Hey LingQ! -- Fix Bugs, Stop Features, Users don't want to be Beta Testers

Open Forum
1

Recently I reupped for another year of LingQ.

I’ve noticed from my own experience and comments on this bulletin board, that LingQ is riddled with bugs and poor design choices, but nothing much changes aside from adding more features, not necessarily ones people are asking for, and breaking older features by releasing poorly tested code.

I’m a retired professional programmer from the Bay Area / Silicon Valley. I’ve never seen anything like this aside from amateur freeware. LingQ seems to be an out-of-control project from top to bottom. I suspect the codebase needs a major rewrite too.

How about it?

13 Likes
2

It’s well said, though I suspect they’re never gonna become accountable to their users. They act like they don’t give a thing for what would be right. Untill you have an alternative. they’re gonna abuse you like badly. Rise the price, break users learning flow regularly. Like, hey, you said something? Zoran, pal, tell them that we will look into it.

image

12 Likes
3

Indeed.

However, alternatives are on the horizon. For instance, Lute (Learning Using Texts):

Lute is a free, open-source PHP-Apache-MySQL project that is a complete rewrite of the core features of LWT and is basically a stripped-down version of Lingq, which is the company headed by the great polyglot Steve Kaufmann.
–Summarizer

I haven’t tried Lute and maybe it isn’t a LingQ killer. But it won’t be long for one to appear, especially one with AI built-in from the ground-up.

6 Likes
4

I’m a retired software engineer. I’ve seen very successful software projects with thousands of reported bugs. I’ve never seen bug-free software – it is either a myth, or it happened on a tiny (one-man?) project.

Is LingQ terrible, compared to perfection? Sure. So is every other large software project I’ve ever seen. That is not a fair evaluation. It is not useful to anyone.

“Poor design choices” is pure nonsense to me. “Best design methods” are a fad: they change every few years. More importantly, different tasks require different designs. How many LingQ-like systems (supporting 47 languages) have you developed?

A complete rewrite is probably 4 or 5 man-years of work, for skilled experts. It is possible (hire a new team to do that for a few years, while the existing team continues to support existing customers, then somehow integrate the new software with the old customer data) but it would increase bugs for a few years. You don’t expect the new code to be bug-free, do you?

I might agree with you about fixing bugs (if possible) versus adding new features. I don’t have enough experience with LingQ to have an opinion on that. So far, I’ve only found bugs in the language modules (supported by other people).

5 Likes
5

Who said anything about perfection?

Straw man.

I would like, for instance, keyboard shortcuts to work reliably. They don’t. They didn’t when I started LingQ a year ago and they still don’t.

Keyboard shortcuts are junior programmer work. Unless the codebase is so screwed that no one can figure out the flow of control.

5 Likes
6

As an experienced software engineer, you should stay here for awhile to judge competently what is happening here. Now you’re just talking out of solidarity and hypothetically. Making “supporting 42 languages” an excuse for doing it poorly, really?
MVP? Nah…
Man, I actually have a screenshot where a support guy (who had been working for LingQ for like 7 or more years already) asked a user for the password to the user’s acc in order to figure out the problem, specific to the user’s device.

image
image820×277 110 KB

Then they unilaterally deleted the thread the user had created.
Nobody’s perfect, but this many years is enough to learn the basics at least. But only if you’re intended to improve, which seems like isn’t the case. Mind you, LingQ 4.0 was better than this version. The only somehow reliably working features introduced to the system (when the system itself works), like AI simplifying of texts and Whisper thing, were developed by other developers as you understand. But even the core of the system fells apart every two weeks or a month.

But, if you think it’s ok to pay the highest price among LL apps to be a beta tester and to not being able to use it like 1/5 to 1/3 of the time because of the crushes and bugs, then I’ve nothing to add, we’re just too different I guess.

3 Likes
7

I’m surprised to see you’re experiencing that many issues…

This doesn’t reflect my experience at all. I use it everyday, at least 30 minutes, and sometimes for over an hour and I have never encountered a problem that makes it unusable in my 3+ years using it. I experience more minor errors, like when trying to highlight a phrase to link, sometimes I have to click away and re-highlight to get it to work.

With Japanese, I’d understand for any one who wasn’t advanced it seems like it would be extremely frustrating due to text parsing and erroneous pronunciation. But because I have enough Japanese to recognize those as errors, it isn’t an issue and I find it a great tool to help me brush up on reading.

I don’t use it with YouTube, or Netflix, etc so I don’t know about that experience.

3 Likes
9

Another retired professional software engineer here … :rofl:

I have made far more progress using LingQ than with any other app, I can now understand native French podcasts, for example on the lack of diversity in French cheese moulds, and I am close to understanding French films. And I am using it for German, albeit at a much more basic level. I am impressed by how well the idea works, so I am loathe to be critical.

But honestly the standard is poor.

Critical bug: For a week recently the app crashed as soon as I opened it, making it unusable. Fortunately I discovered a solution which was to reinstall it. I’ve never had that with any other app. Others reported it, and I suspect this is due to lack of regression testing. Some users might not realise that they have to reinstall the app, so for them the app becomes unusable until an update with a fix appears. This level of bug should never get out into the wild. I’ve never worked on. a software application that does this, and I worked in the industry for 30 years.

Serious bug: Today after one hour revision of a YouTube video, my German known word count decreased by 10 words, and my coin count increased by 5. The statistics don’t work. In short, unusable.

Serious bug: When I watch YouTube videos, they regularly shoot back to the beginning and I have to spend a minute relocating my current position. For an 80 minute film that is not pleasant. I often end up swearing out loud out of frustration and annoyance. Surely this isn’t hard to solve?

Serious bug: When watching a YouTube film, more often than not the current line scrolls off the screen and hence I cannot read the transcript while listening. This is so easy to solve. Just place the current line in the middle of the screen, not at the top.

Quibbles: I don’t like the content in LingQ, some is poor quality, and I suspect some YouTubers see it as a way to promote their products. Some has only artificially generated audio. Some is atrocious. But this isn’t serious, I just import good content.

So for me LingQ is a reader tool only, which I use with YouTube imports. £100 a year for a reader is in a sense a lot of money, especially when compared to Babbel et al. However, I pay the money because it is far superior to Babbel et al. Admittedly it took me six months to figure out how best to use it.

But, if another reader appears on the market which allows me to turn YouTube videos into lessons, I’ll drop LingQ.

It makes me wonder what the problem is. Perhaps they try to support too many languages, and spend time searching for content to put into LingQ. Perhaps they don’t do any regression testing. Perhaps they do inadequate testing. Perhaps the income is too small to employ sufficient engineers. Perhaps the code base has become very hard to maintain, a common problem with older software. Or perhaps they just don’t have anyone responsible for UI design. I’ve worked on countless projects, so I’ve seen excellent practices, and dreadful practices. But I don’t know what the problem here is.

9 Likes
10

Sadly it looks like Lute won’t run under iOS. I use an iPad.

11

Well, yeah. I don’t know exactly what the problems are either.

But I do recognize there are serious problems which are not being addressed by adding more features with very little testing, which seems to be all that LingQ is doing.

As I see it, LingQ has a brilliant core concept of supporting comprehensible input via audio and text, but their codebase and practices are poor and LingQ would rather add features than fix and strengthen that which makes LingQ outstanding.

4 Likes
12

Good points, I agree. Now you mention it I’ve seen large changes to the layout of the iOS app, but the serious UI usability bugs remain.

1 Like
13

In my experience, usually “the problem is” that end-users expect zero bugs. Maybe that happens with small 1-man programs, but it doesn’t happen with programs written by many people. How many programs (written by other companies) does LingQ combine? Interfacing always causes problems.

Is there a better program that does what LingQ does? If not, then LingQ is the best! I consider it baloney to compare things with perfection and then say they are “bad”.

I tried using LingQ for studying Chinese, and it wasn’t for me, so I stopped. Instead I use Language Reactor and Youtube videos. When I started studying Turkish last fall, I found LingQ useful for Turkish, so I use it.

Please read the book “the mythical man-month”, published in 1970. It explains why adding more software engineers ALWAYS makes a project take longer. Maybe that isn’t intuitive to you, but it’s how software works.

2 Likes
14

Please read the book The Mythical Man Month

I’ve worked in software for 35+ years, in many companies, on many projects, good and bad. I read the book ~30 years ago. It is very outdated.

In general the issues apply to large projects, not small ones. Engineers have developed methodologies to mitigate the issues highlighted in that book. For example, Service Oriented Software and Software As A Service. I’ve worked on old style projects, and ones managed with modern methodologies, and the difference is plain. I’ve worked on a project where fixing one bug created another bug, because the UI and the software were so impenetrable. On that same project, one engineer would create countless bugs when adding a new feature, because he reused an existing database table, rather than create a new one, which would reduce the chances of regression to near zero. It was near impossible to fix code because database tables were impenetrable. That was poor management.

It isn’t just the size of the project thay can cause issues, it is also the age. If you have an old project with a rapid turnover of staff, that can be a recipe for disaster if you don’t manage it well. Compartmentalisation is one way to reduce issues. Refactoring is another way to reduce code spaghetti.

As for the number of staff on a project, too few staff is no better than too many. For example, no test engineers will mean little or no testing of the software. Too few test engineers will mean bugs escaping into the wild. Too few UI design engineers will mean a poor UI design.

3 Likes
15

In my experience usually the problem is that end users expect zero bugs.

The LingQ bugs are not minor, they are not slight deviations from perfection. They are major.

The app crashing whenever you open it, rendering it unusable, is a show stopper. It is the kind of bug that should never appear. I discovered that uninstalling the app, then reinstalling it, solved the issue. I had to do that several times an hour. This bug was present for many days, maybe even a week. There are some non tech savvy users who would not figure out how to workaround that bug.

Last night while watching a YouTube video in LingQ, I was unable to read the text while listening, because the current line had scrolled up, off the screen. That’s no minor bug, That’s an unusable product.

When I see an interesting word or phrase, I like to Google, to find extra information. That is interesting and helps me remember. When I go back to LingQ, it resets the video back to the beginnning. For a film it can take me a minute to relocate my original place, so I can continue. That’s not a minor bug, that’s a very serious usability issue. I routinely end up swearing out loud when using LingQ.

I use LingQ because it is the only app I know that allows me to watch YouTube videos, while reading the transcript, and looking up words. However, whilst it looks very slick - and yes it does look very nice and polished - usability is poor. And they don’t fix many of these major issues. I have a feeling they don’t adequately manage regression testing and product releases, but I don’t work there, so I can’t be sure.

2 Likes
16

That is my suspicion. LingQ is a fairly niche application with a fairly small user base, likely making a fairly low amount of revenue*, based on fairly old software. It’s probably stuck somewhere between being a profitable business and an enthusiasts’ hobby project, hence the difficulty addressing underlying bugs and annoyances and the seemingly unprofessional way they release updates without enough testing to spot critical bugs ahead of time.

To put it another way, I suspect LingQ the company is doing a bit of a balancing act, offering an ambitious niche product that’s hard to make profitable enough to address major issues.

(*We individually pay a decent amount, but I suspect to really pay for the necessary fundamental revisions, the price would need to be quite a bit higher - at which point many of us might cancel, so their revenue might not even grow enough from such a price hike.)

7 Likes
17

You make many good points. I think they are all valid, including the importance of test engineers and thorough testing. A programmer self-tests his code, but that isn’t enough.

I just want to add one comment about testing. In my experience, no matter how much in-house testing is done, customers will find some bugs. Always. There might not be any major problems (this isn’t about LingQ), but there will be some bugs. Why? Because customers always use the thing in ways that nobody at HQ thought of. Not engineers, spec-writers, tech writers, QA engineers…nobody.

2 Likes
18

I suspect LingQ the company is … offering an ambitious niche product that’s hard to make profitable enough to address major issues.

Yes, LingQ is indeed very ambitious. They support Windows, Android, iOS, and a web browser interface , and a huge number of languages. And it’s inherently not easy to get to grips with, and that’s not a criticism.

Applications such as Duolingo, Babbel, Busuu are much simpler to create and use, and must make a good profit.

But as someone else said earlier, LingQ needs to focus on fixing existing serious bugs, not adding new features.

2 Likes
19

I completely agree. I think that’s the part where they’re stuck in the “enthusiasts’ hobby project” that I was talking about: they’re not professional enough to focus where focus is needed. (Sorry guys, I do love you nonetheless :sweat_smile: )

4 Likes
20

Please stop saying this. Some here are or were professional software engineers. Even experienced users know there are always bugs.

Speaking as a retired professional, no company I ever worked for would have allowed the rampant bugs in the basic mouse-keyboard-event-handling in Sentence View to pass QA and into the hands of users.

For example, tell me about any commercial software product you have ever used where the keyboard shortcuts were not reliable.

2 Likes
21

However, alternatives are on the horizon. For instance, Lute (Learning Using Texts):

If you’re going to do it yourself, I recommend LWT. I had it up and running and used it for a while, but quite frankly, despite or maybe because I was in software development for so many years, I did not want to host my own tool. And this is just the reader – it’s not a material repository, it’s not going to host and manage playlists for me, etc. – and I’m not going to pay for online hosting and secure access!

Lingq isn’t perfect, I’ll grant you. It seems the web dev team has some code management issues. The iOS app is more solid. However, as long as Lingq returns as much value as it does, I’m OK with the bumps in the road. Here’s what Lingq offers that any other competitor would also have to offer me:

(1) Cross platform online application (and offline in iOS) – Mac, Windows, Android
(2) Single database serving all my access to it tracking my learning
(3) Playlists available to use wherever I am on whatever platform I happen to have at the time
(4) Virtually unlimited languages supported
(5) Amazing amount of built-in material for study
(6) Creating my own lessons from a virtual unlimited amount of external sources
(7) Audio transcription at no additional charge - I’d go broke paying for this myself

I’m not going to do that myself! It works 98% of the time. Any alternative would have to match what Lingq offers.

9 Likes