Many of us will have documents in various languages in their DT database. Those who have, will have to repeat any search in their database for every single language: a search for “car”, for instance, will have to be followed by a search for “auto” and “macchina”, if the database contains many documents in French, in German, in Italian.
Needless to say that it would be very convenient if DT contained an option which made it possible to extend a search for “car” automatically and permanently to words like “auto” and “macchina”. Did anybody ever think about this? Is there anybody who maybe already worked out a strategy in this field?
I’ve been thinking about the same problem, as I have a mostly two-lingual database (Dutch and English). Searching is not a big problem, as long as you are searching for a single word. You can then give the word in the various languages and use the “any word” option.
But it would be very nice if itwas possible to make a list of linked keywords (of even expressions). For example, if I could link “flood” and “overstroming”. That would not only simplify search, but produce also multilingual results for “see also”. It could also be interesting for linking synonyms.
Of course, this ‘simple’ suggestion may be horrendously difficult to program (i’ve no idea), but it would be nice.
Meanwhile, I am interested in experiences with mmulti-linguale databases (I use multi-lingual groups, and “classify” works in most cases surprisingly well).
Thank you for the “simple” suggestion - implementing such a feature is technically quite simple but creating an easy-to-use interface for editing such synonyms/links would need more time. And, of course, suggestions would be welcome (actually I don’t know if there’s a common interest for such a feature and therefore I can’t promise anything at the moment)
And it shoud be kept in mind that this issue is of the utmost importance not just from a multilingual, but also from a monolingual point of view. Just to give some German examples: the concepts of “Architektur” and “Baukunst”, and those of “Medizin” and “Heilkunst” are more or less identical; those of “Archeologie” and “Altertumskunde” maybe not exactly identical, but at least very similar. Anybody searching for hits of “Architektur”, “Medizin” or “Archeologie” will be interested too in finding the hits of “Baukunst”, “Heilkunst” or “Altertumskunde”. It would be great if it became possible to bundle the searching for similar related concepts, in order to extend a search for e.g. “Architektur” automatically to a phonetically totally different but semantically identical entry like “Baukunst”.
like you I am concerned about multilingual aspects of working with DT. Life is hard, but DT does not do as much as it could as soon as one works is two different languages.
But there is a difference with the problem you mention here: When you collect articles about “Medizin”, it is likely that the word “Heilkunde” is part of that document as well, and at least, all of those documents share a lot of terminology. So the DT classifying method and the “See also”-button can provide you with thhe appropriate documents although there are synonyms.
It is different with different languages. When I have Japanese text about “igaku”, it is most unlikely for DT to find synonyms like “Heilkunst” or other terminology which could link a Japanese document about medicin with a German document.
Fazit: Although I agree in your observation about synonyms in one language being the same in principle like a translation of one word into the other language, these synonyms in one language would not affect DT’s classifying engine. In case you search for “Medizin”, just hit the “See also” button afterwards, and you have a nice collection of documents you are looking for.
Btw, I would like to store these search results, and I am looking forward to the “smart groups” in DT 2.0!
I wanted to ask exactly this same question as I’m having the same problem as described by the original poster:
I’m writing my dissertation in German and I’m using mostly english articles and reports as references.
My dream: I’m selecting a text block in my thesis manuscript and the “see also” feature shows me those references which belong to this text block. This would be extremely useful especially for the “state of the art” chapter.
I tried to index all the pdfs and do so, and (not surprisingly) the “see also” feature does only find German references. Not bad, but nor enough for me.
I had the same Idea as Timotheus: a kind of dictionary table which tells DT that “injection mo(u)lding” and “Spritzgießen” is the same thing and that “shrinkage” and “Schwindung” also belong together.
In terms of database theory, this would mean creating kind of a 1:1 relation…
I read about the “sheets” in DT Pro and wondered if such a relation table could be created quite easily with a sheet with different columns.
… Or maybe just by creating a new text entry for each “word field” which contains all translations of the word grouping them together.
As I’m very very late with my thesis I don’t have time to wait for DevonThink 5.x implementing such a feature, so I would like to know if one of the workarounds mentioned above could help to solve this problem?
(as Timotheus asked “my” question already more than 3 years ago, maybe the solution or a good workaround already exists?)
Martin, it is possible to “teach” See Also about synonyms, or correspondences between terms in different languages, by creating “bridge” documents that have those contextual relationships. A nonsense document with a high frequency of the related terms may successfully establish the desired relationship. You could give such bridge documents a recognizable name that you can then ignore when they pop up in a See Also list or in Search results.
I’ve previously discussed how See Also “discovers” that wolves, foxes and dogs are all members of the canine family, so that it can suggest a document about canines or about wolves even though the document being viewed, e.g, one about dogs, may not contain that term.
