Our data is released under various Creative Commons licenses.More information
If you love this content, please consider a donation.
Wall (5,877 threads)
Tips
Before asking a question, make sure to read the FAQ.
We aim to maintain a healthy atmosphere for civilized discussions. Please read our rules against bad behavior.
AlanF_US
2 hours ago
TRANG
2 hours ago
sharptoothed
2 hours ago
AlanF_US
4 hours ago
TRANG
7 hours ago
AntonKhorev
7 hours ago
CK
15 hours ago
carlosalberto
19 hours ago
heo598
19 hours ago
TRANG
yesterday
**Transition towards the new design**
I will need people to test the dev website for the transition towards the new sentence design.
https://dev.tatoeba.org/
On the dev website, the new design should now be enabled for everyone by default. For those who didn't have it enabled yet, you should see an announcement message about it.
There are two new things compared to the main website:
1) The expanded sentence menu will appear above the sentence ID and username (instead of hiding it).
2) There's an option to increase the number of translations visible by default. This option replaces "Display a link to expand/collapse translations" because on the new design, the expand/collapse link will always be there.
You will still be able to switch back to the old design if something in the new design is blocking you from doing the things you used to do. But this option won't be there forever. You are encouraged to try using the new design as much as possible and report any blocking issues so that can fix them.
Thank you for testing!
It seems to take a long time for the sentences to be displayed.
I could see the comments for a long time before I could see the sentence and it's translations.
https://dev.tatoeba.org/eng/sentences/show/1434
I also wonder why it's necessary to hide translations at all on a sentence's page.
All the data is downloaded anyway, so why not just show it by default.
The plain html page for the above sentence was 643 KB when I downloaded it to check.
This doesn't count the external files, but many if not all, will likely be cached on the user's device.,
I can understand why it might be desirable for pages of sentences like what you get when you browse by language or do a search, since some people doing searches may not be so interested in seeing all the translations.
Experimenting ...
For comparison, in the old layout, when downloaded, the plain html file was 1.1 MB. I assume that the extra filesize has a lot to do with all the lists that are part of the page's code. I have a lot of lists, plus all collaborative lists add even more.
In the new layout's page, I changed every occurence of " to " and the file size went down to 469 KB (72% of the original). Perhaps the code will work as well with this, perhaps not.
I also wonder whether you could cut out some of the data that is being sent to perhaps speed things up and to require less bandwidth for those who don't have unlimited bandwidth internet plans.
Here is one of a lot of sentences on that page.
{"id":499788,"text":"Estoy enamorado de ti.","lang":"spa","correctness":0,"script":null,"transcriptions":[],"audios":[],"lang_name":"Spanish","dir":"ltr","lang_tag":"es"},
Is this necessary?
"lang_tag":"es"
Also, couldn't the following be generated easily enough from the "lang":"spa" part.
"lang_name":"Spanish"
All it would take is a 2-field external Javascript file and a bit of code. I can send you an example if you need one. The Javascript file would likely get cached, so there would be less being loaded each time from tatoeba.org.
Overloading a page with information is usually not a good design. If there were thousands of translations, I'm not sure why you would need them all to displayed at once. At some point you have to set a limit and I can understand that 5 translations might be too low for some people, but there has to be a limit.
The fact that we retrieve all translations isn't on purpose, it is just that we never spent time optimizing this part of the code.
Next to that, there are performance issues when it comes to rendering a page that has more and more content. Unfortunately, the new design is reaching this limit much sooner than the old design.
There is a lot of optimization to do. There's no doubt about that.
> but there has to be a limit.
There's no limit now and it doesn't seem to cause any issues.
Also, I believe it would be a good idea to add the "auto-expand" as an option. The default can be displaying 5 sentences, as now, but I would love to be able to see all translations at once without having to click on each single sentence to see more translations.
> The default can be displaying 5 sentences
Are you aware that you have the option to increase to 50 translations?
> Are you aware that you have the option to increase to 50 translations?
I am, but why not make it infinity? This will not really affect the speed the page is loaded because, as CK had noted, you do load all the translations anyway.
The fact that we retrieve all translations isn't on purpose, it is just that we never spent time optimizing this part of the code.
So you're saying once you optimise the code it will get even worse? Not only will we have to click Show 1000 more translations, but it will also take much longer to display them? Repeat 20 times for each sentence displayed in the search results.
One way to optimize, which perhaps you've already thought about, is to not automatically load in translations of translations, but only load them in later on demand.
The loading time shouldn't be any longer, but there is a chance that the time to display the page will be longer. I think that's what TRANG meant.
I guess she's stuck with doing it this way because it's part of Angular.
List the source of this web page to see a much lighter-weight way to send data and render it on the user's side.
http://www.manythings.org/sente.../always/1.html
I suppose something like this could be done on tatoeba.org, but it would require a lot of re-coding perhaps.
> There's no limit now and it doesn't seem to cause any issues.
There are already some issues even with the old design:
https://github.com/Tatoeba/tatoeba2/issues/2041
https://github.com/Tatoeba/tatoeba2/issues/1737
Note that the data is gzip-compressed during transfer, so the page isn't quite as bloated as it seems at first glance. According to the Firefox developer tools, the one you linked above requires 1.72 MB (compressed to 1.22 MB during transfer) but only 62.87 KB get downloaded again when I reload the page and everything else is cached. The bare HTML is 62.45 KB compressed and 585.19 KB uncompressed.
It's probably still possible to reduce the size substantially, but that would require focusing on the parts that don't compress well instead of those that repeat a lot.
Several issues, which I reported on GitHub yesterday:
https://github.com/Tatoeba/tatoeba2/issues/2350
and today:
https://github.com/Tatoeba/tatoeba2/issues/2352
https://github.com/Tatoeba/tatoeba2/issues/2353
https://github.com/Tatoeba/tatoeba2/issues/2354
Some of them affect the production site as well as the dev site.
The issue about the vocabulary count is independent from the new design.
For the other issues I will reply in the comments over there.
Could you please leave the old sentence design as an option?
Not forever. Keeping it has maintenance cost.
Do you have any date in mind when you actually get rid of the old design in favour of the new one? A rough estimate.
I only know for sure that it won't happen before July.
Perhaps it will have to wait until Kodoeba is over because organizing this event is going to drain a lot of time and energy from the dev team. But I don't really know yet.
I see.
To avoid the problem of having the maintenance cost, maybe it would be a good idea to get rid of the new design and to stay with the old one?
Then you'd have to elaborate why you think it's a good idea.
Because the new design has literally zero advantages.
It has many disadvantages though. It's bulky, ugly and it doesn't let you see more than 50 translations at once, unless you click the "Show 200 more translations".
So if the choice is between something not perfect, but functional, and something not perfect but low functioning, I guess it makes sense to choose the first option.
I can list several features and bugs that are solved in the new design but are still an issue in the old one.
https://github.com/Tatoeba/tatoeba2/issues/2051
https://github.com/Tatoeba/tatoeba2/issues/1933
https://github.com/Tatoeba/tatoeba2/issues/1911
https://github.com/Tatoeba/tatoeba2/issues/1874
https://github.com/Tatoeba/tatoeba2/issues/1783
https://github.com/Tatoeba/tatoeba2/issues/1709
https://github.com/Tatoeba/tatoeba2/issues/1678
https://github.com/Tatoeba/tatoeba2/issues/1614
https://github.com/Tatoeba/tatoeba2/issues/1599
https://github.com/Tatoeba/tatoeba2/issues/1528
https://github.com/Tatoeba/tatoeba2/issues/1332
https://github.com/Tatoeba/tatoeba2/issues/536
https://github.com/Tatoeba/tatoeba2/issues/161
Saying that the new design brings zero advantages is a bit unfair. Perhaps for you, it has zero advantages and I cannot argue against that.
But overall, the problem is that the old design was tailored for desktop usage and people are now increasingly using mobile devices. We have to adapt to this shift, somehow.
I know very well that some users (you included) have built years of habits on top of the old UI and it makes it much harder to transition. If we had all the money and brains in the world, then our approach would be different. We would have several teams of developers, each working on a different UI that is optimized for a different device and even for different types of users. We don't have that.
At the moment we can only maintain one UI and rather than sticking forever to a desktop-oriented UI, the decision was to try and build a UI that finds a middle ground between desktop and mobile usage.
In terms of statistics, as far as I can tell, there's a majority of users who are currently fine with the new design:
- 263 users have the option "use_new_design" set to true
- 110 users have the option "use_new_design" set to false
I'm not saying the new design is great, but it is still preferred over the old design by many people. Your arguments for getting rid of it are strongly personal. From my perspective, it still doesn't look like a good idea to get rid of it.
> I know very well that some users (you included) have built years of habits on top of the old UI
That's not only habits. Tatoeba became kind of second hobby to me, and part of the experience that I enjoy is comparing all the different translations to all different languages. Not seeing all those translations unless you click on every single link in the search result is quite less enjoyable.
I do agree it's a personal and very subjective opinion, and I probably get too emotional about one of my favourite hobbies slipping away from me.
I do agree you don't have to accommodate for me. I just expected this feature (being able to display all translations) to be liked by many people, but this might not be true.
I don't mind changing my habits and getting used to the new design, but one of the new features is literally a big step back from my point of view.
> I do agree you don't have to accommodate for me. I just expected this feature (being able to display all translations) to be liked by many people, but this might not be true.
However, two of your main "disappointments" have been expressed by other people: "bulky" design and not all translations visible (cannot retrieve the exact GitHub ticket).
I would just like to mention that the new U.I. and its set of features will not be carved in stone forever. I think the new design will be more scalable and manageable, and once we get rid of old issues preventing this evolution, new features will keep on being added. With a little bit of luck, the feature(s) you find most enjoyable may find its way back into Tatoeba.
> - 263 users have the option "use_new_design" set to true
> - 110 users have the option "use_new_design" set to false
We have many more than 263+110 users.
We have at least 6,000 contributors with sentences in their native languages.
Does that mean 4.3% (263/6000) have the new design set to true, assuming we only have 6,000 users? Of course, we have more users than that.
> Of course, we have more users than that.
But not all of them are really active. For example, since the beginning of this year 1544 distinct users (excluding spammers and suspended users) have logged in at least once.
263 out of 1544 is still not a majority.
Correct, but you can't compare these two numbers.
263 is the number of all users in the database who have set the setting "use_new_design" to "1" (= true), i.e. the changed it from the default value "0", false or "" [*]
1544 is the number of logged in users since 2020-01-01 disregarding of their user settings.
[*] Now why are there three different values for setting "use_new_design" to false?
The default value in the PHP code is "false".
When a new user registers, we don't save anything in the database for all settings and just fallback to the default values. (i.e. the settings column in the database is empty)
Whenever this user closes a banner we display to them, we store this action in the settings. But since we store all settings together in a JSON object, all other settings will be stored with their default values (case 1: "use_new_design":false)
Whenever the user visits their settings page, the checkbox for "use_new_design" is initialized with the empty string due to the default conversion from PHP false to the empty string.
If the user doesn't change that setting but some other one we get back the empty string but don't convert it back to PHP false and just save the empty string (case 2: "use_new_design":"")
If the user changes the setting we get back "1" (in the HTML code we use "1" for true and "0" for false) and save that in the database. And when the user later on changes the setting again, we now get back "0" and save that in the database (case 3: "use_new_design":"0")
Now the value Trang posted is the number of users who fall into case 3, i.e. users who have once switched to the new design and later on switched back to the old design (I'm one of these users because when I test something I always switch between both designs).
We don't really know whether the users which fall into the first two cases just don't care about the design (yet) or don't want to switch (yet).
There is another issue: The settings "use_new_design" only exists since October 2016. So users who registered before and never modified their settings as described above since then, won't have this setting stored in the database.
(There are 43358 users who don't have this setting, 847 of them logged in since 2020-01-01 and 459 of them contributed at least one sentence).
So we will see whether we get more complaints about the new design in the following weeks.
> But overall, the problem is that the old design was tailored for desktop usage and people are now increasingly using mobile devices. We have to adapt to this shift, somehow.
I had used tatoeba on a desktop only a few times. I mostly use it on a tablet. I prefer the old design.
The old design's disadvantages are small buttons.
The new design's disadvantage is *we'll show you translations to five random languages and bury the ones you want to see under 'show more' button*. But I guess this is now fixed by an option to show 50 translations instead, so you'll just have to scroll for longer because the sentences are bulkier.
> we'll show you translations to five random languages
Translations are currently ordered alphabetically by language name.
> But I guess this is now fixed by an option to show 50 translations
> instead, so you'll just have to scroll for longer because the sentences
> are bulkier.
It's more like a workaround for the problem you mention. It does serve the purpose that some people will want to see more than 5 translations at once. It doesn't solve the problem that some users will have to look at the bottom of the list no matter what because the languages they are interested in happen to be last in the alphabetical order. To solve that problem, we have to allow the users to configure which languages they want to see in priority.
For the bulkiness part, you can read further explanations on the related GitHub ticket created by Alan:
https://github.com/Tatoeba/tatoeba2/issues/2354
As you can see on that GitHub page, I was able to use code created by Trang, together with instructions by AndiPersti, to configure my Firefox/Windows setup to display Tatoeba in a custom compact mode, reducing whitespace and eliminating the "Translations" and "Translations of translations" captions. I'm satisfied with this approach, since I understand why Trang doesn't want to add an option for this mode to the standard user interface.
Trang, is there an issue ticket that tracks the feature that will allow users to configure the priority of the translations they see by language?
> Trang, is there an issue ticket that tracks the feature that will allow
> users to configure the priority of the translations they see by language?
I created one just now:
https://github.com/Tatoeba/tatoeba2/issues/2364
Thanks!
> Because the new design has literally zero advantages.
That's not true. For one thing, it makes the relationship between the sentence, the translations, and the translations of translations more clear. That's a big advantage, especially for people who are new to the site.
Especially since I am already familiar with the site, I do wish there was an option for a compact mode, as I mentioned in the issue ticket I wrote:
https://github.com/Tatoeba/tatoeba2/issues/2354
Notice my approach:
(1) wrote an issue ticket
(2) included a screenshot
(3) was specific about what I wanted to change and what I could live with
(4) was polite
I figured that being specific and polite gave me a better chance to have my issue addressed. Also, I like the developer team and don't see a point in antagonizing them.
Alan, if you consider yourself polite you're very wrong. You're constantly bullying other people here, you're a meanie and a bully.
I, too, would like to see the old design stay in place for a while.
I think it should definitely stay in place until the website has been converted over to mobile-friendly, which I think was the main purpose for changing this layout.
Keeping both the old and new design has maintenance cost. For everything we do, we have to care about not breaking the old design as long as it's there. And among other things, it is difficult to move towards a mobile-friendly UI while knowing that there's still an ancient component which will easily break if we change the layout or change part of the CSS.
One of the buttons I'd like to see whenever a sentence is displayed is the "♥" (favorite) button.
Sometimes it occurs to me that a sentence is somehow important and so, I'd like to add it to my "personal gallery" (my favorites) instead of creating a list.
As for now, I have to click on the arrows to be able to favorite a sentence.
====
Perhaps in "Number of translations visible by default" there could be a way to display which sentences we'd like to see first (in which language, for example). It has been discussed before, thought.
For the favorite button, you would have to use the feature to expand all menus. You would then see all buttons all the time.
The collapsed menu can only support 3 buttons and the (arbitrarily) chosen features were:
- translate
- edit
- list
If we display more, the username will sometimes be hidden for people who have long usernames. Unless you can justify that the favorite button is more often used than "translate", "edit" or "list", then it will have to remain in the hidden buttons.
The option to prioritize which language to display first in the list of translations has been discussed before, yes. I don't think it will be implemented before we switch to the new design though, it's not a very easy feature to implement.
How can I link two (distant) sentences together using the new sentence design? I can't find the chain icon that could be easily found in the old design.
And to be honest, I think this new design is not very practical for maintenance purposes. Editing a sentence and changing its flag should be easier (at least for admins and corpus maintainers), but these tasks seem to be much more complicated and time-consuming now than they used to be.
I do like the way things look in the new design, but IMHO it's very unpractical.
To link sentences, you would now have to make use of the mechanism to auto-detect duplicates.
1) You copy the text of the sentence you want to link.
2) You add it as a translation of your current sentence.
Editing a sentence requires the same steps as before:
1) Clicking the edit button.
2) Changing the text in the input.
3) Pressing enter or clicking save.
Changing the language requires only one more step:
1) Clicking the edit button (as opposed to clicking the flag).
2) Selecting the language.
3) Clicking the save button.
If any corpus maintainers has to fix languages so often that it is needed to eliminate the 3rd step, I'd like to know how frequently it happens and why it happens (because it really shouldn't happen often).
Thanks, Trang. I guess I can live with that. ;-)
* Tatoeba Top 30 Languages Interactive Graphs * have been updated:
https://tatoeba.j-langtools.com/igraph/
https://tatoeba.j-langtools.com/igraph/share.html
https://tatoeba.j-langtools.com.../index-14.html
I think it might be a good idea to eliminate the question mark (?) wildcard for searching.
I think that users should be able to paste in any sentence in the search form and get a result if the sentence is in the database. This doesn't happen now for sentences with question marks.
I think this usefulness far outweighs the usefulness of the one-letter wildcard.
If you're a member of this website and frequently use the ? wildcard, perhaps you could tell us how often and whether you'd be greatly inconvenienced if this were removed.
Mi ne ricevas nun retpoŝtajn sciigojn. Kial?
The email notifications are still working on our side, as far as I can tell.
Are they maybe marked as spam on your side?
Yes, they were marked as spam. The problem is solved. Thank you very much.
is it ok to webscrap Tatoeba?
Since Tatoeba doesn't provide an API, I've been developing a simple API myself that scraps the search results and returns them as JSON data. I'm playing with an Android app made with React Native as a way to quickly search for words without having to load the website on mobile browsers. I'm also developing another personal web app which uses Tatoeba Japanese sentences but as I didn't know how to properly download and use the data in the beginning of the project (and how to do it with good performance for the web), I decided to use the same webscraping mechanism.
I read the ToS and there's nothing against it on there. I'm aware that there's so negative views about webscraping out there but in my defense, it would be the same as a user requesting a search page, only difference is my code parses and returns the data in another format with no further stress to Tatoeba's servers.
There's no automated scraping involved.
Perhaps having an API of this type would also help anyone else who wants to participate in Kodoeba or simply use data from Tatoeba in a personal project?
If it's ok and anyone is interested, I would be glad to share the code :)
Yes, it’s perfectly fine to webscrape Tatoeba as long as the scraper is not too aggressive and you give proper attribution (the contents are under Creative Commons CC-BY).
We are aware that there is a need for a proper API. It’s just that nobody had time or energy to create an API so far. You, or anyone else, is welcome to contribute. We think that creating a read-only API could be a good project for Kodoeba (especially if other Kodoeba participants want to build something based on that API, that would be a nice collaboration between participants).
Thanks for the answer!
I've been contributing a few sentences for a while (not consistently though) and have looked at the repo a few times but as it's mainly written in PHP I have no clue how I could contribute to create an API. My experience is mostly with Python and Django, I'm also working with React but it seems Tatoeba uses Angular instead?
Some time ago I even researched a bit about CakePHP and how to make a REST API but couldn't find anything simple for a non PHP dev so I gave up.
In Django it's so simple, I just have to use drf and create a few views and paths and voilà, I have an API. If anyone could point me some resources for creating an API with CakePHP, how it works, etc, I would like to try at least.
> it seems Tatoeba uses Angular instead?
Yes.
> If anyone could point me some resources for creating an API with CakePHP, how it works, etc, I would like to try at least.
Well, there's the cakephp.org site, which has a "CakePHP at a Glance" page:
https://book.cakephp.org/4/en/intro.html
though I don't see anything on that page about creating an API. You could also take a look at the "Contributing as a developer" page:
https://github.com/Tatoeba/tato...as-a-developer
I know the developer team would be glad to have you help them in any way!
How many developers are there in the team and may we know their names?
At the moment, the core developers are:
- myself, since the very beginning
- gillux (jiru on GitHub), since June 2018
- rumpelstilzchen (AndiPersti on GitHub), since April this year
Core developers have write access to the main source code. Any code change submitted by anyone else has to go through the approval of a core developer.
You can the full list of people who contributed to the code here:
https://github.com/Tatoeba/tato...s/contributors
You can see the latest changes in the code here:
https://github.com/Tatoeba/tatoeba2/commits/dev
> it seems Tatoeba uses Angular instead?
AngularJS to be more specific.
> Some time ago I even researched a bit about CakePHP and how to make a REST API but couldn't find anything simple for a non PHP dev so I gave up.
The main docs are at https://book.cakephp.org/3/en/d...ment/rest.html and https://book.cakephp.org/3/en/d...estful-routing
> In Django it's so simple, I just have to use drf and create a few views and paths and voilà, I have an API.
It turns out that with CakePHP it's also pretty simple. I've just created a simple demo which you can find at https://github.com/Tatoeba/tatoeba2/pull/2341
> My experience is mostly with Python and Django,
When I've started contributing to Tatoeba I also didn't know much about PHP and CakePHP but it's not hard to learn if you already have programming experience. You can contact me any time if you need some help.
Hi, rumpelstilzchen, thanks for your answer and sorry for the late reply.
Judging by your demo, it really doesn't look as complicated as I thought it would be.
I'm currently working on a side project and would like to finish a few plans I have for it in the next weeks before starting to contribute, after that I'll take my time to (re)learn PHP and better understand REST and how to implement it with Cake.
I say (re)learn because around five years ago when I started coding I was using Codecademy and went through their PHP tutorial at the time but I forgot most of it because I've never used it ouside of the tutorials. I also think I should step up some of my skills before trying to contribute to other projects. I'm sure you get what I mean, writing code on a personal project is one thing but contributing to such a huge project as Tatoeba is something entirely different!
Although I've learned a lot over the last 2 years, which is basically the time I have been working on my current project, I feel like there are a few things I should go back and study more.
Anyways, I hope to finish my project and start soon, meanwhile I'll keep contributing a few sentences here and there when I can.
** Stats - 2020-05-30 - Native Speakers with Contributions **
http://tatoeba.ueuo.com/stats-2020-05-30.html
Find out who the native speakers are and get links to their sentences.
Will the new design for the sentences eventually replace the old design for everyone, or will the old design be kept as an option, as it is now?
There will be a period of transition, but sooner or later we will get rid of the old design entirely.
Thanks for replying. Sad news, I guess.
Are you being melancholic?
You can put it like this, if you wish. The new design is terrible and clumsy to use. I'd hate to have it, but, obviously, we won't have a choice, which I accept.
Well, you could explain what is terrible and clumsy to use, and we could see how things can be improved, as a first step :)
1. By default, some translations are hidden in the new design and I need to click "Show N more translations" every time in my search results:
https://i.imgur.com/YyBRn8s.png
It is very annoying and very counterproductive. I do understand why someone might have this preference, but can we make it optional? Create another checkbox - "Expand all the translations". By default, if the user doesn't bother to change it, the sentences will be displayed as they're displayed now, but some of us can make it more comfortable for us.
I am aware you can permanently limit yourself to just a bunch of languages in your profile, and this issue will become less important, but I'm going to do it because I enjoy being able to see all the translations.
EDIT: I'm also aware of this setting:
"Display a link to expand/collapse translations when there are too many translations"
But it seems to be working for the old format only. I guess if you make it work for the new and the old format, that would resolve this problem.
2. The second thing that bothers me A LOT are the words "Translations" and "Translations of Translations":
https://i.imgur.com/rZCd1yi.png
I understand why they are there - a lot of users, especially the non-regulars or the relatively new contributors are confused by the whole idea of "gray" links and report the indirect translations as wrong while they're not wrong, they're just translations of translations and they don't have to match the "original" exactly. But for more experienced folks who understand and embrace the concept of indirect translations those words are just a nuisance, completely unnecessary information in words that can be perfectly color-coded. Could we optionally remove those words, leaving them there by default?
3. A less important issue - the new format is just less compact than the old one. Compare (this is the same sentence with the same translations):
https://i.imgur.com/4yuO9YJ.png
Thank you for your always useful feedback. :)
Well, I don't think there is so many terrible things there!
1. This setting needs improvement. Notice that the new design isn't complete yet so functionalities may be behave incorrectly or be missing. Even after the new design implementation, that might still be the case and we hope we can count on people's feedback to correct and improve them. As gillux mentioned, there will be a transition period, but the old design will only disappear when we will be absolutely sure that nothing's gone missing in the new design.
2. That's your personal opinion, so there isn't much to say. I can't say two. The first one is that of course, we can discuss that point. The second is that I'm not sure for how long you've been using the new design, but maybe after a while, you'll get used to it and those words won't bother you anymore. Once you're used to the new design, they might get easily ignored by your brain.
You also mentioned color-coded sentences. That's is also currently under discussion.
3. Yeah, I've noticed that too. Now, we need more scrolling. Not a really impacting issue, but it has been reported.
I hope I could answer your points clearly.
As a final, more general, note, change is always a difficult thing, and we expect a lot of complaints from a lot of people. However, if people could start by "here what I like and dislike now" before the "this crap is terrible", that would be appreciated and we would be able to discuss with a better set of mind. Thanks.
Hi Aiji,
I really hope you're not part of the development team, or at least not a decision maker of what will be implemented or not.
Your smug, dismissive and passive-aggressive way of dealing with our concerns is not ideal.
Your active-aggressive way of expressing your concerns wasn't ideal either.
Now, I hope I could give you an overview of what was currently being discussed on GitHub (mainly) and that a part of your concerns were already discussed or will be, and if my answer seemed too aggressive, I apologize. My thanks for your feedback were sincere, as I think you're one of those who very often gives precious information regarding how we could improve things.
How do I get specific languages to be above "show more"? Do I need now either to always click "show more" or to hide most of the languages completely?
You can deactivate this link in your settings: https://tatoeba.org/eng/user/settings
If you want it activated but would appreciate it to behave differently, please elaborate on what you would like to have and *why*.
From the same settings page, you can also choose a list of languages to only display sentences in those languages (all others will be hidden).
If I choose a list of languages, all other languages will be completely hidden. They won't be displayed if I click "show more". Here are my options if I don't hide languages completely:
Best option (not available) - show my preferred languages first; if there's a lot of translations, put everything else under "show more".
Worse (old design) - show everything; it's up to me to find sentences in my preferred languages among all others
Even worse (new design) - hide some translations no matter if they are in my preferred languages or not; I have to always click "show more" and then do everything I had to do in the previous option, so it's more work. Also, the list takes more space.
As noted above by deniko, it's impossible to disable "show more" in the new design. But even if it would have been possible, new design would still be worse for sentences with many translations.
> Best option (not available) - show my preferred languages first; if there's a lot of translations, put everything else under "show more".
That would actually be a cool solution. Thanks for this idea.
I would still prefer to be able to expand the sentences by default without clicking, but if you can actually list the languages that you want to be displayed first that would really be awesome. And would be cool if you can specify the order in which they're displayed.
For example, if I specify
eng,spa,fra,ita
In my profile, I would see the translations in those languages in that order, and only after those 4 languages I'd see all the other translations in their default (alphabetical, according to the language code) order.
The existing solution of completely hiding all the other languages is just not satisfying at all, at least for me.
I agree with Denis and Anton. The new design is extremely frustrating.
Is it for the same reasons as them or do you have other reasons?
The reasons are the same. They described them very well and I have nothing to add so far.
Then please see my reply to deniko below.
https://tatoeba.org/eng/wall/sh...#message_34492
I have the same questions for you as I have for deniko.
Most importantly the last question:
> Could you provide a scenario of your activities in Tatoeba that can make
> us understand "Okay, indeed, if the translations were not expanded by
> default, it would be annoying"?
When I see a sentence in a foreign language I know, first of all I look whether it has already been translated into Russian or not.
If yes, I proofread the existing Russian translation(s) (not only for translation accuracy but also for punctuation, spelling and grammar). Sometimes I add my own translation(s) as well.
If not, I add one or several Russian translations.
If I see that my newly added sentences can be linked to the existing translations in the other foreign languages I know, I do it right away or later from this page:
https://tatoeba.org/rus/activit...ces_of/marafon
Besides that, I check the indirect Russian translations and link them to the main sentence if they match.
I also check the indirect translations in the other languages. Sometimes it helps me find some linking errors like this one:
https://tatoeba.org/rus/sentenc...omment-1165179
Just looking at the indirect link here:
https://tatoeba.org/rus/sentences/show/5918489
helped me find this error.
I'm also interested in some languages that are not listed in my profile. For instance, I enjoy seeing the Ukrainian, Czech or Italian translations.
It's just one of my activities in Tatoeba. I'm not sure I described it well enough but I hope it helps you understand why I find the new design and particularly the button "Show more" so frustrating.
I used to love the random sentence on the main page. But not anymore.
p.s.
> Now it works super weird - I add
> 3 Ukrainian translations out of habit, only to realize 20 seconds later
> it already had Ukrainian translations, they were just hidden.
That was my experience, too.
I created the following issue on GitHub:
https://github.com/Tatoeba/tatoeba2/issues/2210
If you were able to increase the limit of translations that you can see, what would be your limit? Would it be higher than 50?
On a side note, your description actually brings up other fundamental issues in the overall design of Tatoeba. It's a problem with the way Tatoeba is structured and organized. But addressing that problem is a whole other story so I won't delve into the details here.
50 would be enough, I think.
Btw, now I don't see the way to link the indirect translations to the main sentence. That's what I used to do all the time.
The linking feature has not been implemented yet in the new sentence design.
OK
And also the feature of reviewing?
Here an example:
https://tatoeba.org/epo/sentences/show/8604386
This sentence has been corrected, so my review has to be changed from red to green. For doing this I need to switch back to the old design, no?
Sorry if that has been answered before. I didn’t see it.
That will be implemented later :)
> Best option (not available) - show my preferred languages first;
A similar idea has been mentioned before:
https://github.com/Tatoeba/tato...ment-505311261
Additionally there has been some debate regarding the order in which translations should be displayed:
https://github.com/Tatoeba/tato...ment-577132284
> And would be cool if you can specify the order in which they're displayed.
That's also my opinion.
Based on the various discussions on GitHub, it became quite clear that we need to provide a better way for users to define the languages they want to see in Tatoeba and what is the priority of each language.
I'm still wondering however what is your use case for preferring to have all translations displayed by default. Do you confirm that it is because most of the time, the translations you are interested in are not displayed at the top? Or is there another reason?
Also, which of your activities does it affect the most? Is it slowing you down for translating sentences? For linking sentences? For proofreading? For something else?
Could you provide a scenario of your activities in Tatoeba that can make us understand "Okay, indeed, if the translations were not expanded by default, it would be annoying"? Right now I must say I can't imagine this scenario.
Joskus, yksinkertaisten lauseiden tapauksessa, löydän uusia kieliä joita voin linkittää. Esimerkiksi suomalais-ugrilaisten kielten numerot ovat hyvin samanlaisia, enkä tiennyt niistä kaikista ennen kuin selasin kielilistaa läpi sen lauseen kohdalla, joka listaa numerot ykkösestä kymmeneen.
Samoin olen joskus linkittänyt satunnaisen latinankielisen lauseen, vaikka en suuremmin latinaa osaakaan, jos olen sattunut tietämään mitä se tarkoittaa.
Usein selatessani Tatoebaa luen lauseita kielillä, joita en osaa tarkasti, saadakseni tuntumaa niihin ja kenties oppiakseni jotain. Tämä kattaa monet romaaniset ja germaaniset kielet, suomalais-ugrilaiset kielet ja latinan. En osaa kääntää näistä tai edes lukea näitä itsenäisesti ja näiden näkemin on huomattavasti vähemmän tärkeää kuin niiden kielten, joita osaan jonkin verran.
> Also, which of your activities does it affect the most? Is it slowing you down for translating sentences? For linking sentences? For proofreading? For something else?
Linking mostly, not proofreading.
When I link, just browse through Ukrainian sentence using the search feature to give me random sentences, and link them to all sentences they can be linked directly.
Also, when I stumble upon some interesting expression I really like searching for it here and check out translations to all languages I kind of understand, even those not listed in my profile just because I'm curious. I'd say that activity is very important for me.
Also, I used to enjoy having the random sentence on the main page. I don't know why, but I've always thought it to be a fascinating feature. I've stumbled upon some real gems there a few times, something I would have never found if I was trying to search for something. I would occasionally translate them. Now it works super weird - I add 3 Ukrainian translations out of habit, only to realize 20 seconds later it already had Ukrainian translations, they were just hidden. I turned on the random sentence because all this was very annoying.
So, in general, I understand why someone might prefer the new format, and I understand why you would prefer this to be the default format for the unregistered users and the default format for everyone else, but I kindly ask either to leave the old format for us, or at least add t those two features as optional features:
1. Expand them automatically, if my settings say so.
2. Remove the words "Translations" and "Translations of translations" as completely unnecessary and distracting (again, leave them there by default, but please allow us to get rid of them).
Everything else doesn't really bother me too much. Making the new format more compact or dense is nice, but not really a deal-breaker. Same about the feature mentioned by Anton - allow users to configure which translations to display first. That's a cool feature, I'd love to have it, but purely psychologically I feel like it's less important for me.
One thing to get a cool new feature - yeah, that's nice, but I understand why it's not there - another thing - taking a nicely functioning design and breaking it to something barely functional - that feels bad, like a giant step back.
Olen samaa mieltä kohdan 1 ja motivaation kanssa.
Kohta 2 olisi myös minulle parempi, mutta toisaalta ylimääräisten sanojen aiheuttama harmi on hyvin pientä, ja liika valinnaisuus käyttäjillä aiheuttaa kasvavaa työtaakkaa ohjelmoijille. Niinpä ehdotan, että vaikka tiiviimpi käyttöliittymä olisi mukavampi, ei ehdotus kaksi luultavasti ole pitemmässä juoksussa toteuttamisen arvoinen.
Thanks, Thanuir, you can be right about point 2, but I just feel they take space and they don't do anything useful for the experienced users. The direct and indirect links are already colorcoded by the color of the arrows (or > in the new design). It feels right to use words where they belong, for the important information - the sentences themselves - and icons and other mnemonics for technical information and actions.
All that feels consistent.
Now, using words to distinguish direct and indirect links just seems to be very inconsistent and it makes me uncomfortable. I will absolutely get used to the less compact design, but I just liked the consistency of everything else.
While this might be way less important for others, I still feel like it's something easy to implement and hopefully I won't be the only one who chooses to opt out from seeing the words "Translations" and "Translations of Translations".
I wonder whether I can join the development team just for this single task? I'm a software developer by profession, after all, although I do lack any web-development experience.
> I wonder whether I can join the development team just for this single
> task? I'm a software developer by profession, after all, although I do lack
> any web-development experience.
Tatoeba is open source. You (and anyone) can make a pull request on GitHub.
Please see our guide for people who want to contribute as developers:
https://github.com/Tatoeba/tato...as-a-developer
You can by the way use the Stylish add-on to make some readjustment of the UI to better suit your personal needs.
https://userstyles.org/
You would need to know CSS, but hiding certain elements or making the UI more compact would be totally possible with Stylish.
You can find various styles that people already created for Tatoeba in the Stylish website, by searching "tatoeba":
https://userstyles.org/styles/b...eba&type=false
> You can by the way use the Stylish add-on to make some readjustment of the UI to better suit your personal needs.
Thanks, that actually sounds cool. CK also mentioned Stylish to me today, so I guess I'll give it a whirl.
You may be interested to check the CSS code that I posted over here:
https://gist.github.com/trang/0...33407ccd794716
Related comment on GitHub:
https://github.com/Tatoeba/tato...ment-636214078
Just to be clear, I was not asking you those questions in defense for keeping the new design as it is. We will always try our best to provide the best possible user experience for everyone. But to do so, we need to understand how people use Tatoeba.
Users tend to go straight off saying "I would like things to be this way" or "Is is possible to implement this", but that is not helpful to us. We don't make changes just because someone prefers things to be a certain way.
Unless you are a trained UX/UI designer, you will not necessarily think of the best possible solution for a UI problem. You will suggest solutions based on your own habits and your own biases and the things you suggest may have a negative impacts on others without you realizing it.
The best thing you can do to help us provide you, and everyone, the best possible user experience, is to describe how you use Tatoeba, a little like you are telling a story. You can always provide a solution, but without the story behind it, it doesn't help us figure out the best course of action. That's what my questions were aimed to do: to extract this story.
Now more concretely, you shared this bit of story:
> I've stumbled upon some real gems there a few times, something
> I would have never found if I was trying to search for something.
> I would occasionally translate them. Now it works super weird - I add
> 3 Ukrainian translations out of habit, only to realize 20 seconds later
> it already had Ukrainian translations, they were just hidden.
So this is useful information to me. It is a lot more useful than hearing you say you want to be able to have all translations expanded by default.
It explains why you prefer to have the translations expanded and it also explains why you prefer to have a more compact design. And it helps me understand where precisely are the problems.
Now what I can say is that having the translations expanded by default isn't great if you have tons of translations. The more translations, the more effort for you to spot the existing Ukrainian translations.
For that specific problem that you described, a better solution, in my opinion, would be that whenever you select the translation language, it displays the existing translations in that language.
Having all translations expanded does solve that problem but it is not a solution that will scale in the long run. It works only because at the moment, most sentences don't have more than 10 translations.
But one day, if Tatoeba continues to grow (which I hope it will), we may have 100+ translations on the majority of the sentences, and displaying them all is not exactly usable. There's still time until we get to that point, which means if expanding all translations is the cheapest solution, we can still go for that solution. But that shouldn't stop us from trying to figure out better ones.
For instance we could make an option to allow users to choose how many translations are displayed by default. We chose 5 arbitrarily, but some people like you may need this limit to be higher, perhaps up to 20 or 50 even. We have a similar option for the number of sentences per page (we chose to have 10 by default, but users are able to increase it up to 100). I think that would be longer-term solution than keeping the "expand all translations" option.
Bug :-(
After entering and submitting a search string, on the results page that follows, the search string is missing in the search box.
Where do you enter the search string?
I cannot reproduce that neither using the search bar nor the advanced search form.
Looks like it has been fixed. It's not happening anymore. I was using Chrome browser on my iPhone 6. It's the same search box where you would type in the text string that you want to search. When the results are displayed, that search box should retain the search string you had typed, but it was missing. Anyway, it seems that it is no longer exhibiting this behavior.
I really do think that we should make the tag facility of Tatoeba more efficient and a better place to be.
Before even daring to ask our developers for some structuring of the current system, only a few simple considerations could drastically reduce the amount of existing tags and improve the conveyance of their intent and thereby the visibility and accessibility of every single one.
————————————————
Sorted by Count:
The first 64 of 153 pages of tags are tags with one occurrence only. I think many of them are not eligible to be treated as tags but should rather be a comment on the sentence.
A few days of review and sorting out could go a long way towards a major improvement.
————————————————
Sorted by Name:
Looking at the 5 pages of tags that don‘t start with a letter of the alphabet the following observations seem to be worth mentioning:
The 18 tags starting with # are uniformly monopolized by ‚Almanac de Interlingua‘ (itself a tag - with 410 occurrences) and do only occur in tandem with it.
Because of typos (2104 instead of 2014), (missing) empty spaces and differences like ‚Dec. vs Dec‘ - 5 out of the 18 are only attached to 1 sentence and are redundant, what leaves us with 13 #-tags in total.
Even if the ‚Almanac de Interlingua‘-tag with 410 occurrences makes sense, the 13 #-tags (with Issue number and Release date only) don‘t. They cannot be reused for any other application and should explicitly be prefixed with ‚Almanac...‘!
‚Almanac de Interlingua - # 1 - Oct. 2009‘
The tag ‚Almanac de Ainterlingua’ has only 1 occurence and seems to be a typo.
————————————————
* “man“ as person (3)
* “man“ meaning person (1)
* “man“ meaning human (1)
looks like being candidates for one single tag. About starting with quotation marks -> see System-Tags below...
————————————————
* „that that“ (1)
* that that (47)
search for ‚that that‘ sentences result is several thousands which could be tagged by someone ;-)
————————————————
* s-a (1)
* ‘-‘ it is used for linking two different parts of (1)
* ‘s‘ = reflexive pronoun (1)
do not make sense as tag (no language information etc.) and are only used in one common sentence
https://tatoeba.org/eng/sentences/show/5130610
————————————————
* (Appius Claudius Caecus)
occurs in 2 sentences and should be conformed to hundreds of similar references made with
* by Appius Claudius Caecus
————————————————
All Esperanto grammar tags used for denoting tense/mode etc. with groups of
* -as
* -is
* -os-
* -us
* -u
* -i
* -nte
should be prefixed with Esperanto or a language code
————————————————
Numbers like
* 1762
* 1790
* 1816
* 4072
make no sense without context and are always used in conjunction with other tags.
They should either be incorporated in the contextual tag or explicitly and uniformly be extended with
* 1762 (year)
* 1762 a.d.
or what ever the number tries to convey.
Some ‚body’ has to decide which format is recommended for years e.g.
————————————————
* 1/1
* 1/2/1
* 1/2/3
* 2/1/2
etc.
are all monopolized by Esperanto and have no obvious meaning as to be eligible for being tags
Not even a prefixed country code would make the intent of these random numbers any clearer.
————————————————
What is the ‚official‘ language of tags?
* 2 sentences (1159)
* 2 orações (2)
————————————————
* 3 Henry VI (1)
What is the intent of the prefixed 3?
And it should be ‚from Henry VI‘ - just like all other citations are marked uniformly
————————————————
Tag ‚?‘ has one occurrence ;-)))
————————————————
SYSTEM TAGS:
There are 40 @-tags at the time of writing this.
I understand that the @-tag should be reserved for System information.
Maybe user tags should be restricted to starting with numbers, letters and maybe certain „“‘ punctuation only.
Tags like @NNC, @Japanese puns, @Islamophobic etc. don‘t match these requirements.
@check translation
@check A-to-B translation
is not necessary as a tag (endless number of combinations) and should for individual needs be done with a comment anyways and a List for collaboration.
etc., etc., etc.
————————————————
The set of @-tags or other potential System tags like a #-tag etc. should be maintained by a very restricted group of admins to guarantee their small number and unified use because of recommended applications.
————————————————
There are still a number of open tag-related issues on GitHub.
https://github.com/Tatoeba/tatoeba2/issues/961
Allow admins/corpus maintainers to merge tags (2015)
https://github.com/Tatoeba/tatoeba2/issues/333
Add categories for tags (2014)
https://github.com/Tatoeba/tatoeba2/issues/330
Add a way to delete tags from the set of possible tags through the UI (2014)
https://github.com/Tatoeba/tatoeba2/issues/305
Prevent users from adding unnecessary new tags (2014)
Search "tag" https://github.com/Tatoeba/tato...+is%3Aopen+tag
You have obviously spent a long time considering this problem, so perhaps you should open several issues on GitHub, explaining the problems and perhaps suggesting several approaches to solving the problems.
Well, almost all of the above mentioned problems could be solved in 20 minutes by someone with the permission to rename (for typos) and delete tags (duplicates). I have them already properly laid out and justified.
Others that require some knowledge of policies or decision-making (how to extend year numbers ‚1848 a.d.‘ or ‚1848 [year]‘ in a multi-cultural environment) could follow suit.
The wrong audio recordings, hinted to with tags like
• incorrect audio (2)
• wrong audio (2)
are not seen by the right people with the responsibility to change such cases and remain in the corpus since as far back as 2014.
So I have already addressed the audio-tag issues by providing comments and the official @-tags to all sentences. So they are just waiting for some admin to delete the wrong and redundant tags to be left with a clean set of 3 remaining ‚official‘ tags.
https://tatoeba.org/eng/wall/sh...#message_35382
The Wall is a shared space. It is not the place for one person to post daily lengthy bug reports or enhancement requests. You have now had two people (CK and me) suggest to you that you open these as issues on GitHub. It's not just a suggestion anymore.
By the way, telling our busy developers and admins (or any other members of the community) the order in which they should tackle issues, or how long it should take to fix them, is not the way to get them to the front of the queue, or to make friends.
Please repost this at GitHub. This thread is now closed.
Searching the tags for ‚audio‘ I get the following result
• has audio (559)
• Audio (1)
• by Cl-audio Magris (1)
• incorrect audio (2)
• wrong audio (2)
in addition to the officially supported
• @change audio
• @move audio
• @wrong audio
I have added comments and official @-tags to all entries, except for the (559) ‚has audio‘ tags, but I think they are redundant anyways these days.
So I think a nice clean-up of the unnecessary tags by somebody with permission to delete tags could be performed.