Thank you.
If I try to load the word ććæ on this āflfl-cliā page it seems to try to load every katakana word in existence and the page takes forever to load and eventually I have to stop the execution. The screenshot is only a partial a image because to capture the whole area was too large to upload.
edit: The book was shirokuma cafe.
2 Likes
What in Godās name
Edit: nevermind, it has nothing to do with ććæ and everything to do with the card afterward, which is blank.
Anyway, itās fixed but Iāll need to find something to prevent that from ever happening again.
1 Like
You work fast! Thanks as always.
1 Like
Update 9/29 - Extremely boring update
Iāve updated the libraries that the website runs on. This is extremely uninteresting and means nothing to every single user unless it breaks the website.
Iām prepping to try and upgrade to Vue 3 (from 2). I might not if it seems too annoying (it will definitely be a process).
After that, Iām going to work on an Anki flip card mode.
5 Likes
Just a general life update. Iām currently working a full-time and a part-time job and it kind of sucks. I am not a āsigma grindsetā kind of guy. Iāll be doing only content updates for a bit.
At the end of December Iām planning on either quitting my full-time job or moving it to part-time so that I can āgo back to school.ā By which I mean starting a metric ton of programming courses and projects. So at that point Iāll probably do some crazy shit on this website under the guise of ālearning data analyticsā or something.
Itāll probably be on the scale of that time I randomly decided to recreate the entire website in Node in a month.
Iāve noticed that my website has, in ways, fallen behind the curve of stuff people have innovated in the four years since Floflo launched so Iāll be reflecting on whatās required to bring it up to par. If anyone has any ideas that they want to throw out so they can marinate for the next two months, go for it.
If anyone has any sites that they think are superior in X way or have potential, feel free to share. Itāll help me figure out specifically where I need to expand my skills in order to match.
10 Likes
10/10 Content Update
New anime (8)
- 86
- Arslan
- Beastars 2
- Fruits Basket (all)
- Nisekoi
- Tensei shitara Slime Datta-ken 1 + 2
- Yuukoku no Moriarty
Iāve also figured out how to make anime compatible with the readability formulas. So anime will have that from now on. On avg it seems to hover from 75-82 while books avg 60-75. Completely making these numbers up but yeah
5 Likes
Please all the take the time to you need. Youāre working for free to help strangers after all.
The best example of a competitor is https://jpdb.io/ that was posted earlier above.
jpdbās design is a little more cohesive and it runs much faster. It also doesnāt require an email to sign up which I like (Although I understand the some of the many reasons for not doing that). I like that it shows you an example sentence with the card that can be easily changed or translated from a prebuilt list. I like the stats for each book too.
I donāt like that it takes two clicks to move the next card, and how slow it feels (time from clicking āshow cardā to rating it is long), but since itās anki style they show the ābackā of the card every time.
Koohi has a better lesson creation and review process I think. You can add the words you want. jpdb makes you add all of the words at once which is kind of annoying and itās also hard to remove known words quickly. Koohi does that well. Moving between reviews is faster too.
I started using jpdb in an attempt to try to learn some words for a separate book I was reading. I didnāt want that other bookās lessons to āpolluteā my current list. That said adding yet another place to do daily SRS was too much for me and I basically stopped using it.
4 Likes
So for when I come back to the site, Iāve been thinking of adding a āreview setā kind of idea.
So the idea is to split the dayās reviews into more doable chunks (10-30 items per set, out of 170 items or however many you have for the day). The ideal size would be to have a set doable in one quick sitting.
The sets would be divided as granularly as possible based on the following kinds of criteria:
- Leeches
- Book/Book series
- Kanji/Vocab
- Specific Parts of Speech
So for instance, letās say youāre doing 2 books. Book A has 20 cards (across kanji/vocab) and book B has 80 cards.
You might get the following sets:
All Books - leeches (4 cards)
Book A (20 cards)
Book B - kanji (15 cards)
Book B - adjectives (25 cards)
Book B - leftovers (30 cards)
Book B - leftovers (6 cards)
When you go to review you can:
- See all the available sets
- Choose one to do (or have it pick a random one)
Finished sets would stay in the review ui and you can quickly see what words were in them, which ones were wrong, accuracy, etc. Then thereād be an overview for what was accomplished that day (so like: 40 cards reviewed from book A at 80% accuracy, 90 cards from book B at 90% accuracy).
The idea is to accomplish the following:
- Encourage users to do a meaningful chunk of work within a sitting and not feel intimidated by the total number of reviews (esp true for people who donāt review every day)
- Allow users to focus their studies on what interests them at the moment (ex: you can leave book A to rot if you donāt care anymore)
- Group words into groups more thematically meaningful than the current kanji-vocab distinction
Anyway yeah just brainstorming
6 Likes
Just wanted to say thank you for making and maintaining such a cool resource!
3 Likes
Thanks. Iām extremely excited to come back next month
4 Likes
Three day work week secured bois. Iām back on January 3.
5 Likes
Enjoy your time off! 
1 Like
Just saying (again) this site is super usefulā¦I love it and Iād like to have time to use it much more
1 Like
Okay Iām back. Iāve taken a 3-6 month period with reduced work hours to āgo learn stuff.ā This means that Iām back on development for Floflo/Koohiās yearly āf*** this website, Iām going to break everything I spent the last year attempting to stabilize.ā
In this post I will be talking about plans for me and also the website. It may get slightly technical. You may or may not find it boring. Every section will have a summary: the SUMM will avoid any technical speak so that normal people and/or skimmers can read it.
1. Overhaul w/ Emphasis on SEO
When I swapped over to Koohi (a single page application), I had no idea what I was doing. The website has been almost completely invisible on Google because I didnāt set up SEO correctly. At the time, I ignored this because I didnāt even know how to make an SPA and that was enough for me.
However, over time the lack of SEO has made it impossible for the site to grow organically. It is, for all intents purposes, invisible except for when Vanilla or someone tells people on this forum that itās cool.
SUMM: You literally cannot find this site on Google and that is bad. The following sections outline plans to fix this.
1.A Blog
The first thing Iāll be doing is creating a blog. I miss doing posts and find the work somewhat soulless when Iām literally just going full code monkey on the site.
The blog will cover topics related to both Japanese and coding. I donāt believe Iām a crack coder or (even a very good one, to be honest), but I think I have hit a point where I can explain my thought process and be confident that Iām not wrong (even if not optimal).
Like how Floflo was the application that I wish that I had when I was learning Japanese, the coding articles will be the kind of articles I wish I couldāve read when I was learning coding.
Theyāll be focused on solving specific problems ex: āHow was the v2 api designed, what was introduced to the new api and why?ā āHow do we do dictionary word lookups?ā And then the lessons from the articles will be abstracted out so that they can be used on other peopleās projects.
Japanese articles will be the same as usual - opinions, stats (I will be gathering new data), resources, etc.
Iām planning on using a headless CMS (sanity.io) and maybe NEXT.js or something.
SUMM: I will be (re)creating a blog. It will be hosted on its own subdomain. It will cover Japanese and coding, and its goal will be to pull in traffic from Google towards the main application.
1.B Main App
The main application is in an alright state. Planning on overhauling it a bit.
Appearance-wise, I would like it to look less sterile. Maybe just reintroduce some colors or something.
One of the biggest things Iāll be doing is rebuilding using Vue 3 and a state management library (Vuex). Vuex simplifies the process of passing data to child and sibling components and means that I donāt have to start crying into my hands when I decide itās time for the 1500 line component to be split into different pieces.
This should make the app easier to expand. It should also make the app more debuggable (aside from the host of new bugs that will introduce xd) since you can do time travel and snap shots on the app. Also since the state is only mutatable in predictable ways (ex: āADD_A_WORDā), you can be sure that thereās something wrong with adding a word, not adding a word on X page using Y method.
SUMM: In order to handle the SEO thing, it will have a host of new features. For example, the book and dictionary pages will have to become fully fleshed out. This means individual book pages, individual series pages, individual dictionary pages. All of these should be discoverable on Google.
2. Application Overhaul w/ Emphasis on Context
Even during Floflo, we knew that there was significant overlap between certain kinds of content. Consuming content that was related made it easier to improve your reading because you could spend less time learning words in isolation and therefore spend more time reading.
Previously, this was only available by using same-series books, but the next overhaul will expand this.
2.A Genre Lists
So one issue is that categorizations for content does not currently exist. I will introducing user voting so that we can categorize things by genres and tags (ex: romance, action, isekai, etc.). I cannot catalogue this by myself and I wonāt try.
An automated process will use this information to manufacture āgenre listsā. These will serve as something similar to the core 6k but for specific genres. They will contain 1) a word list and 2) a catalog of books used to comprise them. Words will be organized by frequency.
In addition to word frequency in book, you will also be able to see how a word ranks within your chosen genre lists, meaning you can determine if a word will be useful to learn based on the genre youāre focusing on.
The catalog will help you find a next book to read thatās has a genre similarity (and therefore probably a word similarity).
SUMM: Allowing users to tag books by categories/tags. Frequency-based lists will be created based on the books that belong to a tag. You can use these lists to find books with overlapping vocabularies, making reading new books easier. You can also learn from category lists directly.
2.B Sentences
In order to facilitate reading, Iāll be emphasizing context sentences in the next iteration of Koohi.
Weāll be doing this by leveraging Tatoebaās corpus to create automatic context sentences. You will be able to lock in useful example sentences for specific cards.
In addition, Iāll also be introducing a parser (yes, failed to do that last iteration). The parser will automatically create a list out of text. Anything that you add from said list will have the sentence it belongs to added as a context sentence.
SUMM: Adding Tatoeba corpus for context sentences. Adding a parser which will create lists from text. Context sentences will be gathered from said text.
2.C Review Sets
Talked about this here: Weow! Koohi.cafe - A WK friendly SRS [300 vocabulary lists!] - #550 by Raionus
Anyway that pretty much sums up the plans. Currently working on step 1 and reworking the API so that itās flexible enough to do what I want.
19 Likes
I dont know if you are oppen to suggestions, but going to post them just in case,
Hi, just started using the website like 2 days ago, and I have a few suggestions that I have been writting over this days
Everthing has to hight of font weight, I would recommend reducing it the ones at 700 to 500, and the ones in 500 to 300, from the test I have done, just look for example at this, is impossible to see it right
Some letters with lots of stroke you canāt see the strokes because of the hight number font-weight has
In the lection seccion, when you click in the quiz buttom to learn new words, the word should be a lot bigger to obserb better the radicals and strokes of the kanji
Also, the rest of the info would be nice if it could be a litter bigger, but the main word should definitely be bigger and idealy, centered
Also a feature when we can train items that we just learned or have a SRS wait time below 1 or 2 days. I need to force brute it lots of time before I can remember it usually, even more without mnemonic or radicals, so it would be userfull to be able to train items you are having problems with or items you just learned (biggest issue, I already need somethign similar in wanikani to learn some kanjis, and they have mnemonic and radicals, so here is even worse). I usually after learning like 20 lessons in wanikani, quiz myself on it like 5 times all the items, so something similar to that would be great
The Schedule, it would be great if instead of soon, we could get how many hours until it, Even better if there was a table with all the upcoming reviews next 24h or one week, but at least it would be nice to have a concrete time for the next review
Its cool that we can exclude the wanikani level when generating lists, but it would also be userfull if we could include those vocabulary as learned, so if it appears in a book, it shows you the stat of how many words you know correctly. While it would be enought to just enter the level manualy in the case previusly mentioned, I know it would be possible to automatize it, since bunpro has it automatized using a wanikani api key
And of course, if you could get something like radicals for learning the kanjis, would be great, but I guess this is asking for to much
Personally, I give it the next priority, as more important something to train new items, then wanikani vocabulary count as learned on stats, then maybe font and other style things.
But do whatever you want, is still your app, this is just a list of things, that I, as a new user, think would make the app better
1 Like
This would be so helpful!
1 Like
Thanks for sharing your thoughts. Iāve been asked by many people for more detailed scheduling info and I think itās high time it was added. Training mode sounds useful and could definitely be included in the next iteration.
Kanji stroke order drawings could be included and I have a database for it just figuring out how to implement.
Radicals I do not have any info on but many sites have included it so obviously the information exists somewhere
Yeah itās a nice feature. I nearly had it implemented last iteration of Koohi but couldnāt reconcile some key differences between my personal computer and the computer my actual server was running on 
2 Likes
Developed a style to help me see clearer the words, posting it here just in case someone else find it helpfull Website Themes & Skins by Stylish | Userstyles.org
1 Like
