Initial vibes

jurgen85

榜眼
(3.99.1.13)

I got the impression that you were more interested in vibes based feedback at this stage, so I’ll start with some things I really like:
  • The tree style components are amazing.
  • I love that flashcards can contain multiple entries, but the UI for customizing it is daunting. I imagine something like checkboxes instead of the plus signs for every sub-entry, or allow multiple checks in Card Info -> Remap to dictionary.
  • Importing 30,000 flashcards only took a couple of seconds.
  • Karmina Sans is a great font, compact, clear, and still has some cute details.
  • The boxes around examples are much clearer than the lines from before.
  • Flashcards being their own dictionary solves so many of my old problems. The card Info/Edit/Dict tabs also make card management a lot smoother.
I have gotten quite used to version 3, but set version 4 up with the new defaults out of curiosity. So while I eventually figured out some of these, and at the risk of yelling about bikesheds, here is a mix of things that confused or annoyed me:
  1. In the search settings, what does “detect patterns” mean?
  2. What does the “separate entries by simplified + traditional” setting mean?
  3. What does the flashcard “is rare” option do? Should I use it for words I don’t care about, or ones that I won’t encounter in the “wild” and therefore want to prioritize higher in tests?
  4. It took me a long while to realize you can long press the tabs at the bottom. A hint about that would be great.
  5. The initial search screen had me confused for a while. Why can I search single dictionaries from that screen, but also from another slightly different screen accessed from long-pressing the search tab? It’s also easy to forget that you are limited to any dictionary you (accidentally) press there, since the title collapses when you start typing.
  6. There’s also the inverse problem: can I start searching for something in all dicts, and then use that same search in a single dictionary without having to copy-paste?
  7. Is there a way to apply tone colors like before, only where the dictionaries themselves provide pinyin? I can get it for headwords, but not for examples, without also coloring everything else.
  8. Can’t delete some “created” settings, like External Lookups or Presentations.
  9. I have a small screen. In definition screens there’s a lot of space, both horizontal and vertical. The individual dictionary titles also take a lot of space, and not that important information to me.
  10. The dict slider is hard to hit.
  11. Can I customize dictionary abbreviation labels, and hide the separator dots?
  12. Can I always hide examples for certain dicts while keeping examples expanded for others?
  13. Can I customize the colors of definition backgrounds, labels, links and pinyin-less examples?
  14. Can I increase the font size of Chinese-Chinese definitions? I tried messing around with the expert settings but can’t figure it out.
  15. Can I customize the reader popup toolbars, beyond the legacy presets?
  16. Why does restoring a flashcard database also reset settings unrelated to flashcards?
  17. Custom tone color preview does not follow dark mode (the “1 2 3 4 5”)
  18. In expert settings for Fields, I can change the font style of “Mandarin” which does affect some parts of entry body, but I can’t find a way to make it apply to lists pinyin, example pinyin, variant lines at the end of entries, measure words in CCC, etc.
  19. Links to other entries only show that single dictionary.
  20. With traditional as the main character set, links containing traditional characters don’t work.
  21. I want to be able to close the “small” card info window by swiping it away or clicking outside the window.
  22. Can’t edge swipe to go “back” in checkbox menus.
  23. It took me a while to figure out how to filter flashcards by custom definitions. “result dictionaries”, “links to dictionaries”, “transclusion” have either confusing or technical names.
  24. Is there a way to reset only expert settings to default? They’re easy to mess up.
  25. Re-ordering flashcards in their categories doesn’t appear to be saved. Maybe related to changing things in the organize and search tabs at the “same” time? Need to go up a category level after making changes?
  26. In imported custom flashcards, two single-quotes appear where one single-quote is expected. (only sometimes)
  27. Can I set up the reader popup to always start from the top dictionary?
  28. Flashcard import automatically converts some links and line breaks but not all. (those hacky PUA characters)
  29. Outlier is very unpredictable in how it combines TC/SC. In a card test with 際〔际〕, only the SC version is grouped with the other dicts in the list, but in a search, only the TC version is.
  30. The search list contains what looks like exact duplicates, e.g. 2 of the same SC entry in a row. Just searching ji4 in traditional mode I find 2 of each of 记、计、继、剂、际、etc. from OSC, all separated from their friend dicts who have both TC/SC. So here are 2 problems: how to handle Outlier that doesn’t specify the other character set, and why are there duplicates. (If it’s relevant, I have set up only 2 dict groups, one C-E and one E-C)
  31. The flashcard Edit tab is annoying to scroll. I suggest reducing the font size to match the rest of the app and/or have toggles to hide fields we don’t use.
 
Last edited:

jurgen85

榜眼
Also I just had an unrelated idea: what if you could swipe with two fingers in a handwriting test to scroll the definition without exiting input mode?
 

mikelove

皇帝
Staff member
Thanks for all of this!
I love that flashcards can contain multiple entries, but the UI for customizing it is daunting. I imagine something like checkboxes instead of the plus signs for every sub-entry, or allow multiple checks in Card Info -> Remap to dictionary.
Yeah, eventually we plan to replace it with little inline previews instead.
In the search settings, what does “detect patterns” mean?
It means that it'll look for grammar patterns in your search, like 不...不...; also does this in the reader.
What does the “separate entries by simplified + traditional” setting mean?
Pretty much that - if there's an entry with a single simplified version but multiple traditional ones, this determines whether we show them as a single entry or split them into different ones. (people studying simplified only probably want this off, people juggling both probably want it on)
What does the flashcard “is rare” option do? Should I use it for words I don’t care about, or ones that I won’t encounter in the “wild” and therefore want to prioritize higher in tests?
At the moment it's actually just about dictionary display - it determines whether we bother showing its variant traditional / pronunciation / etc in the header if it differs from whatever word it merged with. (we use it internally to keep obscure variant characters from cluttering up headers)
The initial search screen had me confused for a while. Why can I search single dictionaries from that screen, but also from another slightly different screen accessed from long-pressing the search tab? It’s also easy to forget that you are limited to any dictionary you (accidentally) press there, since the title collapses when you start typing.
Yeah, still debating what to do with that screen. In general the goal is to a) make it easier to do single-dictionary searches and b) make it easier for new users who've just bought a dictionary to see that they can actually access the thing that they bought.
Is there a way to apply tone colors like before, only where the dictionaries themselves provide pinyin? I can get it for headwords, but not for examples, without also coloring everything else.
No, though we could probably add an option for that. In general our impression is that people who like tone colors like to have them everywhere possible, even when imperfect, so now that we have a decent system for auto-generating them we weren't sure whether we'd ever not want to if tone coloring is turned on.
Can’t delete some “created” settings, like External Lookups or Presentations.
Eek, yeah, that's an ugly one - thanks.
I have a small screen. In definition screens there’s a lot of space, both horizontal and vertical. The individual dictionary titles also take a lot of space, and not that important information to me.
Those are as much a visual separator as a conveyer of information.
The dict slider is hard to hit.
Hmm, have to see if the touch area is different than 3.0.
Can I customize dictionary abbreviation labels, and hide the separator dots?
The separators show up if we can't fit in all of the abbreviations; I'd be reluctant to remove them because then it won't be clear why only some of them are appearing. I suppose we could offer an option to only show however many we can squeeze in and skip the rest but then you'd have to manually scroll down to the bottom after you get past the last visible one and I'm not sure if that's better. Abbreviations aren't currently customizable, though we could revisit that.
Can I always hide examples for certain dicts while keeping examples expanded for others?
Not at the moment but it might make sense.
Can I customize the colors of definition backgrounds, labels, links and pinyin-less examples?
No, though internally it's all Pleco Markdown and so we could probably let people attach different styles to it all at some point.
Can I increase the font size of Chinese-Chinese definitions? I tried messing around with the expert settings but can’t figure it out.
This also goes in the 'dictionary-specific stylistic override' potential feature category.
Can I customize the reader popup toolbars, beyond the legacy presets?
Yes, if you make an Expert Mode reader screen. You can use that in a document via its expert settings, and can use it in the dictionary definition screen by making a custom one of those, switching Chinese Word Card Type to a custom type and picking that as its default definition screen. (this is very involved and we're probably going to streamline it since a TON of people have very specific workflows they want from those toolbars)
Why does restoring a flashcard database also reset settings unrelated to flashcards?
It's all one database now. The new JSON export/import format is the intended way to back up + share flashcards with all of their attendant data but not any other data, but 'backup database' is really 'back up all of your user data.' I suppose when restoring a 3.0 database that lacks settings we could add some code to carry over anything that wasn't in that database but I don't know if enough people will care to justify adding that.
In expert settings for Fields, I can change the font style of “Mandarin” which does affect some parts of entry body, but I can’t find a way to make it apply to lists pinyin, example pinyin, variant lines at the end of entries, measure words in CCC, etc.
Yeah, that's kind of a vestigial option that we should probably clean up or get rid of - it pre-dates the current 'presentation' system. (though the related text processing options do still matter since they're how we efficiently switch from Pinyin to Wade-Giles or whatever)
I want to be able to close the “small” card info window by swiping it away or clicking outside the window.
You mean the duplicate card alert?
Can’t edge swipe to go “back” in checkbox menus.
Sorry, checkbox menus? You mean popup screens like the one where you choose categories? (those are done as modal dialogs with done buttons)
Is there a way to reset only expert settings to default? They’re easy to mess up.
If you switch one of them back to 'builtin' type it'll go back to its defaults, more or less. You can then convert it to custom again.
Can I set up the reader popup to always start from the top dictionary?
Yes, option for it is in the 'dictionary order' subscreen - 'preferred dictionary.'
Flashcard import automatically converts some links and line breaks but not all. (those hacky PUA characters)
Could you send some examples of what it's not converting?
Outlier is very unpredictable in how it combines TC/SC. In a card test with 際〔际〕, only the SC version is grouped with the other dicts in the list, but in a search, only the TC version is.
Yeah, Outlier is kind of our test case for 'dictionaries that only cover one character set' so we're still working out some kinks in that behavior; there's not really a great obvious answer for where to stick entries from a simplified-only dictionary when you're in traditional mode. Since there aren't really any characters that are *only* in Outlier and can't be found in any other dictionary, in multi-dictionary searches we might simply fetch + merge in matching opposite-set Outlier entries for a particular entry, like we do with Unihan.
The flashcard Edit tab is annoying to scroll. I suggest reducing the font size to match the rest of the app and/or have toggles to hide fields we don’t use.
You can customize the editing field order in the relevant Card Type. I'd be reluctant to shrink the font size since it seems like larger fields are helpful when editing.
 
Last edited:

jurgen85

榜眼
The separators show up if we can't fit in all of the abbreviations; I'd be reluctant to remove them because then it won't be clear why only some of them are appearing. I suppose we could offer an option to only show however many we can squeeze in and skip the rest but then you'd have to manually scroll down to the bottom after you get past the last visible one and I'm not sure if that's better. Abbreviations aren't currently customizable, though we could revisit that.
I think it would make sense to have a little “...” label added at the “end”, which would both make it clear that the list is cut off, but also only take that one little space instead of once for every dictionary.

Yes, if you make an Expert Mode reader screen. You can use that in a document via its expert settings, and can use it in the dictionary definition screen by making a custom one of those, switching Chinese Word Card Type to a custom type and picking that as its default definition screen. (this is very involved and we’re probably going to streamline it since a TON of people have very specific workflows they want from those toolbars)
I managed to create a custom reader screen, but I can only figure out how to apply it to the reader tab, i.e. not when looking up in the dictionary tab.

I also can’t include external lookups in the “share selection” menu.

I want to be able to close the “small” card info window by swiping it away or clicking outside the window.
Yes, I mean the Duplicate Card window. I’d actually rather be taken directly to Card Info when pressing the [+] buttons. I’m also not sure about the red color coding of Cancel and/or Delete Card.

Sorry, checkbox menus? You mean popup screens like the one where you choose categories? (those are done as modal dialogs with done buttons)
Yes, e.g. when you use a filter in a custom search, you get a checkbox menu to select dictionaries. Why not make it a normal window? Most other parts of those menus allow the back swipe gesture.

Flashcard import automatically converts some links and line breaks but not all. (those hacky PUA characters)
I couldn’t find any patterns for whether it would convert successfully or not. I used basically the same custom format for most cards, but in some cases one card converted successfully but its almost identical twin didn’t.

Yeah, Outlier is kind of our test case for 'dictionaries that only cover one character set' so we're still working out some kinks in that behavior; there's not really a great obvious answer for where to stick entries from a simplified-only dictionary when you're in traditional mode. Since there aren't really any characters that are *only* in Outlier and can't be found in any other dictionary, in multi-dictionary searches we might simply fetch + merge in matching opposite-set Outlier entries for a particular entry, like we do with Unihan.
I think for the most part the behavior in version 3 was fine for Outlier.

Perhaps related, GR sometimes has the headword in the simplified field (e.g. 云 or 胡), and I might have seen the same in MOE.

Is this related to auto-conversion of entry body as well? In the case of 胡, that body is already traditional but still gets converted, confusing 鬍and 衚.



I also have some new pieces of feedback:

  1. Some headwords have superscript numbers (e.g. OCC 借) and this isn’t stripped when sharing, which means you can’t look it up in Apple dictionaries.
  2. That same superscript determines whether a card is treated as “mirroring” a certain dictionary (whether the flashcard button becomes [+] or [↓]).
  3. The “Share” menus could use a more fitting name like e.g. “More Actions”.
  4. Some combination of
    1. a custom Card Type for Chinese Entry
    2. changing the order of that Card Type’s TC and SC fields
    3. using TC[SC] mode
    4. hide simplified if longer than 5
    5. something else ;)
    caused flashcards to import with empty TC fields.
  5. “Max definition lines” does not belong in the “Adjust font sizes” menu.
  6. Newlines from markdown double trailing spaces do not get collapsed in compact lists (when not followed by a markdown list), e.g.
    Code:
    @noun@  
    clothes
  7. What can I use the Variant Info field for?
  8. I had a hard time finding the component input button since it is hidden behind the handwriting button. Can I e.g. replace the voice input button with component input instead?
  9. What am I missing out on if I use “Legacy search screen interface”?
  10. Selected text has a weird highlighted area, especially for English text.
  11. I can change fonts between the “Basic Fonts”, “System Fonts” and “Pleco Add-On Fonts”, but I can no longer open the “Custom” menu. Clicking it simply resets to PingFang SC and returns to the main settings menu. Some possible culprits:
    1. I had already installed Source Han Sans SC from Creative Cloud. It showed up in the “Pleco Add-On Fonts” section even though I hadn’t installed it from there. I could select it and it worked.
    2. When I could still access the “Custom” fonts menu, I downloaded some fonts from the “More Fonts” button at the top right, notably Lantinghei, which after download only listed the SC version, but then after a restart also the TC version. When I selected that it triggered the error.
    3. I had installed some other fonts via custom hand crafted configuration profiles. I suspect I might have accidentally caused a UUID collision, perhaps corrupting an index or something, but other apps like Pages still accept those fonts.
 

jurgen85

榜眼
Update on the font misadventures: I exported my settings and manually replaced Lantinghei TC with the default. After restoring that, I can pick other fonts again. Lantinghei TC is only available to “use” as an A-Z font, but Lantinghei SC has no issues as a Chinese font. Both trigger the problem when used as an A-Z font. SimSong appears to have the same problem.
 

mikelove

皇帝
Staff member
I think it would make sense to have a little “...” label added at the “end”, which would both make it clear that the list is cut off, but also only take that one little space instead of once for every dictionary.
The problem is that then you wouldn't be able to navigate the whole list; you'd be limited to the first however many dictionaries, tapping on that ... might leave you only halfway down and having to scroll manually for the rest.

I managed to create a custom reader screen, but I can only figure out how to apply it to the reader tab, i.e. not when looking up in the dictionary tab.
For the dictionary tab you'd have to create a custom content screen, then set that as the default for Chinese word flashcards (Card types, Chinese Word, convert that to custom, then pick your new content screen as the default).

I also can’t include external lookups in the “share selection” menu.
Looks like a bug, there should be a menu item configuration screen for that like there is for share entry.

Yes, e.g. when you use a filter in a custom search, you get a checkbox menu to select dictionaries. Why not make it a normal window? Most other parts of those menus allow the back swipe gesture.
If I remember correctly, this was a workaround for a really nasty UIKit bug, but it was probably from iOS 15 or 16 and might be fixable now.

I couldn’t find any patterns for whether it would convert successfully or not. I used basically the same custom format for most cards, but in some cases one card converted successfully but its almost identical twin didn’t.
Could you maybe send me a sample file with a working + a non-working card?

I think for the most part the behavior in version 3 was fine for Outlier.
Yeah, I'm getting the impression from a number of people that the old system was better; I don't think we appreciated that people were actually downloading + using both Outliers rather than just the one for their preferred character set.

What can I use the Variant Info field for?
That's for the little clarifying remarks around variant characters / readings / etc, "OLD VARIANT OF" etc.

What am I missing out on if I use “Legacy search screen interface”?
It doesn't show you the search history menu when you tap the input box, and you don't get to instantly see which other search profiles have results (e.g. for typing in 'ran' and seeing that that has results both for Chinese and English).

Selected text has a weird highlighted area, especially for English text.
Could you be more specific? Both as to how it's weird and to which selected text you're referring to (long press highlights, full text search results, etc)?
 

mikelove

皇帝
Staff member
Some combination of
  1. a custom Card Type for Chinese Entry
  2. changing the order of that Card Type’s TC and SC fields
  3. using TC[SC] mode
  4. hide simplified if longer than 5
  5. something else ;)
caused flashcards to import with empty TC fields.
Regarding this one: is there a chance you were importing a simplified-only card that mapped to multiple traditional versions? When you do that, the new default behavior is that we leave the traditional blank and give you the definitions for all of those traditional versions, since we don't want to drop an important meaning if, say, somebody imports a list of simplified characters that includes 历.
 

jurgen85

榜眼
Regarding this one: is there a chance you were importing a simplified-only card that mapped to multiple traditional versions? When you do that, the new default behavior is that we leave the traditional blank and give you the definitions for all of those traditional versions, since we don't want to drop an important meaning if, say, somebody imports a list of simplified characters that includes 历.
I’m usually meticulous about including both SC and TC in my cards, but I guess there is always a possibility that I messed up the exported list somehow?
 

jurgen85

榜眼
Anyway, here’s a fresh new batch of feedback:
  1. The Search Screen setting button “toggle character set” forces a single set and hides the other. (no 〔…〕)
  2. Customize separators between SC/TC? (〔〕) Show on separate rows without separators (if long enough)?
  3. I think splitting entries by neutral tone makes sense for single-syllable words, but the engine is a little aggressive for multi-syllable words (eg 困难 or the many …上来, …下去 etc). Yes, there are some words that have quite different meanings (e.g. 告诉), but usually they are related and I think it’s useful to group them, if only to avoid the (easily forgotten) chore of having to search twice for both readings. (I think I could make the same case for erhua pairs.)
  4. The expert settings are mostly self contained, but it’s very confusing to figure out where they should all finally be hooked in to replace the normal settings, e.g. reader screens need separate set up for clipboard, documents, search screen, etc. I took a long while to figure out where to override the content screen. (“Card Types” is not where I would look. I think it’s better to rename it Entry Types since it isn’t limited to cards.)
  5. There are a bunch of on-off settings phrased as negatives (“Don’t …”, “Exclude …” etc.) that would be clearer visually as positives.
  6. The dict slider moved to the side of the screen is better, but it currently makes it impossible to click things that are under it. It could be moved all the way out to the edge of the screen, like Apple did e.g. in lists of contacts. I might prefer if it was hidden until I started scrolling.
  7. KEY has a bunch of unformatted labels like {physiology}, {physics}, {chem}, {music}, {onom}, etc.
  8. A Custom Presentation for lists does not respect “maximum # of lines” for certain values of “compact entry structure” and “Field layout” (in overall settings section). Lists also behave strangely with “show definition text only”
  9. Setting up head+reading+defn on the same line will scroll past the headwords for long pinyin or full text searches.
  10. Many (or even all) text input fields could benefit from using a monospace font, e.g. regex fields.
  11. In a saved search, setting up “field text” “entry body” “contains“ “regex” without “multi-line match” only treats the very end of the field as a proper $
  12. When I change the expert settings for a field, does it need to process every single card that has that field in order to save the change? It takes a long time.
  13. I had some trouble after restoring settings and/or cards, and somehow it made Unihan appear as a normal dict.
  14. Rearranging dictionaries by dragging and dropping has a tendency to get them stuck, you will appear to keep dragging them, but they will end up somewhere else.
  15. Drag and dropping card categories triggering scrolling can get stuck scrolling upwards after supposedly ending the drag and drop.
  16. The Presentation type Simple seems more advanced than Custom.
  17. A lot of editing expert setting is doing the same change to a number of sibling fields. E.g. sorting entries with flashcards first in the words and chars tabs requires 10+ of the same change. Regardless if that becomes a non-expert setting. It would be really cool if you could set up defaults, more fields getting an inherit option, “set/add field/list to parent/children/siblings” etc.
  18. Not contradicting the above, some big things like reader screens should have a way to “set this reader screen as the default everywhere”, or maybe being able to edit the default reader screen directly.
  19. In presentations, to specify a “default” language and translit system for tone colors? (use whatever is set up in layman settings rather than forcing a system)
  20. “Font size scale” for custom font sets had no effect, until I changed the system Dynamic Type font scale.
  21. Custom insert text buttons for field editing styles can’t contain newlines?
  22. When changing the name of a card category, dragging the text anchor and lifting my finger while it is above the Cancel button will actually press that button, rather than just moving the text anchor horizontally.
  23. The custom reader popup button “next dictionary” has a weird signpost-like icon where I expect the current dictionary label. I also can’t long press it to open the list.
  24. The reader popup when selecting 转 will default to KEY’s weird 转〔転〕 (same with 台 that picks 台〔㙵〕, or OTC’s 號) despite the normal characters being available in dictionaries with higher priority.
  25. I find bold fonts very hard to read for Chinese characters. Perhaps a system wide setting to switch to background highlighting instead? Or even change the default
  26. Unrelated to the above, there is a lot of redundant highlighting, e.g. Chinese characters in E-C defs. (They already stand out from the rest of the definition by being Chinese characters.)
  27. After messing with card type and field expert settings, and going back and forth between the settings tab and the flashcards tab, the main settings menu became completely black. I was able to long press the settings tab to access addons, registration etc., but not the main screen. The app did not crash. It returned to normal after an app restart.
  28. I miss the old “shrink selection” behavior when only one character was selected.
  29. Importing a .plecoflash containing only flashcards will scramble my custom order in Manage Dictionaries. I did a settings backup before and after import and diff’ed them and that’s seemingly the only change.
  30. In operation preview lists, add prev/next card buttons.
  31. Sort the “internal” file pickers by date modified?
  32. If I set up Default Categories in global settings, I don’t want long-press to change that whenever I add a card to another category. The current behavior should be a separate option like “remember previous category”.
  33. At least in the reader tab, using a font set up in expert settings with a custom scaling, does not fall back, missing glyphs are just ignored.
  34. Presentation definitions should have a “Chinese character set” > “Keep original” option.
  35. A custom presentation for examples with “don’t derive missing” results in completely blank Chinese, rather than falling back to the other character set (e.g. NEC examples in TC mode).
  36. The Outlier “system” buttons () should use the link color.
  37. Multi-syllable searches are fine, but simple searches like “yi4” take a good second or two, while in v3 it’s basically instant. If I’m searching for a single syllable, it’s in 99% of cases because I’m interested in a single syllable. How about loading those first and then the rest? (I guess this might only work well with my custom sorting imitating v3.)
  38. The anchors for selected text (the blue handles) are hard to hit.
  39. Can’t toggle compact mode from the Organize tab’s dict tab.
  40. The variant fields that appear when one dictionary has a unique value causes all other dicts to also repeat the value from the top dict. See e.g. 严惩不贷 and 档案
  41. Some dictionaries like GRD spell out sandhi, but it’s treated like normal.
  42. Can I filter cards by part of speech, including transcluded definitions?
 
Last edited:

mikelove

皇帝
Staff member
Thanks for all this meticulous feedback!
I think splitting entries by neutral tone makes sense for single-syllable words, but the engine is a little aggressive for multi-syllable words (eg 困难 or the many …上来, …下去 etc). Yes, there are some words that have quite different meanings (e.g. 告诉), but usually they are related and I think it’s useful to group them, if only to avoid the (easily forgotten) chore of having to search twice for both readings. (I think I could make the same case for erhua pairs.)
Sorry, you mean that if you do a search for 困难 you see two separate results for kun4nan5 and kun4nan2? Could you possibly send me a screenshot of that and a backup of your settings? (there's a bug relating to this sort of over-aggressive entry splitting that we're currently working on)
The dict slider moved to the side of the screen is better, but it currently makes it impossible to click things that are under it.
The margins were supposed to be big enough with it enabled that that wouldn't be a problem - could you send me a sample screenshot of that too?
When I change the expert settings for a field, does it need to process every single card that has that field in order to save the change? It takes a long time.
What sort of change were you making? It should only be triggering on a few specific cases; at the moment those are:
  1. change to derived source list (since it has to re-derive them)
  2. change to search index (has to remove from old / add to new)
  3. change to default content field (since that requires rebuilding the indexes with items in the new field indexed as such)
  4. change to alternate index field or processor (rebuilding that index)
  5. change to class ID (since that's used to tag text in a content field with it)
  6. change to any of the merge transformers (since it has to reindex with the new one)
For anything other than those, let me know what you were changing and we can investigate.
The Presentation type Simple seems more advanced than Custom.
There was a bug where a lot of the options in custom were missing but it should be fixed now; however, custom is kind of 'deceptively simple' because all it's really doing is specifying what text to fetch, when to fetch it and how to format it; the complicated part is figuring out all of those rules. If you convert a simple presentation to custom you can see how this works in practice - we do a lot of nesting of presentations inside of other presentations, among other things.
A custom presentation for examples with “don’t derive missing” results in completely blank Chinese, rather than falling back to the other character set (e.g. NEC examples in TC mode).
That shouldn't be happening with NEC - it has TC versions of them too. Could you do a json export of this custom presentation and email me that too?
Multi-syllable searches are fine, but simple searches like “yi4” take a good second or two, while in v3 it’s basically instant. If I’m searching for a single syllable, it’s in 99% of cases because I’m interested in a single syllable. How about loading those first and then the rest? (I guess this might only work well with my custom sorting imitating v3.)
They shouldn't be nearly that slow; what model of iPhone is this? Could you give me a rough estimate of how many flashcards + user dictionary entries you have? If there's a configuration where those searches are *not* instant we can certainly add an option to limit 1-syllable reading searches to 1-syllable results as we do in 3.0; our testing experience was that the 4.0 search engine was fast enough that that was not necessary.
Can I filter cards by part of speech, including transcluded definitions?
Filter where? We do index parts of speech for searches now, and we alias part of speech tags in C-C dictionaries to their English equivalents, so if you type in "亦 noun" for example you should get the 规范词典 entry for 亦 with the 名 highlighted. A quick check in Organize suggests that this works OK for transcoded definitions too but if there's a situation where it's not working for you let me know.

I'll follow up on any other issues once we have a chance to check them all, thanks again!
 

jurgen85

榜眼
(I sent you my settings in a PM.)

Aggressive entry splitting – I checked again and now 困难 is merged for me too, but there are still others that aren’t, e.g. 反正、事情、运动.

Dict slider – Upon closer inspection, it’s mostly fine, I actually had another issue here. My custom content screen defaulted to “ICON:share” as the abbreviation for Commands, which I have now found and changed too. I didn’t look too close and assumed the slider was right-aligned but the length of that abbreviation confused me. Still, if someone uses long abbreviations it would be best to right-align the slider, and then may still be an overlap issue.

Field expert settings – it was probably one of those ones then.

Slow searches – iPhone 8 with 27000 cards and no other custom dicts.

Filter cards by part of speech – I meant filters used in Custom Searches. The regular searches work fine.
 

mikelove

皇帝
Staff member
Aggressive entry splitting – I checked again and now 困难 is merged for me too, but there are still others that aren’t, e.g. 反正、事情、运动.
Thanks.
Dict slider – Upon closer inspection, it’s mostly fine, I actually had another issue here. My custom content screen defaulted to “ICON:share” as the abbreviation for Commands, which I have now found and changed too. I didn’t look too close and assumed the slider was right-aligned but the length of that abbreviation confused me. Still, if someone uses long abbreviations it would be best to right-align the slider, and then may still be an overlap issue.
Ah, yes, seems like you converted it to custom before we made the sidebar switch and got rid of that no-longer-necesary ICON: bit.
iPhone 8 with 27000 cards and no other custom dicts.
Thanks - our current baseline test system is an XR but that's only about 20% faster than the 8 so the difference shouldn't be so dramatic.
Filter cards by part of speech – I meant filters used in Custom Searches. The regular searches work fine.
No good way to do that at the moment, though we could look at adding an explicit filter for it.

If you convert a search profile to custom, you could theoretically add a new search type that just searched 'part of speech' and no other field, add it to the search profile, and (with 'span multiple blocks' turned off) turn on 'require filter prefix' and give it a prefix like 'ps'; then, when typing a search you'd simply add a "-psnoun" and it should (assuming all of this is hooked up correctly) limit the search to entries with noun in their part of speech field.
 
Top