color for source text that was inserted without being matched with a potential translation ("failed text'). [+BT]
5. Option to prevent sentences shorter than X words from being sent to the MDB by AutoSend, where X is user-adjustable.

EDITING

1. Capitalization: The Shift-F3 key cycles between (a) all lowercase, (b) all uppercase, (c) First letter only in uppercase, balance lower case. I want at least one more option, such as:
   
   (d) Title Caps: first letter of every word in uppercase.
   Better yet:
   (e) Real Title Caps: Do not capitalize 'unimportant' words. This would have to be implemented by language, of course, but considering all the magic that DVX already does, this would be straightforward.
2. Keyboard method to hide or display formatting codes in the source. This was available in DV 3.0. It would still be  handy to see the source text without being obscured by the codes in a heavily formatted sentence.
3. Option to lock the source column and prevent accidental changes.

USER INTERFACE

1. Disable mice scroll buttons so that they do not scroll in the row selector.  If you select a type of row to view, then forget to click elsewhere before scrolling, the display changes to a different type of row. Drives some people nuts.
2. When working on sets of rows accessed via the Select feature of the SQL command, when changing from one set of rows to another one, the Sentences Found window is never updated; it remains with the information of the last row accessed in the previous SQL window, which is rather a nuisance. To update that window, one has to click on any other row, and then click back on the first row. [entered in DVX2 bug tracking system]

QA, TERMINOLOGY CHECKING

1. Add an option to each MDB and TDB that is attached to a project: whether to include it in a terminology check. I use the TDB as a source of possibilities, and to save typing, but there are so many entries that a term check is time-consuming and almost useless. I have included a 'check TDB' * as an additional TDB, which helps. I want DVX to use that TDB during the term check, and totally ignore the other one (the 'big Papa' TDB). For now I have to 'remove' Big Papa before running the term check.
   More discussion here: http://www.techlanguage.com/tips/term_checks.html

ADDING TO / RETRIEVING FROM TDB

1. When DVX produces an Assembled translation, it would be straightforward for the program to remember what source term(s) resulted in what target term(s). It should also be easy to keep track of the new word(s) that I type to replace the initial choice(s) made by DVX. What would then be 'nice' is, when I select a word or phrase on the target side, for DVX to automatically highlight that choice on the source side. That way, I could then press F11 to send the pair to memory. Even if it's wrong 10 or 20% of the time, this would save time.
2. Better control over capitalization when (a) storing terms and (b) retrieving from the TDB. For example, one keystroke should send the terms with all characters in lowercase.
3. What's missing is a step to add the term pair to TDB with Client and Subject, but *without* adding the whole source and target pair as Context, and thus also without an intermediate dialogue step.
4. An option/function that would enable the user to see the context information for the term selected in the Autosearch - Portions window, without having to go through Ctrl-L. [JV]
5. Display Definition field in TDB entries in the AutoSearch window.

IMPORT / EXPORT

1. DVX asks whether we want to add spaces at the ends of our target sentence. Please provide the option to always say 'Yes!' without asking.
2. The current default location for target export is the folder that was defined during the last project. I keep exporting translations to the last folder. Instead, DVX should offer the folder that was defined as the source for the current project. Same for External View.

MISCELLANEOUS

1. Could Atril please explain exactly what the 'Reverse' attribute does when we attach a database? It's not covered in the 622-page Manual or Help system.
2. The task 'Send project to MDB' takes all computer resources, locks up. Can this be set to a lower priority to allow us to use other programs (email, web browser) during that process?
3. Full two-column display of database entries, as we had in DV 3.0. Now we see several source entries but only one target entry at a time. More difficult to compare duplicate entries and decide which one to delete.