This is a tough one. Ideally there is no input lag and the lightbulb system as it is now works amazing. I get why you opted for that. I’m far from a web-tech expert, but I’d imagine that would be expensive in hardware / ISP costs to improve that.
The others options do indeed come with their own downsides usability wise. I think all of these options are about equal to me. If I had to choose I’d just not hide the row after marking as known. That one has downside if you are just there for marking, but also upside in the sense you can see which one you DO know. Maybe a filter on the page in combination with that could work wonders?
4a+b. I think simplest would be to add a page to the ‘Articles’ section with this information. And maybe rename that section to ‘help’ or ‘FAQ’ or something similar.
4c. Sounds great. I have no real suggestions here since I know nothing about building a website. But that sounds like it might do the trick. Just don’t bankrupt yourself on hosting costs, that’s not worth it
These are smaller issues for sure. It would just be a shame if people go to your site and bounce of it confused because they didn’t find the menu’s. (like me the first time I used the site) Sounds like you have a good handle on that.
I definitely think having a multi-select would probably work best. Web-dev wise, its the smart thing to do, as you don’t bombard the server with hundreds of requests each with a single state change.
Sometime in the future, maybe also add support for WK API keys and importing your current guru’d list of vocab? I definitely find it pretty tedious going through and clicking every single word manually, especially when theres thousands.
To be honest, maybe its better to rethink the account and database approach and take a more Anki-like approach? Something like storing all words and data locally, on the browser, and then adding some sort of sync mechanic.
Although not currently on my to-do list, it is very doable. It’s something I’ll keep in mind as a potential future item.
Do you have examples of where something like this might be used for a site like Manga Kotoba?
One area I could see is when marking words as known, initially store all markings in the browser’s local storage, then after x words marked or y amount of time passes, or a link is clicked, push the word changes to the server all in one go.
I think the main idea with this is that you completely remove the server-side rendering aspect which would remove a lot of the problems with sticky bars, marking stuff taking a while to process, etc. It would basically be like how the site was initially (stuff was stored locally it seemed), but just with an additional sync button to let you sync progress across computers.
I’ve been purchasing manga off of Kobo and running them through Mokuro for a little while now, and wanted to help contribute. However, it seems that none of my OPF files contain the <spine> element, and there are 0 expected files when I drag the OPF file into the browser. It looks like this is the case for all the volumes I have. Do you know if this is something I might be doing wrong on my end, or if this is just the kind of OPF file that won’t work?
I’m currently in the middle of a massive hard drive recovery that’s been going for over a week and is maybe 10% complete. (I have a NAS with eight drives that doesn’t lose data even if two drives fail, and three all failed at the same time.) So that’s slowing me down on a lot of things.
But I’ve mostly got things recovered for working on the site.
If you have Discord, you can also catch me there for other submission options.
I have a small suggestion: would it be possible to add a button (or similar) in vocabulary lists where we could report that an entry is a name, and should be removed?
Otherwise, I’ve been loving this site and I’ve been using it pretty much every day as I’ve been continuing to read and submit more books. Amazing job with it!
I’ve noticed a strange bug with the statistics for this series, not sure if it’s been affecting others.
In general I’ve noticed some “errors” showing in percentages of known vocabulary when opening the series and then reloading it after maybe looking at the vocab pages it would show wildly inconsistent percentages.
It’s not a matter of a couple percentage points where it takes a moment to recalculate after marking words as correct, it was things like jumping from 80% to 20%
It’s on the to-do list! But maybe not just yet, as the site’s getting some structural improvements at the moment including with how it interacts with the database. (This specific feature would require modifying an existing database table or adding a new one.)
When I first add a volume, it defaults to 100,000 words and “undefined” for various stats, until I run a separate process to update those.
I should probably add something that hides entries like this so they don’t show until I run the stat calculation. I’ve been thinking of how I want to implement not showing a series before it’s ready, and hiding based on these values would be a good automated solution.
I’ve updated the stats for the one you showed. Thanks for letting me know about it! I did a database query for any others and didn’t find any, but don’t hesitate to let me know if you happen upon one. I’m always grateful for help catching things I miss!
The issue is I was using a height of “100vh” (100% viewport height) which for mobile devices doesn’t factor in the height of UI elements. Thus the page is taller than the browser’s height. Add in hiding the overflow, and the bottom of the page cuts off without a scrollbar.
The solution? Using a height of “100dvh”. This “dynamic viewport height”, which I hadn’t heard of before, resolves the issue with “vh”.
Apart from that, the site’s up to 8,771 volumes across 3,410 series, with more on the way!
The icon is a relic of the original admin-only design (so I don’t accidentally block words unintentionally). Eventually, I’ll probably remove the icon, but for now, you’ll need to select the icon to reveal the block feature.
Once enabled, the “block word for this series” icon appears:
Select a reason, then select “Submit Recommendation”. If the submission succeeds, you’ll get a brief message saying so, and the block icon becomes disabled for that word.
Interesting, will check it out!
So a blocked word won’t appear in vocabulary lists anymore for all series? And it doesn’t get counted for statistics anymore?
A block applies only to a single series and does not impact other series that may use that word. (In some cases, I may opt to apply a block for all series.)
Correct.
Consider a volume where you know 1,245 out of 1,491 words (83.5%).
Let’s say this includes a character name that appears 25 times in the volume.
Once the name is recommended for blocking, and I approve it, the stats change to knowing 1,245 out of 1,466 words (84.9%).
Yesterday evening on mobile I’ve had a bug where, even after marking a word as “known” (lightbulb button) it disappears from the list, but after reloading the page it’s not actually getting added because it appears again. Haven’t tested it on pc because I don’t have it available at the moment
Also the formatting of the page has changed on mobile making it a bit uncomfortable to use, before scrolling on mobile I could easily see the word, translation, and mark it as known all in one glance, now the mark as known button is barely out of screen forcing me to use desktop mode (which zooms out and makes it easy to accidentally misclick and mark the incorrect word)
I’ve attached some screenshots.
Using mobile mode, vertical (the viewport actually doesn’t move at all and the rest of the page is cut off, making it impossible to mark words):
