| 주제 내 페이지: [1 2] > | DVX 8.0.505 스레드 게시자: Marcus Malabad
|
|---|
The added features of DVX2's latest build (505) are fantastic. The Code Zapper, in particular, is a great tool as it saves you the time and energy of cleaning up Word files, especially those converted from PDF. Download the latest build, folks!
Thanks Atril for the great job! Please still using Trados should migrate to DVX now.
Marcus
| | | | | Here's what I think | Nov 18, 2011 |
I have downloaded the trial version of this build and tried to use it on a real project of mine, consisting of many small Excel files. Here follow some of my impressions (some concern DejaVu in general, some conern the latest build in particular):
THE GOOD
— The Advanced Project Explorer is brilliant! It makes importing the files of your project very efficient.
— The "Parse embedded HTML" import option is fantastic: it not only converts HTML tags to Dej... See more I have downloaded the trial version of this build and tried to use it on a real project of mine, consisting of many small Excel files. Here follow some of my impressions (some concern DejaVu in general, some conern the latest build in particular):
THE GOOD
— The Advanced Project Explorer is brilliant! It makes importing the files of your project very efficient.
— The "Parse embedded HTML" import option is fantastic: it not only converts HTML tags to DejaVu tags, but it actually parses tags such as line breaks and segments the files based on them.
— I love all those navigation options when I'm translating: Ctrl/Alt and up/down, Alt and left/right.
— The AutoComplete feature proved to be of help in most cases; not as an aid to make typing faster (I already type very swiftly), but as an aid to improve stylistic consistency (this is the whole deal with "sub-segment leveraging", I believe).
— I loved the ease with which I could edit source segments. No need to hit a function key; you just put the cursor in a source cell and edit it. Plus, you can run search & replace in source segments, which I haven't been able to do with other tools.
— I like that you can have your projects and TMs and TBs stored in the folders YOU chose and then back them up or transfer them between computers easily, without having to back up any hidden files in the Program Data folder of Windows (I really hate this folder). Plus, this 'architecture' allows for making use of Windows 7 Jump Lists: you pin DejaVu X2 on your Task Bar, right click on it and have easy access to your most recent projects and TMs and TBs.
— I like that you insert the names of your clients once and DejaVu stores them in a neat list, which can save you lots of inconsistencies later on. I also like the list of fields/subjects, although it could be organized in a much better way instead of being a tiny, cramped drop-down list.
THE BAD
— Still not the slightest formatting: no bold, no italics, no underline, nothing. All these are represented by codes, which can lead to an overwhelming amount of them in complex files.
— The Excel filter is still lacking: you can't specify columns that contain comments or context IDs, which has proved to be of tremendous help in software/smartphone/website localization projects.
— Inserting hits from what AutoSearch has found (I don't know how this list is called, in memoQ it's called "Translation Results") is not very efficient: are all these keyboard shortcuts necessary?
— There is a general bug which mistakes punctuation marks for letters: as a result, when you hit Ctrl+Shift+N to shift a word to the right, there is no way you can put it before a comma or a full stop, because the program thinks the comma or the full stop belong with the word that precedes it. Also, this bug means that you get results like word) (i.e. a word with a right bracket attached to it) in the AutoComplete list.
— The DeepMiner is such a slow-moving monster, even on my speedy system with an i7 processor, 6GB of RAM and fast HDDs partitioned with care.
— Undo works in a very, very esoteric way: you expect it to undo one thing and it always seems to undo another (e.g. it can't undo splitting of a segment).
— A simple SQL command I used against a TM of mine worked when I inserted it in one "command line" and it threw an error when I inserted it in another!
— There is no way to create views.
— "Fuzzy terminology recognition" is too fuzzy a mechanism in its own right. The user should be allowed to separate the stem of a word from its suffix himself. Even better, the user should be able to use all sorts of wildcards when entering a term in the database, since not only suffixes but also the accented letters can vary, in the different cases of a noun/adjective or in the different persons of a verb.
THE BOTTOM LINE:
— DejaVu still has a long way to go before I consider it as a viable alternative to the tool I currently use.
— For what it offers, I find it quite overpriced.
So, fellow translators, what's your take on this latest build or on DejaVu in general?
[Edited at 2011-11-18 11:03 GMT] ▲ Collapse
| | | |
What tool are you using now?
| | | | | My current tool | Nov 18, 2011 |
I'm using memoQ 5.0
| | |
|
|
|
Thank you for your reply.
- The issue about an SQL statement working in one place but not in another could by explained by the fact that in some places DVX2 fills in, so to speak, part of the complete query for you. For example, the SQL filter in the Row Selector automatically add the "SELECT Sentence FROM TranslationMemory WHERE" part of the SQL clause, so you can just type in a set of conditions and it will find the segments that fulfill them. However, if you open a TM and use the... See more Thank you for your reply.
- The issue about an SQL statement working in one place but not in another could by explained by the fact that in some places DVX2 fills in, so to speak, part of the complete query for you. For example, the SQL filter in the Row Selector automatically add the "SELECT Sentence FROM TranslationMemory WHERE" part of the SQL clause, so you can just type in a set of conditions and it will find the segments that fulfill them. However, if you open a TM and use the "Execute SQL" Déjà Vu will require that you type the complete query (since there you can use other SQL commands, like DELETE). Thus, if you were to use the SQL string you used in the first case with the second case, Déjà would tell you that the SQL statement you typed is not valid. ▲ Collapse
| | | | | You make some good points Epameinondas | Nov 18, 2011 |
This is what I'd say about your "THE BAD" list:
- Bold, italic formatting is supposed to be coming soon for Office files
- I agree that the Excel filter needs to be improved
- For inserting AutoSearch portions, you usually only need to press control + 1, control + 2, etc. You can also just double click on the portion.
- I believe the shift left/right bug is being fixed
- DeepMiner does indeed need speeding up. Try switching it off for the time being. It does ... See more This is what I'd say about your "THE BAD" list:
- Bold, italic formatting is supposed to be coming soon for Office files
- I agree that the Excel filter needs to be improved
- For inserting AutoSearch portions, you usually only need to press control + 1, control + 2, etc. You can also just double click on the portion.
- I believe the shift left/right bug is being fixed
- DeepMiner does indeed need speeding up. Try switching it off for the time being. It does continue to work with AutoWrite though and can sometimes retrieve real jems well-hidden in the TM.
- Undo is being overhauled
- SQL is being made more user friendly with a “library” of statements
- Have a look at “view” in the help (All files view and SQL statement) and and “rows selector”. I think you'll find that DVX2's “View” feature is better than any other tool's
- Yes, the terminology recognition mechanism does need some attention ▲ Collapse
| | | | | The most serious bug | Nov 18, 2011 |
I forgot to mention the most serious bug I came across, which forced me to abandon my trial of DejaVu X2 (build .505): from a point onwards, every time I moved to a new segment (by hitting Alt/Ctrl+Down/Right) I had to hit the first letter of my translation twice! At first, I thought that I might be typing too fast for DVX to catch up because of AutoSearch or something like that. But the problem persisted even at ridiculously slow speeds.
Atril's support denied the existence of this... See more I forgot to mention the most serious bug I came across, which forced me to abandon my trial of DejaVu X2 (build .505): from a point onwards, every time I moved to a new segment (by hitting Alt/Ctrl+Down/Right) I had to hit the first letter of my translation twice! At first, I thought that I might be typing too fast for DVX to catch up because of AutoSearch or something like that. But the problem persisted even at ridiculously slow speeds.
Atril's support denied the existence of this bug, and that was the end of my experience with this tool. So, for the time being I'll deny the existence of DVX2 and possible consider asking for a new trial code when DVX3 comes out. :-\
[Edited at 2011-11-18 17:11 GMT] ▲ Collapse
| | | | | Possibly a problem with your keyboard | Nov 18, 2011 |
Epameinondas Soufleros wrote:
I forgot to mention the most serious bug I came across, which forced me to abandon my trial of DejaVu X2 (build .505): from a point onwards, every time I moved to a new segment (by hitting Alt/Ctrl+Down/Right) I had to hit the first letter of my translation twice! At first, I thought that I might be typing too fast for DVX to catch up because of AutoSearch or something like that. But the problem persisted even at ridiculously slow speeds.
Atril's support denied the existence of this bug, and that was the end of my experience with this tool. So, for the time being I'll deny the existence of DVX2 and possible consider asking for a new trial code when DVX3 comes out.
I work with DVX2 for almost 10 hours a day and I have never experienced this problem. Moreover, I don't remember anybody mentioning it at Yahoo Deja Vu group. So it looks like a keyboard problem. But if there is really such a bug, then I am sure Atril will fix it immediately.
Do not wait for DVX3, it may take long years. Atril's policy is different from other CAT tools. The others launch a new version almost every year and ask for upgrade fees. But Atril releases a new build every two months with new features and does not ask you to upgrade. Moreover, if you wait for DVX3 then you will have to work with MemoQ. How do you manage your TMs and TBs with MemoQ
Selcuk
| | |
|
|
|
| Not a keyboad problem | Nov 18, 2011 |
Hello Selcuk, thanks for your reply. No, my keyboard works fine, no key is stuck or anything like that. I work long hours with memoQ and many other applications and do a lot of typing without problems. This problem only appeared in DejaVu.
As for managing TMs and TBs, I think it'll take me about a decade to learn how to deal with them using SQL, so I'll be ready just in time for DVX3!
[Edited at 2011-11-18 17:3... See more Hello Selcuk, thanks for your reply. No, my keyboard works fine, no key is stuck or anything like that. I work long hours with memoQ and many other applications and do a lot of typing without problems. This problem only appeared in DejaVu.
As for managing TMs and TBs, I think it'll take me about a decade to learn how to deal with them using SQL, so I'll be ready just in time for DVX3!
[Edited at 2011-11-18 17:30 GMT] ▲ Collapse
| | | |
I will test it again, and if I find any problem report it to Atril.
As for SQL, this is an additional feature in DVX. Normally you do not need to use SQL Statements or Filters for daily operations.
And you do not need to create SQL commands every time, you can save them in DVX and use whenever you need them.
One of the SQL Filters I use frequently and therefore saved in the settings is
ID IN (SELECT ID FROM Translations WHERE Client = '001')... See more I will test it again, and if I find any problem report it to Atril.
As for SQL, this is an additional feature in DVX. Normally you do not need to use SQL Statements or Filters for daily operations.
And you do not need to create SQL commands every time, you can save them in DVX and use whenever you need them.
One of the SQL Filters I use frequently and therefore saved in the settings is
ID IN (SELECT ID FROM Translations WHERE Client = '001')
It displays all sentence pairs entered for Client 001
Subject and Client metadata is important in DVX, when used properly you will get results only for a specific subject and client. ▲ Collapse
| | | |
Yes, SQL gives flexibility, but it should not be thrown in as-is; it should have some 'friendliness' around it. I'm neither a machine nor a programmer to like having to type a statement like the one you mention, referring to a client by their code. The number should be 'interpreted' by DejaVu and show the corresponding client's name for you to select it in a friendly expression builder, in which expressions will be closer to natural, human language rather than to artificial language.
... See more Yes, SQL gives flexibility, but it should not be thrown in as-is; it should have some 'friendliness' around it. I'm neither a machine nor a programmer to like having to type a statement like the one you mention, referring to a client by their code. The number should be 'interpreted' by DejaVu and show the corresponding client's name for you to select it in a friendly expression builder, in which expressions will be closer to natural, human language rather than to artificial language.
DejaVu should gather the most frequently used 'statements' and provide them to the user from within the application. What's the point of listing them in the user manual? Isn't it a fuss to have to open the PDF guide, find the expression you want, copy it, paste it, modify it and execute it? Even the thought of that procedure gets me tired and frustrated!
[Edited at 2011-11-18 17:54 GMT]
[Edited at 2011-11-18 17:54 GMT] ▲ Collapse
| | | |
Epameinondas Soufleros wrote:
Yes, SQL gives flexibility, but it should not be thrown in as-is; it should have some 'friendliness' around it. I'm neither a machine nor a programmer to like having to type a statement like the one you mention, referring to a client by their code. The number should be 'interpreted' by DejaVu and show the corresponding client's name for you to select it in a friendly expression builder, in which expressions will be closer to natural, human language rather than to artificial language.
DejaVu should gather the most frequently used 'statements' and provide them to the user from within the application. What's the point of listing them in the user manual? Isn't it a fuss to have to open the PDF guide, find the expression you want, copy it, paste it, modify it and execute it? Even the thought of that procedure gets me tired and frustrated!
Client Names are stored in the settings file, both with a code and full name. And they are accessible with 3 clicks only. Yes, just 3 clicks!
We both look at the same thing but see different things. What you see is difficulties, new challenges, something new and not available in other CAT tools. (Actually, SQL commands were available since the launch of DVX1, but they are more friendly in DVX2).
And I see opportunities. I have a large TM or TB and I can make better use of it with SQL. I can add new fields to my TB and search/edit it with SQL.
But as I have mentioned before, TM and TB management even without SQL commands is much more better compared to other tools. In Trados 2007 and 2009, you cannot move to the last segment with just one click. It may take hours depending on the size of the Trados TM. In MemoQ, deleting some hundreds of segments means hundreds of clicks.
Most frequently used 'statements'? I don't think there is such a list. There are some SQL commands/statements that I never use, but other DVX users may use them.
Just try to see the extra opportunities. Your large TM/TB may store and display more information than those displayed by your favourite CAT tool.
| | |
|
|
|
| Varia... DVX2 vs memoQ... | Nov 20, 2011 |
Epameinondas Soufleros wrote:
I have downloaded the trial version of this build and tried to use it on a real project of mine, consisting of many small Excel files. Here follow some of my impressions (some concern DejaVu in general, some concern the latest build in particular):
THE GOOD (...)
— I loved the ease with which I could edit source segments. No need to hit a function key; you just put the cursor in a source cell and edit it. Plus, you can run search & replace in source segments, which I haven't been able to do with other tools.
In fact, I think the way DVX handles the source segment is rather a problem.
I.e., I think the source cells should be protected by default 'cause if they become corrupt, the file can't be exported or is damaged.
DVX has sometimes focus problems and a source cell content may be replaced by the content of another cell or simply overtyped.
It happens to me in a regular way.
Not very often, but it happens (I suppose the problem is the Edit in the separate test area option I use).
Nonetheless, I think the locking/unlocking should be global and should permit the S&R operations in the source cells.
THE BAD
— Still not the slightest formatting: no bold, no italics, no underline, nothing. All these are represented by codes, which can lead to an overwhelming amount of them in complex files.
Frankly speaking, I would have both options, something like in Trados Studio.
Nonetheless, generally, I prefer codes but it's very personal.
BTW, these abstract codes permit a better leverage of the TM data.
— The DeepMiner is such a slow-moving monster, even on my speedy system with an i7 processor, 6GB of RAM and fast HDDs partitioned with care.
Indeed, the DeepMiner is damn slow for huge TMs but, unlike in memoQ, you can open the same project twice in two DVX2 instances.
In the first one, you run the pretranslation with the DeepMiner enabled, in the second one you translate interactively with the AutoAssemble option disabled.
— Undo works in a very, very esoteric way: you expect it to undo one thing and it always seems to undo another (e.g. it can't undo splitting of a segment).
True.
And some undo operations performed outside of the active segment may cause segment/project corruption.
memoQ is far better here.
— There is no way to create views.
The views can be created using SQL.
The problem is the SQL is neither easy nor intuitive, especially for the beginners.
mQ is far easier here.
But the incontestable DVX advantage is one can open simultaneously a view and a file containing the same segment.
— "Fuzzy terminology recognition" is too fuzzy a mechanism in its own right. The user should be allowed to separate the stem of a word from its suffix himself. Even better, the user should be able to use all sorts of wildcards when entering a term in the database, since not only suffixes but also the accented letters can vary, in the different cases of a noun/adjective or in the different persons of a verb.
Indeed, memoQ is far more precise but the term definition is very slow for heavily inflected languages.
E.g., recently, my wife did her first project in memoQ.
Generally, she found it nice but the way mQ handles the terminology was simply unacceptable for her.
IMO the best solution would combine both approaches.
THE BOTTOM LINE:
— DejaVu still has a long way to go before I consider it as a viable alternative to the tool I currently use.
Well, memoQ has multiple incontestable advantages but DVX is the most flexible zero administration tool for small teams and it's core "linguistic" features like EBMT and AA (using both TMs and termbases) are still simply the best.
Another point is mQ has some basic internal architecture flaws which are crucial for the effective use, i.e. it doesn't permit concurrent data access and its word definition (a character chain between two spaces) seriously affects the term recognition for "compound" words, including the typical case of contracted articles (e.g.) in French, e.g. if you have homme in your termbase, l'homme will not be recognized which is a shame.
DVX has no problems here, so, it's very unlikely I'll shift definitively to mQ, at least in my main language pairs.
AFAIK, it's very unlikely the Kilgray guys will change their word definition in a predictable future, I asked for it two years ago with no success.
— For what it offers, I find it quite overpriced.
Well, I don't think so.
The general TCO of DVX was always low because it's no frequent/annual upgrades.
I upgraded DVX to DVX2 after 8 years, so the initial 1000 euro purchase corresponded to 125 EUR per year, in fact, it was one of the cheapest CAT tools I ever bought.
The DVX2-DVX3 shift may be different but I expect, let's say, 3 or 4 years with no upgrade/maintenance fees.
So, fellow translators, what's your take on this latest build or on DejaVu in general
Decidedly far from being perfect but getting better.
I simply can't imagine to shift back to DVX.
Catspeed
GG
[Edited at 2011-11-20 11:24 GMT]
| | | | | Separate text area? | Nov 20, 2011 |
Selcuk Akyuz wrote:
Epameinondas Soufleros wrote:
I forgot to mention the most serious bug I came across, which forced me to abandon my trial of DejaVu X2 (build .505): from a point onwards, every time I moved to a new segment (by hitting Alt/Ctrl+Down/Right) I had to hit the first letter of my translation twice! At first, I thought that I might be typing too fast for DVX to catch up because of AutoSearch or something like that. But the problem persisted even at ridiculously slow speeds.
Atril's support denied the existence of this bug, and that was the end of my experience with this tool. So, for the time being I'll deny the existence of DVX2 and possible consider asking for a new trial code when DVX3 comes out.
I work with DVX2 for almost 10 hours a day and I have never experienced this problem. Moreover, I don't remember anybody mentioning it at Yahoo Deja Vu group.
IMO it may happen when one works in the separate text area.
I don't experience the bug described by Epameinondas but, e.g. when I type Alt+0132, Alt+0171 etc, I must sometimes do it twice or type a space.
It looks as a variant of the Epaimeinondas' case.
So it looks like a keyboard problem. But if there is really such a bug, then I am sure Atril will fix it immediately.
Normally it's enough to restart DVX but it should not happen.
The separate text area is simply handled incorrectly, it's a recurring problem.
Catspeed
GG
| | | | | You don't need SQL to "create" views... | Nov 22, 2011 |
The views can be created using SQL.
The problem is the SQL is neither easy nor intuitive, especially for the beginners.
mQ is far easier here.
Views aren't really "created" at all in DVX like they are in mQ (named views). They're just there at the click of a button. Just click the project icon for the "All files" view or use the rows selector to select status based views (all fuzzy matches, etc.), the A-Z button to sort alphabetically, the filter on selection buttons to obtain a view of all source and/or target rows containing a particular text or the checkboxes in the rows selector SQL statement option to display other views.
You can, of course, also use SQL to display more advanced views.
I'd say that it's much easier to display (and use) views in DVX than mQ. Views are one of mQ's weakest points.
| | | | | 주제 내 페이지: [1 2] > | To report site rules violations or get help, contact a site moderator: You can also contact site staff by submitting a support request » DVX 8.0.505 | CafeTran Espresso | You've never met a CAT tool this clever!
Translate faster & easier, using a sophisticated CAT tool built by a translator / developer.
Accept jobs from clients who use Trados, MemoQ, Wordfast & major CAT tools.
Download and start using CafeTran Espresso -- for free
Buy now! » |
| | Protemos translation business management system | Create your account in minutes, and start working! 3-month trial for agencies, and free for freelancers!
The system lets you keep client/vendor database, with contacts and rates, manage projects and assign jobs to vendors, issue invoices, track payments, store and manage project files, generate business reports on turnover profit per client/manager etc.
More info » |
|
| | | | X Sign in to your ProZ.com account... | | | | | |
Login to reply/comment 
