evought

For what it is worth, I am currently a Premium User and have been either Plus or Premium for some time. The Evernote folks should know that this is STILL an issue. I will talk a little about use cases and maybe some alternatives for making support easier. Many people learning a Semitic language (including Hebrew, Arabic, Aramaic, etc) need to write notes, vocabulary, add glosses (short definitions) to text, or record references (like the gentlemen in this thread with Strong's concordance numbers (LTR) intersperesed with RTL text, Syriac was it?). This is also clearly going to be true of anyone doing original language Bible or Koran study. It is currently true of researchers studying ME terrorism, the Israeli-Palestinian crisis, or other similar issues. Evernote, in some ways, is a perfect tool to organize, index, and search that kind of thing. As it turns out, I do all three of these things and I am building up a good amount of mixed-language notes. Almost all of my paper notebooks on language studies, Bible Studies, and Terrorism research have moved to electronic text, either typed or hand-written with a stylus. Evernote seems to have little problem indexing text and tags in multiple languages (other than some issues searching and matching pointed text--- but we'll talk about that another day ;-) ). Actually ENTERING the text in EverNote, at least on Android (5, 6, and 7) is a *royal* pain in the rear. Much of the time, I end up putting the tablet down in frustration, syncing over to NixNote on Linux, fixing the text, and syncing back. NixNote's RTL support is *not* perfect either, but it is MUCH more reliable than the Android app. Here is some detail on things I try to do (cut and paste from my notebooks), numbered for reference, then I'll talk a little about ways Evernote might either improve OR *avoid* the issues. [1] Language exercises: ---- Translate Hebrew To English: אַתָה הוּא הֳֽאֱלֹהִים וַאֲנַ֫חְנוּ עַמְךָ בְּנֵי־בְרִיתָֽךָ Thou art the God and we your people, sons of your covenant. שָׁלַח מֹשֶׂה מְרַגְלִים אֶת־אֶ֫רֶץ כְּנַ֫עַן וְהָֽאֲנָשִׁים הָיוּ אַנְשֵׁי־שׁם, רְאֹשֵׁי־יִשְׂרָאֵל Moses sent out spies into the land of Canaan, and the men were men-of-name, chiefs among the Israelites. --- (You may see boxes where some of the characters don't display properly in your font: that isn't a problem in the app.) You've got alternating RTL and LTR text on different lines. [2] Vocabulary --- כְּנַ֫עַן - Canaan מְרַגֵּל - spy לֶ֫חֶם - bread --- [3] Mixed language notes This is closely related to the Arabic practice of mubahlaga ("مُبالَغة": the "ah" is an 'a' as in 'father' and the 'g' is an aspirated sound like in "gomorah"), which is more or less a habitual practice of excessive, almost fantastic exaggeration. There are a few reasons why this practice happens, but... [4] Definitions --- אניה: 'onîyâh Ship Men of ships, seamen [BDB ] From H590 Strong's Definition: Feminine of H590; a ship: - ship ([-men]). יונה/Jonah 1:3 & 4 Q: Why is seamen feminine? --- With #2 you've got mixed LTR and RTL text on the same lines, often separated by punctuation. This could be RTL terms with LTR definitions or vice-versa. With #3 and #4, you have mixed free-form text which may suddenly contain text in the reverse definition. Notice that in the definition of oniyah above I have Hebrew, English, phonetics, links, and keywords all tied together. I can type the characters very easily by switching software keyboards and, in theory, have it all neatly indexed in Evernote. I can even attach hand-writing to add Ancient Hebrew pictograms (I haven't figured out how to type those yet!) all neatly packaged. What Goes Wrong: When entering the text, cursor handling becomes totally bizarre. In the best of programs, it can be confusing to move the cursor in and out of the runs of bi-directional text and keep track of where you are (do you move the cursor logically such that the forward arrow takes you to the next logical letter in the sequence, or does the right arrow always move one space right and the left arrow one space left? Evernote doesn't even try. It leaves the cursor in one place and the next character typed goes wherever it feels like. If the cursor is at the end of the line, RTL text is added at the beginning; if I put the cursor at the beginning of the line, text is entered at the end! Using the arrows to move through the text (say, on a bluetooth keyboard) is an exercise in frustration. With either the software keyboard (say Hebrew Nikud Icon-Me or Multi-Ling O Keyboard) or a hardware keyboard, it often forgets where it is and just starts adding new text at the end of the line but a Backspace then deletes in the other direction! This is especially true when pointing (add in vowel marks and accents) in Hebrew or Arabic. The app sometimes loses its mind in this case and starts creating total gobbledy-*****. I have to then select the block of text, delete it, and retype. Sometimes just selecting the most recent RTL character and retyping that seems to let the app figure itself out, but that leads me to the next issue: Selecting bidirectional text is completely nightmarish. When I sweep out a block of text with finger, mouse, arrow-keys, or stylus, it highlights a more or less arbitrary run of text with the selection handles in completely meaningless positions. I then flail at things randomly until it decides to select the right letter. Moving a handle up, down, right, or left changes the selection in some unrelated fashion. This is often where I give up and go to NixNote or type the text in something else and paste it in. The text above with יונה/Jonah in English and Hebrew was one of those spots I fiddled with for a good while in the app before giving up but it only took me a second or two in this web form, including the time to switch keyboard modes. I often have to type a reference like that in both languages so **either will hit on a search.** With text selection broken, not only is it hard to enter text, it is almost impossible to even fix the problems the app causes. For the most part, ONCE I GET THE TEXT ENTERED, it looks fine on Android, the web app, and NixNote (given appropriate choice of fonts), so somewhere the Unicode must be handled correctly by your back end. I suspect your back end doesn't really care, which is good and the way it should work. Other approaches: Doing bidirectional text right is hard. (I am a retired programmer, believe me, I know!) and sometimes broken APIs underneath make it *really* hard. But, there are also some ways to cheat which might make some of these folks (like me) happ[y/ier]. I am happy with cheating if it gets the problem fixed and reduces my frustration level. A. Be able to add equivalent search terms. If I can file somewhere that when I search for Jonah, I want to find references to יונה as well, then I don't have to write both terms as often like I did in #4. If I only need to write one term (and I don't care which, mostly), then I have less mixed text and both of us have less frustration. It's less to go wrong. This also would fix problems with searching pointed text by accident (e.g. מלך and מֶ֫לֶךְ don't currently match in your searches, one with and one without vowel-pointing. Strictly speaking, they should search equivalent, just like King and king, but if I can work around it... I really don't care.) This is also VERY important in researching ME history and events anyway as the same name, say of a historical figure, can be transliterated 8 different ways in 8 different articles. B. Be able to either split LTR and RTR into multiple columns or C. simple tables. When I do vocabulary lists like #2 or translations like #1, it could be split into columns. In fact, when I do it on paper, I use reversed legal paper with a very wide margin on the right. I write RTL in the main body and LTR in the wide margin. With Greek, I use regular legal paper with the wide margin on the left. You could make the same thing easy to do in software and people would have to mix text less often. I have also done them as simple tables in NixNote (Don't seem to be able to in the Android app) which puts the LTR and RTR text in separate boxes: Hebrew/Arabic terms go in one column, English glosses in the other. Again, less mixed text, less problems. Either solution would have side-benefits for non-bidirectional users. Lots of folks take class notes in columns and that's why they make legal pads with those wide margins in the first place. Really the only thing left then is #2, free-form mixed-language text suddenly switching directions. If it comes down to that and I have to type it in another app and paste it in, sure, that would still be frustrating, but MUCH less so. As it is, however, since I bang into this problem all the time, I am seriously thinking of giving OneNote a try. I'd rather leave my data where it is and just have the problem go away.