I really would like to constantly read these updates and know what you change, implement and how new features work. In the past, it happened to me that I went to the webapp settings to find new features that I didn’t even know were added there.
These notifications I received about timestamps, now I know why.
You have Twitter (X), use it for this, or I don’t know, if you don’t want to receive and deal with comments on your technical updates, use another format, like the Newsletter that you regularly send, but more communication would be very nice and useful.
Imho.
4 Likes
I second wanting to receive updates on new features and updates (no matter the version ios, Web, Android, etc,).
I’d love them just to stay on top of (as best as I can) the LingQ program development.
May I suggest the category Updates, Tips and Known Issues would be the appropriate place in the forum for these?
Note: It looks like this was the place for them in the not-too-distant past.
Thanks!
5 Likes
Can you please solve your problems without touching my lessons?
Whether you guys have a problem with old files, or new files, or files that drop off the internet, solve the problem and stop creating problems for me. Formulated differently, stay away from sneaky, underwater, automatic processes. You guys don’t have the software process implemented to support updates that work for everyone involved. You have no systematic communication, no two way communication and you certainly do not ask whether we like your changes before you implement them. Please stick to explicit actions only: on request of the user. That way, I control what happens to my lessons, not some runaway process that someone, somewhere thinks is a good idea.
In summary. No matter the excuses, never touch my stuff without my explicit consent. Keep sneaky, underwater, automatic tasks away.
I don’t care about files that nobody owns, or if someone requests you do something, like through pressing a regenerate button, but please stop sneaking around, changing stuff that I never asked you to. Return the control to the user, don’t take control over something you know nothing about.
P.S. Another thought, if you want to fix piping in the basement, you don’t rebuild the whole house because its convenient for you. I.e. solve the problem at hand, don’t involve the rest of the world in your problems.
6 Likes
I have been struggling with the aftereffects of this update since Friday when I found myself locked out of Chinese Traditional. Fortunately this got quickly fixed. But for me two concerns remain:
- I still can’t access many of my lessons
- According to the notifications I have been receiving, hundreds of my lessons have gotten their timestamps newly generated. Since I cannot possibly check them all, how do I know that the timestamps actually improved? Were manually adjusted timestamps bulldozed and replaced by computer generated ones?
The Chinese expression “帮倒忙”, appears fitting here. While I appreciate the attention to my lessons and I’m sure you had the best of intentions, I didn’t request this being done and would’ve preferred to opt-out, tbh.
2 Likes
Its quite obvious what has happened.
- LinGQ introduced a new feature to regenerate the timestamps of files that had ‘Not available’ ‘NaN’ Timestamps.
- The code had unexpected consequences and wasn’t scoped enough to complete the task effectively.
- They ended up regenerating lessons with languages that can’t be time-stamped and also lessons with manually set time-stamps.
- These lessons that can’t be time-stamped are stuck in a death loop of server-side time stamping.
You have to wonder why all of these new features are tested on the live server. Especially when the problem is so egregious and this has only happened because of poor implementation.
15 Likes
I’ve got a youtube video I imported that seems to be stuck in a forever-timestamp loop. It’s been in that state since yesterday. I can’t edit the lesson or add a translation to it because of the timestamping job. Granted, the video is music with heavily distorted audio, so it’s probably giving their Japanese algorithm a heart attack. Just wish I could cancel the job and do it manually.
3 Likes
I’ll push a patch to Lesson Editor Premium tonight that lets you open it from the library page by lessonId. Should be able to cancel their timestamping process by forcing a new timestamp or manually updating them. Edit: This patch is now ready if you want to see if you can fix that song
3 Likes
Thanks! I’ll check it out.
1 Like
I completely agree with your request: “Never touch my stuff without explicit consent”. I just want to clarify what did you mean by saying “my stuff”. Did you mean the content that you created using components like an audio, transcript etc and explicitely made this new content unit/lesson “private”? If some automated background processes were run on the live servers to “improve” something in privately owned content units without the owner’s consent this is a disaster for all of the ordinary users of LingQ system. We know that developers/admins of the system always have some backdoor to assign themselves whatever access privileges they need for modifying any content but normally this is kept as an emergency measure.
But if this is the case in everyday practice then it defeats the point of creating any content by the students themselves. I would never spend a significant time by creating content for my personal studies when I know that it can be destroyed any moment without my explicit permission. Sharing new content by the owner is another issue with its own caveats and requiring very careful design considerations.
So can you please confirm that the content owned by you and explicitly marked by you as “private” was being ruined without your consent?
2 Likes
Well, to answer your question, yes it happens both for shared and for private content. Stuff in my sentence is referring to content. When you create a lesson (private or not), immediately after you add audio as soon as you start view lesson i.e. working on the lesson, a background process starts generating time-stamps and lets you know through a short notification. So the trigger is the first time you open after having added audio to a text that did not have time-stamps. In other words, import a subtitles file (srt) does not get regenerated.
I have learnt to deal with it as follows. I add the audio after having assured the lesson text is ok. Then I add the audio and open the lesson. I get the notification of time stamp generation and wait for the notification of being finished. In order to get the notification, I have to leave the lesson and later re-enter the lesson.
After that I can change the time-stamps to my liking. As far as I know, after the first run, they will not be overwriting. I am not overly sure though.
As for the main message of my post being “give users control over what happens” clearly fell on deaf ears. Lingq has no intention of handing control to users.
As long as the time-stamp generation is a one-time thing and after that Lingq refrains from changing my work, I have less of a problem. If for any reason, Lingq starts overwriting my changes again, as far as I am concerned Lingq is handling out of pure malicious, control-hungy greed.
The same goes for the idiotic saving during a “edit lesson” session, where every change is saved every time, but it waits till you want to do something, after which it annuls what you wanted to do, and finally starts saving the last change. Up until now, Lingq has been unwilling to hand control of saving to the users. Like letting me do ctrl-s to save when I deem it useful, not every change. Or letting me determine how often it gets saved and when.
So, to summarize, it is not as bad as it used to be, where the background process was out of control and kept changing content back to what it was, but it is certainly not good. The time-stamps warrant a lot of correction to be useful in sentence mode. The explicit button “generate time-stamps” suggests user control, but there is none. It will run irrespective of user wishes. I am not even sure what happens if you use that button. In the past it would overwrite changes long after finishing.
In conclusion, not worth a war on the forum in my opinion. Having said that, if anytime soon a competitor offers something equivalent I will have no issues switching to the competitor. Lingq is getting very complacent.
Hope this helps.
3 Likes
I agree with a lot of what has been said in this discussion. I am also VERY disappointed with LingQ.
When I contact support, sometimes I receive a reply, mostly not. When I do receive a reply, it’s just to say ‘we will look into it’. Which to me means ‘get lost’.
The idea of LingQ is great. That’s why I have spent money to try it out for a year. But the program is not very professional in my opinion. And they do NOT listen to the users.
I am also reluctant to continue with this when my current subscription ends.
5 Likes
Thank you for your detailed response.
-
It is good to learn about your methods of protecting your work from damages and thank you for that. But at the same time I don’t think that constantly looking for counter-measures against developers “deep power” might work in the long run. If LingQ developers do not really cooperate with the customers they will always come up with some new ways to destroy what we are doing. Not intentionally, of course, but just by not taking us into account in their mindless experimenting with software and enjoying their full control over the system. No, we should be on the same side to succeed.
-
I can easily see at LingQ well-known syndrome “Developers know better what the customers want”. This approach was a cancer for many companies in Silicon Valley and around the world when designers/developers created software based on their own arrogant views on what is wanted, needed and required. This cartoon was true many decades ago and its still true today:
-
I see that the software engineering process in LingQ is very far from being a healthy one. Instead of publicly discussing new proposed features with the potential users at the early phases of new development, finding the weak points and flaws in proposed solutions, clarifying what a “must have” functionality should be, LingQ (even with outsourcing) spends very significant resources/money on implementation of functionality without having a clue how people are going to use it in real life. The concept of “Use cases and their analysis” seems to be Greek to LingQ Product Managers. This attitude must change or LingQ doesn’t have many chances to survive in a foreseeable future.
-
I don’t look at the discussions like this one as being “a war on the Forum”. I think the realistic purpose might be to make the owners of the company, Steve and Mark, at least to become aware of what is going on and to realize that this is not acceptable. If they really care about LingQ they can try to do something about it. If they don’t care, then you are right and maybe it’s time to look for a good competitor. I’m starting to see some publication about building language learning systems based on ChatGPT and other promising technologies. But the key point is not technologies per se but having really good, well-educated software engineering team, real professionals (product managers, developers, designers, architects, QA and Tech Support) who beleive in solid good software engineering practices and who know how to use them.
-
But for now I would suggest to keep talking and to involve into this discussion more and more current users of LingQ. By reading comments on this Forum I see that there are people here who have an appropriate experience in computer-related areas. I would also suggest that each of us would get in touch with LingQ owners by messages or some other way to draw their attention to this situation.
After all, LingQ - is a brilliant idea and it would be a waste to loose it due to everything mentioned above.
Cheers,
Ted.
5 Likes
I certainly think better communication wouldn’t go amiss. Most of the time, we are kept in the dark as to what is going on. Nothing is ever said about upcoming updates and new features.
I also get the clear impression the team doesn’t really listen to users’ input and will only improve or change what they see fit, with the rest falling by the wayside.
5 Likes
I can just assume that if all the members of design/development team would be actual users of LingQ as students learning some language we would probably see today much different and much better LingQ system. That’s the problem. The system is being designed by the people that probably don’t use it in their everyday life. But we do.
4 Likes
As a retired professional programmer, that’s my take. LingQ is a brilliant conception, but the crummiest software I’ve ever paid money for.
Everyday I use LingQ I find the keyboard shortcuts don’t work in unpredictable ways. Nor do bookmarks. This is Programming 101 stuff.
Initially I tried some of the more advanced features like timestamps or even setting up tutor appointments and it was so painful, I gave up.
My guess is that the codebase needs a major rewrite and something, probably several things, are out of whack with management and the programming team.
3 Likes
Without looking at the current codebase its hard to say what exactly needs to be done. It could be anything from a complete re-design and re-coding to maybe just a reasonable restructuring and refactoring if the the system core has been designed and coded properly. Actually it might be the case because there is a chance that the LingQ system has been created in several layers by several engineering teams in the last 15-20 years that LingQ exists.
But what is absolutely clear to me that at this point LingQ Engineering needs very strong, knowledgeable and experienced leadership that will be able to put things in order, to implement some strict and disciplined engineering processes in product management, design, coding, testing and release processes. That’s hard thing to do, its painful and pretty expensive. But continuing wild-west-style hacking around will kill the whole thing pretty soon.
1 Like
I don’t see that LingQ owns anything special here. Display a sentence, connect words and LingQs to a database and dictionaries plus TTS for audio. Add user incentives like streaks and challenges.
If a serious software company thought it was in their interest, they could steal LingQ’s market.
However, at this point I suspect AI tutors coming up are the real threat.
3 Likes
I completely agree. There is nothing special here and actually it is absolutely not necessary to have some special functionality or patentable ideas to create a successful commercial software product. What is really challenging is to create a reliable, robust product with well thought-over functionality that its easy to evolve by adding new functions on the top of the solid foundation.
But the most of the challenge would be to create highly usable product with a lot of time spent on user-experience design and interaction design. Those serious software companies of even good startups with the right people who are ready to invest time and money in implementation of LingQ-type functionality and who will do it in the right professional manner from the very beginning will be the winners.
4 Likes
My guess is that at some point in the past they’ve build pretty decent LingQ prototype and got their initial financing etc. But after that instead of building a commercial production system from ground zero as it is supposed to be done, they started to enhance the prototype by adding bits and pieces of functionality to this prototype. And now, many years later we use this pretty awkward system. But there is a huge difference between prototype/demo and solid production system. And it requires solid software engineering experience to see it and deal with it. A lot of startups died after getting into this trap.
6 Likes
That’s my impression. It’s a lot more fun to add features than to fix spaghetti code.
I am boggled that LingQ can’t make keyboard shortcuts or bookmarks work reliably. Aside from the elementary nature of such bugs, they also degrade the basic everyday user experience of LingQ.
2 Likes