The Interpreter’s Toolkit: Intragloss - a useful glossary-building tool for interpreters
Intragloss particularly excels for its glossary-building tools, which make preparing for a conference fast and easy. Its document annotation feature is innovative, but somewhat buggy.
- Mis à jour
To do so, Intragloss offers a suite of tools for reading and comparing documents, identifying terms, searching for equivalents online, and creating term lists and multilingual glossaries. Innovative features allow you to quickly generate assignment glossaries from preparation materials and automatically annotate a document with translations of terms from your glossaries.
Nevertheless, Intragloss is a versatile, easy-to-use program that significantly decreases my preparation time. Despite a few minor bugs, it’s my go-to tool when preparing for nearly every interpreting assignment.
The nuts and bolts
Intragloss is available for Mac computers; a Windows version is in the works. Intragloss supports over 180 languages and is Unicode-compatible, which means it can handle diacritics, special characters, and non-ASCII languages, like Russian and Hebrew. The software includes a user’s guide and fast customer support via email.
Intragloss is one of the most expensive pieces of interpreting software I use; at $49 per month, $99 for three months or $269 per year, an Intragloss subscription is rather pricey. Rates drop to more manageable levels for renewals, especially if you’re willing to shell out $359 for a three-year subscription. Before you take the plunge, try out the free month-long trial. Intragloss also offers a hefty discount for students enrolled in interpreting programs.
Kenig and Pohoryles argue that “Intragloss is a professional glossary-maker that gives you everything you need to prepare for assignments.” While I find this claim to be slightly hyperbolic – I’ll discuss some shortcomings over the course of this article – Intragloss is indeed a robust, time saving tool.
Kenig’s short video provides a quick visual introduction to Intragloss.
A quick overview
Intragloss is organized hierarchically, into domains and assignments. Domains generally reflect specializations, and assignments usually represent a single interpreting job. In this example from the user’s manual, “Cell Phones,” “Cars,” and “Psychology” are domains, while “Paris Expo 2015” and “Madrid Expo 2015” are examples of individual assignments.
The Intragloss interface
Every Intragloss assignment is by default bilingual, and includes a bilingual assignment glossary. Each assignment glossary has at least two columns, for the source and target language terms. You can customize your assignment glossary to include columns with acronyms and remarks in both the source and target languages.
Each domain has a domain glossary, which is automatically populated from the individual assignment glossaries in the domain. Since a domain can include assignments with different language pairs, a domain glossary can be multilingual.
In the example above, “Paris Expo 2015” and “Madrid Expo 2015” are two bilingual assignments – an English/French assignment and an English/Spanish assignment. Both are part of the “Cell Phones” domain, which has a trilingual English/Spanish/French glossary including all of the terms from every assignment in the domain.
Every Intragloss assignment can include one or more documents, such as the “Mobile phone.pdf” file in the example above. I usually import everything I receive from conference organizers and any other materials I find while preparing for an assignment into Intragloss, and use these documents to develop term lists and multilingual glossaries. In theory, Intragloss supports several input formats, including PDF, Word, PowerPoint, Pages and Keynote. In my experience, however, the software often struggles with formats other than PDF, so I tend to convert all my documents to PDF format before importing them.
Intragloss has a built-in web browser, which means you can import one or more webpages into a given assignment. This is an especially useful feature, since interpreters often prepare from online sources in addition to the documents provided by conference organizers.
It took me some time to wrap my head around the idea of domains and assignments, and I find it especially confusing when preparing for a multilingual interpreting job.
I’ve also found that Intragloss creates duplicate entries in the domain glossary when the same source language term appears in two assignments. For example, if I input “attachments” --> “pièces jointes” in the English/French “Paris Expo 2015” assignment and “attachments” --> “documentos adjuntos” in the English/Spanish “Madrid Expo 2015” assignment, instead of merging them into a single trilingual entry in the domain glossary, Intragloss creates two separate bilingual entries in the domain glossary. If you work with multilingual assignments often, like I do, this will probably get on your nerves. My workaround is to create a bilingual glossary with Intragloss, download it, add the third language manually, and upload my glossary to another glossary manager for easy reference while on the job.
Building glossaries using the online search function
Building glossaries with Intragloss is quick and easy. Create a new assignment (and a new domain if needed), open any related documents or websites in that assignment, and start reading through your documents and fleshing out your glossary.
Intragloss offers two different tools for building glossaries: online terminology search and terminology extraction from parallel documents.
The screenshot below, from the Intragloss demo video, demonstrates how the online terminology search function works.
Searching for a term
Highlight an unfamiliar term – “bi-metallic” in this example – and click on the “search” button. Intragloss adds the term to your assignment glossary, highlights it in the document, and searches for it using your pre-selected websites – which can include online dictionaries like the Collins multilingual dictionaries or WordReference.com; parallel corpora tools like Linguee; and terminology portals like IATE, Termium or ProZ.
Once you find a good translation, highlight it, click “Translation,” and the translation is automatically added to your bilingual glossary.
Look through the various search sites (displayed as orange tabs above), highlight the translation, and click to add it to your glossary
You can also highlight acronyms or remarks in the source or target language and add them to your glossary, or manually edit cells to type in translations or remarks by hand.
Overall, Intragloss’ online search feature is a real time-saver. It simplifies the process of searching for, copying and pasting terms and switching between programs, which can be especially time-intensive if you have to search for a term on various terminology sites.
However, the online search tool is not without its flaws. Intragloss struggles with diacritics, multi-word terms, and plural forms. Although IATE pulls up a good translation for a term like “informes periódicos,” the two-word term gets converted into “informesperiodicos” on Collins and the accent gets mangled on its way to Linguee, which normally provides a good translation for this phrase. On ProZ, the source and target languages are not shared with the website, so you see translations in a variety of languages not relevant for your current project. If the way Intragloss handled multi-word terms, diacritics, and plural forms were improved, the program would end up saving you even more time.
Comparing documents and building glossaries from parallel documents
Since I do a lot of work for international organizations, I often receive sets of parallel documents in various languages.
With Intragloss, you can build a glossary from parallel documents in no time flat.
After you attach a translated file to a document you’ve imported, Intragloss displays the original and translated files side by side.
The process for adding a term is similar to the one described above. Highlight a given term and click “entry.” Then highlight the translation in the parallel document and click “translation,” and voilà – you’ve created a glossary entry without extensive copying and pasting or jumping back and forth between programs.
Comparing documents and adding a translation from a parallel document
Intragloss offers handy keyboard shortcuts, making the process of searching for a term and adding entries, translations, acronyms and remarks even faster. It is also cleverly programmed to eliminate extraneous punctuation when you enter a new term (although it would be even better if it could remove articles, e.g. the l’ in French, and if didn’t occasionally remove spaces between words).
Intragloss’ synchronize scrolling feature makes it easy to scroll up and down in both documents simultaneously. This feature is useful when preparing glossaries and even more helpful when a speaker quotes from a text and you can pull up the original and translation side by side.
Comparing documents is probably my favorite part of Intragloss, and the feature I use most frequently. It works nearly flawlessly, although I have occasionally found a slight glitch in the scrolling feature. Of course, comparing documents is even better with more screen real estate, so I especially love using this feature to prepare on my desktop at home. But when I’m on the road or in the booth, it can still be a godsend.
Creating a monolingual terminology list
If you’re reading through a document you just received, but don’t have much time available, you can also use Intragloss to create a monolingual terminology list. Simply highlight a term, click “entry,” and your term is added to your monolingual vocabulary list. You can download and print out this list, which is useful if you want to look up these terms at a later time or discuss them with a colleague.
You can also use the “merge” feature to create documents where these terms are highlighted, providing context if you’d like to ask a colleague or the speaker about difficult terms.
Merging a monolingual terminology list produces a document with terms highlighted
You can download assignment and domain glossaries from Intragloss with just a few clicks.
The program exports glossaries in .txt and .doc formats, which the developers claim are “spreadsheet-” and “word-processing compatible.”
On this front, I think Intragloss falls a bit short.
First, Intragloss exports domain glossaries with the entry, acronym and remark columns, even if you created the domain without these columns. The ability to pick and choose which columns to export would yield more visually appealing results.
Second, although spreadsheet software can open .txt files, this format isn’t generally used in third party applications. I would love to download my glossaries in .xls, .xlsx, or even .csv format – which are compatible with other glossary management tools, like Interpreters’ Help – instead of having to convert the format before I can open my glossaries in another application.
Intragloss also allows you to import pre-existing glossaries. Since these are imported as domain glossaries, you can import multilingual lists, too.
Intragloss can import Word (.doc or .docx, although .docx is preferred), Excel (.xlsx) and text (.txt) files. Column headers can be language names, two-letter ISO 639-1 language codes, acronyms (e.g. “English Acronym”) or remarks (e.g. “English Remark”).
I tested out the import glossary feature using a 500-word and a 1000-word glossary I had created for previous assignments. Intragloss imported the 500-term glossary in 26 seconds and the 1000-term glossary in 52 seconds, for an average import speed of 0.05 seconds / term in both cases. Regardless of whether I labeled column headers using language names or ISO codes, Intragloss correctly read and imported these headers.
Viewing, sorting, and searching through glossaries
You can view domain glossaries in flat and tabular layouts. Examples of both (from the Intragloss user’s guide) are reproduced below.
Viewing a domain glossary in flat mode
In flat mode, click on any given term to see its translation in all the languages in your glossary. (The Spanish and French translations of “attachment” are displayed in the example above.) You can also search for a full or partial entry, acronym or remark – although your search is limited to the language displayed on the screen (in this case, English).
Viewing a domain glossary in tabular mode
Tabular mode displays all the terms, acronyms and remarks in your glossary. You can also search all the entries, acronyms and remarks in every language in your glossary. In theory, you can also sort your lists alphabetically by a given language; in practice, I’ve found this feature to be buggy, since the terms in one language sometimes temporarily disappear. I would also appreciate the option to hide the “acronyms” and “remarks” columns while in tabular mode.
Finally, you can search an assignment glossary by order of appearance or synchronize your document(s) to see where the term first appears. This is handy for seeing a term in context, but the search feature is slow, and alignment in parallel documents is often off.
Synchronized search with parallel documents in an assignment glossary
Although synchronized search might be useful at times, I generally find that Intragloss’ search features aren’t up to snuff. It tends to be slightly slow, which is the last thing you want when you’re trying to come up with a term quickly. Intragloss doesn’t ignore diacritics when searching, slowing you down even more, since you have to type out every accent and umlaut. Columns cannot be reordered, and you can’t easily search across multiple domains at once. Intragloss’ competitors offer these features, which is why I use Intragloss for preparing glossaries, but other tools for searching through them while on assignment.
Quickly preparing an assignment glossary
Once you have imported one of your pre-existing glossaries, you can take advantage of two innovative features: quickly preparing an assignment glossary and merging translations of terms into documents you have received to create an annotated PDF for sight translation or simultaneous with text.
By default, when you import one or more documents into Intragloss, the program automatically highlights all the terms that are already in your domain glossary and creates an assignment glossary including all these terms and their translations. Click on a term to see it in context, scroll through and search this glossary, see a glossary for a single document, or extract this glossary – an incredibly useful tool for quickly deriving a small, assignment-specific glossary from a much larger one.
Automatically extracting terms from a document to create a searchable assignment glossary
This feature is relatively fast, too. I tested this out with my 500- and 1000-term glossaries on a series of different real-world documents. Intragloss prepared these assignment glossaries in 16 and 29 seconds, respectively (about 0.03 seconds/term) – regardless of the length of the document!
The screenshot above is from a 73-page background paper. Intragloss processed the document and extracted a 334-term assignment glossary from it in fewer than 30 seconds – impressive indeed!
Merging documents with glossaries for simultaneous with text or sight translation
Imagine being sent a speech that will be read out during a meeting, but with translations for all of the key terms provided. That’s what Intragloss’ merge feature aims to do – annotate a document to include the translations of every term in your glossary.
Here’s a screenshot from the Intragloss demo video:
Merging documents with glossaries
To put Intragloss through its paces, I took my 500- and 1000-word glossaries and tested out the merge feature with two 1000-word and two 2500-word speeches from a recent interpreting assignment. After importing the speeches, Intragloss created assignment glossaries with between 45 and 95 terms. I then hit the “merge” button, and … the results were mixed.
Intragloss took an average of 0.75 seconds per term (and 0.25 seconds per unique term) to merge these documents, which meant I waited between 8 and 106 seconds per merged document – on top of the 16 – 29 seconds to import the document, and a lot of other assorted clicking. All in all, Intragloss took anywhere between 90 and 180 seconds to spit out these annotated documents.
And it wasn’t particularly accurate, either. Although Intragloss identified hundreds of unique terms, it only came up with their translations 32% of the time.
The merge feature identified tons of terms, but often didn’t propose their translations
As you can see from the images above, this feature can be exceptionally useful when it pulls up a translation that might not roll off the tip of your tongue. But Intragloss struggled with multi-word terms (e.g. officers vs. prison officers), polysemy (e.g. report as a verb vs. report as a noun), singular vs. plural forms (worker vs. workers), and long phrases. It was inconsistent, coming up with the correct translation at times, but omitting it at others. It produced a color-coded text that was difficult to follow, and took its time doing so. It also occasionally printed the English term below the original and the translated equivalent above it.
I think that the merge feature has great potential, but would encourage the developers to fine-tune the algorithm so this feature is more reliable. For the moment, it’s still somewhat helpful if you have the time to run the speech through the program, but faster, more accurate processing would make the merge feature an incredible boon in the booth.
Intragloss sets its sights high, and is packed with tons of useful features. Given the breadth of these features, and the strong implementation of many of them, it comes as no surprise that practitioners and researchers alike consider Intragloss one of the leading glossary tools for interpreters (Drechsel 2016; Rütten 2015; Costa, Corpas and Durán 2014).
Personally, I think Intragloss’ strengths lie in the suite of features for reading and comparing documents, identifying terms, searching for equivalents online, and creating term lists and multilingual glossaries. These work well, are well designed, and cut down my preparation time significantly.
However, the glossary search features in Intragloss are weaker and buggier than those in other software on the market, and implementation is inconsistent for some of Intragloss’ most innovative features: the assignment glossary tool is killer, while the merge and annotate function is ripe with potential but too slow and spotty to use on a daily basis.
Despite these minor glitches, I’d strongly recommend Intragloss. It does what it claims to do: cut down preparation time significantly, and make it easier to prepare glossaries for an assignment. And what busy interpreter doesn’t want to prepare faster, smarter and better?
Costa, H., Corpas Pastor, G., & Durán Muñoz, I. (2014, August). A comparative user evaluation of terminology management tools for interpreters. Paper presented at the 4th International Workshop on Computational Terminology (CompuTerm ’14), Dublin, Ireland.
Drechsel, A. (2016, January 29). Dan Kenig and Intragloss [Audio podcast]. Retrieved from https://www.adrechsel.de/langfm/dan-kenig-intragloss.
Rütten, A. (2015, June 30). “Booth-friendly terminology management: Intragloss – the missing link between texts and glossaries | die Brücke zwischen Text und Glossar” [Blog post]. Retrieved from http://blog.sprachmanagement.net/?p=655.
Josh Goldsmith is a UN- and EU-accredited interpreter and AIIC pre-candidate working from Spanish, French, Italian and Catalan into English. A lover of all things tech, he frequently shares his experience using technology for interpreting at conferences and courses, as well as on Twitter as @Goldsmith_Josh.
Le présent article n'engage que les opinions de l'auteur et ne reflète pas nécessairement le point de vue de l'AIIC.