DT Pro with different languages


I’m completely new to the program and eager to start my own database. I’m a bit confused, however, about the use of Devonthink with entries in different languages (mostly German and English in about the same frequency). I found a post which recommends to start the database with one language, do manual classification and then gradually add other languages. At least that’s what I understood. This would be really hard for me, since all my material is in both languages. I’m sure most of the users collect information in more than one language. My problem might arise from a poor understanding of the program. For the database there might be no such thing as “German” or “English”, only strings of text. But for a human it’s hard to understand that you might be able to use the program’s semantic analysing features in different languages. Can you give any recommendation from your experience? Thanks a lot!


Hi Birgitt,

good point. I live in at least three languages and collect about the same topics mainly in Japanese, English and German.

Yet I cannot cope with the speed my collection is growing, so I just drag stuff into folders not knowing about how DT would classify. Recently I started using the Classify button from time to time, it works astonishingly, but of course, “See also” never shows related Japanese documents to an English document.

I am afraid we have to live with these insufficiencies unless indices are included which show which words have to be treated as identical in different languages. But as far as I understand the developers, they do not plan to do so.

With best wishes to all those who do not accept a monolingual desert,

Hi, Maria:

I’m envious of those who have multiple language skills.

As for me, I’m language-poor, pretty much limited to English and so forced to wander in my one-language desert. :slight_smile:

Hi Bill,

no reason for being envious. Sometimes I feel I cannot think properly any more because alles turns around in my atama, and I cannot speak any language properly. And please don’t take the desert too serious, I express myself nicer in – other – languages :wink:

Best, Maria


Do you ever suffer from mojibake (can’t think of an appropriate English expression) with Japanese texts? It has happened several times to me and rendered the texts seemingly unusable (i.e. I haven’t yet found a way to convert them back to any human-readable form, though I’ve tried SubEthaEdit in all its Japanese encodings).



yes, although it has become less in the last years thanks to Unicode-aware applications. I would not talk about the mojibake of people who use Eudora (though I loved it) on a Windows 95 machine. But there are other problems, and some of them concern DevonThink:

HTML pages are not nicely rendered in many cases, because they do not show the encoding, as they should. Many are EUC or JIS, and to watch them I have to go to the preferences (!) to change the encoding for this page and later back for other pages. I asked for a menu command “encoding” that is accessible on the fly. Never got it.

Then there are pages which have all Japanese characters encoded as entities, they show up nicely, but are not searchable. I would like a command like in BBEdit, where I can re-encode these to unicode.

I cannot remember other cases since a year or so. My work flow is now NOT to use DA or DT for web browsing, but to capture all important information as RTF from within my favourite browser.

So, there is only a segment of what DT and DA offer that we can use.

If you have more of these problems not related to DT, just send me a pm.


Thanks, Maria. Most of my problems are specific to DevonThink, though I occasionally have problems even using Mail when corresponding with people on Windows 95/98 systems. My impression is that Apple have become a tiny bit cavalier about maintaining compatibility, as these problems first surfaced in Panther, were solved with a Terminal command that I forget, and then resurfaced in Tiger. When mailing correspondents with whom I’ve had problems in the past I create a PDF of my Japanese message and enclose it in a short English email.

The “encoding” menu command you mention sounds like a great idea.


I forgot to mention a crucial fact:

Most of my problems with “mojibake” (garbling of Japanese characters) have happened with RTF documents created in TextEdit, or even created within DevonThink. A typical pattern, as far as I can recall, is that I would create a document such as exam questions in DT or drag it into DT, edit it on and off for the next few days without any problems, and then forget about it. On coming back to the document a year later (to make sure not to make the same questions again, or to reuse the rubric) I’d open up the document and it would be unreadable. The only other application on my computer that ever has similar problems is Microsoft Word (Japanese version) that sometimes garbles its own menus.


This thread raises a topic which seems to me of absolutely vital importance; and I don’t understand why DT’s helmsmen are so reluctant to fully admit it, and (especially) to do something about it.

My own DTP database contains documents in at least four languages: Dutch, Italian, Latin and English. At the present state of things, searching and studying the occurrencies in my database of, for instance, that known large animal which in Dutch is called “paard”, in Italian “cavallo”, in Latin “equus” and in English “horse”, is a very time consuming and rather frustrating experience. The find-feature in DTP, indeed, is still rather elementary; there is no keyword feature; and it is not possible to save the search results in an easy way.

It might be interesting to know how other ‘mutilingual’ DTP users cope with similar issues.

Thanks for all your comments! As I understand it, some users store keywords in the comments field. I agree, keywords are really important. On the other hand, they always reflect your point of view at a given time. And keywords can become either a rather rigid system or you are lost in their multitude. Only a handful of keywords bring up to many search results and with to many of them you are completely lost when you want to assign them in a consistent way. You have to put a lot of effort in the maintanace of a really good keyword system. You need features like changing keywords globally, merging keywords or differentiating them. What I like about the idea of Devonthink is that it is a much more flexible database compared to the ones based on a keyword system. But I agree: Keywords are useful to connect entries made in different languages. But in Devonthink you can also do this by using groups, don’t you?

I’ve just stumbled upon another thread that is relevant to this one:
devon-technologies.com/phpBB … php?t=2671

Correspondance between keywords in different languages might prove fairly difficult. No one word is really translatable into another language anyway. :slight_smile:

What I would like to see is classify work with other languages like Chinese. Since there’s often no spaces between words, maybe DEVONthink can work by character or groups of them. It doesn’t have to understand the content to be able to classify them.