Bad PDF Converter: ABBYY 14

Anyone thrilled with ABBYY 12 for converting pdf documents to Word processor format may wonder why Amazon does not sell version 14. Version 12 is all sold out, and you have to go to the company website to even order 14. For specific information be prepared to paw through some inelegantly translated descriptions.  It turns out that ABBYY 14 is to ABBYY 12 as Windows Vista is to Windows 2000. See for yourself.

Here is how a chunk of Brazilian legalese converted with ABBYY 12 loads into memoQ (a standard translation-assistant tool):

Here is how that same section imports after conversion with ABBYY 14:

The reason I converted again with ver. 12 was that it would take ten times longer to unmangle the ver. 14 output document and make it translatable. The mess of formatting codes you see was left over after I polished up the fonts for consistency, leaving no font stretching, shrinkage or mixing of sorts. ABBYY 14 produces conversions that are absolutely useless for translation and nearly impossible to clean up using the current version of Winword. Small wonder nobody but ABBYY wants to sell it! But there is hope.

While I was struggling to find a way to buy the program (after exhausting all possibilities of purchasing another ABBYY 12) I asked about the possibility of converting without MS Office installed. Nobody at ABBYY had a clue whether their new product would work with Apache Open Office. Omega-T, a free, open-source translation tool, works fine with Microsoft’s competitors. As soon as I installed ABBYY 14 on a clean machine, I tested it with Apache Open Office and no MS Office installed. The conversion worked! The resulting word processor file looked presentable, needing only the usual touch-ups

Once I get caught up I’ll reopen the ABBYY 14 file and save it using the Apache open source program to see if it makes any improvement. Some pdf converters litter the output with “text boxes” that move unpredictably and mess up translations. Opening and saving the resulting Word file with Microsoft’s simple Wordpad gets rid of those, and the resulting file is often salvageable for translation with professional warez. Stay tuned. In the meantime, Caveat Emptor! One sweet solution could be for ABBYY to rename ver. 12 as ver. 14 and forget all about the hideous v.14 miscarriage–kind of the way Microsoft did after releasing Windows Vista.

UPDATE: ABBYY techs suggested using the formatted text option to export converted files. That “solved” the most distressing problem. The interface, however remains user-hostile and even with that work-around all progress is much slower than in ver. 12

2+2=4

Orwellian quality translations

If in need of translations involving oil, dams, power plants or contracts, look us up.
Visit the immigration-related blog too…

My book on the Crash and Depression is out in Portuguese and English on Amazon.com

Houston Translator Association Irregularities

The Houston Translators and Interpreters Association has in recent years been a model of competence in the industry. Yet the current bylaws amendment ballot looks more like a model of practices to avoid. The online bylaws dated April 14, 2010, define members as follows:

Article III – Membership…

Section B – Classes and Qualifications

The Association has three (3) membership classes: individual, corporate and institutional.

  • Individual: An individual who is engaged in translating, interpreting or related work (and may include students)

  • Corporate: A business with an interest in translation or interpretation

  • Institutional: An institution with an interest in translation or interpretation.

Directors elected in single-candidate elections now propose to change those member classes by creating a special class of students who at this time would not be allowed to vote to elect their teacher nor be listed in the online members directory (where the public expects to find professional linguists). To propose the change, voters were told that “new text is indicated by underlining, deleted text by strikeout.” But for the ballot proposal sent to members to change the bylaws, the board of directors approved the following:

Section B – Classes and Qualifications

The Association has four (4) membership classes: individual, corporate, institutional and student.

The above introductory sentence (followed by four, not three bullet items) appeared with no underlining for the new text nor strikeouts for deleted language. It gives the incorrect impression–instructions elsewhere to the contrary notwithstanding–that the student category already exists whether one likes it or not, and that there is mere quibbling to be decided on some trifling point of verbiage in the last of four preexisting bullet points.

In an association of quilt-makers, brewers, basket-weavers or kickboxers, the omission might be brushed aside as simple incompetence, the result voided and new ballots produced. Indeed, one such error in ballot translation into Spanish for the Texas State government had precisely that outcome and cost taxpayers about $100,000 to reprint.

The bad ballot language at issue, however, is presented as approved by the very people immigrants depend on for legal defense of their individual rights in courts that order execution by letal injection. Credible fear reviews can shield dissenters or whistleblowers from extrajudicial execution or torture by junta-style dictatorships, and HITA hosted a presentation on those. Professionals educated abroad want their syllabi competently translated with all legalities accurate so they may exercise a profession despite entrenched lobbyists erecting barriers to entry.  Our newsletter and web tips just now alerted linguists of at least a dozen different fraudulent scams. But more perfidious scams are perpetrated from within the profession. Must we circulate ballots that are an indictment of the board’s competence to frame and edit a simple bylaws amendment proposal?

For over a decade beginners were advised by prominent HITA and AATIA members not to bother to apply for municipal and county court interpreting in Texas. From a position of public trust they emphatically proclaimed that a license was required as a prerequisite. Nevermind that this was a law urged by three individuals claiming to represent the profession without their lobby efforts appearing in our trade publications. The persistent lie was finally exposed at a regulatory meeting at which a government regulatory attorney explained on the public record that the law meant nothing of the sort.

The old law merely formalized a procedure for showing an incompetent interpreter the door and ordering up a substitute, typically someone grandfathered in irrespective of real credentials or ability.  The dissembling was a sales platform for quickie diploma mill courses pushing test answers, podiums for grandfathered insiders to talk down to aspirants, and a loophole enabling agencies possessed of counsel to quietly and without fanfare exploit inexperienced youngsters at pauper rates. The law was only repealed after a libertarian interpreter put up a website playing a recording of the regulatory lawyer’s explanation in language too clear and simple to falsify.

If sidetracking students from earning a degree liable to make them employable is deemed a good idea, it ought to be passed by honest vote of fully-informed members using a ballot prepared in conformity with its own instructions and specifications.  Leaving out the underscores and strikethroughs is a demonstration of lack of competence or subterfuge that can only lead to the outcome being challenged. That is not the sort of attention the board needs to be focussing on the Houston Interpreters and Translators Association.

Any association of actual linguists can raise revenue and provide a public service by hosting interpreting contests. Winners selected by the attendees could thereby earn credible credentials by live testing. A similar competitive approach is used to select and rank athletes, speakers, dancers, writers–even tire-changers or jugglers performing at association events. An interpreting contest need be no more complicated than a live debate or a spelling bee, and its results would carry weight with the membership, judges, attorneys, doctors and honest regulators interested in an objective assessment of competence in performance.

If you are an interpreter or translator interested in the honest defense of individual rights, by all means do get in touch.

 

Certified Portuguese Translators

A statistical breakdown of ATA-certified Portuguese translators. These are translators who have passed a relatively simple test by translating some 700 to 800 words in three passages selected out of a total of five. Three major errors or 20 minor errors suffice to fail a passage, and one has to pass two to pass the test. From 1981 until 2004, passing at least one of these tests was a requirement for voting in the association. People who have passed the tests are usually listed on the ATA website and hence are verifiable.

More and more entities are exercising responsible stewardship by checking translator qualifications. So, what is an ATA-certified Portuguese translator? Of the 987 persons claiming the ability to translate to or from Portuguese with professional competence, only about 16% have passed either of the two separate and distinct tests. Four out of five alleged Portuguese into English translators have never passed that specific test and only 14% of those claiming the ability to translate from English into Portuguese have demonstrated that ability by passing the other specific test. Most of the people who pass either test are native speakers of the target language, meaning that is their dominant language.

If someone says “I am an ATA-certified Portuguese translator,” that doesn’t tell you very much unless they mean they have passed the tests in both directions. Only one percent of all of ATA members claiming some competence as Portuguese translators have passed the certification tests in both directions. That works out to exactly 12 listed translators at the time of writing (2 more are unlisted). Although certification in one direction is better than no certification at all, only one in about 18 certified translators can reliably work in both directions for Portuguese. Here is a breakdown of the numbers:

ATA members who Claim E-P Claim P-E Claim Bidirectional
Claim direction ability 441 546 643 (estimated)
Certified f/direction(s) 86 75 12 (or 14)
Unverified 355 471 359
% unverified 80 86 36
% Certified 20 14 1

Most certified translators only assert that they are certified in one particular direction. The ATA, for reasons of internal politics, goes to great lengths to suggest that certification tests have nothing to do with interpreting ability. I have observed many interpreters, and every one of the certified translators who has passed the test in both directions has turned out to be capable of interpreting with professional competence in both directions (without necessarily liking the work). Somewhere in between 284 and 643 of these interpreters claim competence in both directions. The data tell us nothing about overlap, but the ratios of certified to uncertified (as translators) appear to be in the same ballpark for interpreters as for translators. As you might expect, most (but by no means all) of the better bidirectional interpreters in the ATA have passed at least one of the tests as near as I can ascertain.

Looking at the ATA as a whole, one is struck by the tiny number of people are certified into three different languages. When total membership stood at 8000, there were three such members. It is probably a safe bet that there are approximately half a dozen translators certified into three target languages today.