Beta - don’t get rid of typing tests

OHMYGOSH don’t get rid of the typing tests! I only just realized that the new deck that I made and recently reset my progress did not have typing tests, despite my having checked that the typing box was ticked. I like the new version of memorize- but I need to have typing tests.

Can anyone of the Memrise-Team write here if they still want to get rid of the typing test or not? It drives me a little crazy, cause i have such big courses here and a lot of my friends already switched to ANKI and other apps, due to that. Thanks a million!

We do not, and have never wanted to get rid of typing tests.
Understandable confusion occurred when the web beta launched because it was using the proportion of typing tests that the mobile apps used because of a shared software library. When we realised that we’d inadvertently produced that effect we updated the website beta so it has a much higher proportion of typing tests compared to the mobile apps.


Thanks James. I would still like to have an “All Typing” option, though. Multiple Choice and Tapping tests are helpful during the initial learning stage and for Difficult Words but I find that typing tests are the best way to reinforce my learning.


Honestly, I do not understand as to why you refuse to let users decide how they learn. It doesn’t make any sense to utilize an automatism to predict what would be best - typing, cards, etc. - it doesn’t work as it is today (for instance, reviewing a level will always produce the exact same order of question types on Android). And you actually went ahead and incorporated an external library to do the math for you? Why go all the extra steps for something that makes little to no sense in the first place?
Why do you think this is was or is a good idea, can you share any insight? I’d really be interested in how you get to all the decisions that you make and for which you receive quite more than just a bit criticism here in the forums. Also, you never really respond to any proposals made by the community in a way that we’d understand why you decide differently.

FWIW, it makes even less sense to implement different approaches (or a different logic) across platforms - why would anyone see this as a good decision really beats me.

1 Like

Hi Olaf,

Sorry I think I was unclear in my comment and gave you the wrong impression there, let me attempt to be clearer.

It’s not that we’re refusing to let users decide how they learn, it’s that we have to prioritise what we can focus on, the current focus is rebuilding the technology on all 3 of our apps so it loads quicker, is more reliable and a modern software framework will let us build much faster and focus on improving the service rather than working around the technical debt which restricts improvements we try to make currently.

Using shared software libraries is key to this, all three apps do things in common, such as fetch a list of items for a lesson from the database, so rather than have iOS, Android and Web build their own code, we build code once into a library so they can all use it. That’s what I meant by shared library. The issue you mention on level review is not part of the shared libraries, that is legacy code.

Despite the shared software library we do still alter the default experience between apps where it makes sense; the reason the website has more typing tests than the mobile is because typing tests are a lot more pleasant if you have a physical keyboard.

That’s not to say we’re not making improvements while working on the technical rebuilds, it’s just that some areas are more difficult to work on than others until the rebuild is complete, with the learning sessions being the most difficult. Right now the team I’m in is working on a new feature on Android that should be with you in a fortnight. I can’t say much about it in this post but as a user it’s something I’ve long wanted and I hope you’ll all like.

We don’t have time to respond to every comment and question, but everything is read and collated with feedback from other sources. In terms of how we make decisions, our primary focus is our mission to help as many people as we can speak a language, the secondary focus is to make enough revenue to be able to do that. We can’t explain every decision we make, but they are always made with these goals in mind.


The problem with Memrise is that many of your improvements are in fact are anti-improvements. You are saying you don’t have resources for this, don’t have time for that, and then do something absolutely useless or destroy something fun and helpful. One example, how does this glitter help anyone speak better? As Olaf said in the other thread it looks as if you are not using Memrise apps yourselves.

1 Like

First of all: thanks for your clarifications. The “shared library” sounded like a 3rd party piece of software.

This is again an assumption that you shouldn’t make - you don’t know how your users’ devices are equipped or what their personal preferences are - some may be using a tablet with an external keyboard, for some it might be absolutely fine to use an on-screen-keyboard. Thus, your assumption/decisions often provide an obstacle for their learning process rather than anything else.

(Speed) Reviews for difficult words would be what I would like to see above anything else! :slight_smile:

Without more detailI can’t be 100% sure what you’re referring to, but I suspect it would have occurred as part of the technical rebuild I mentioned where on every app we’re building the learning sessions in new technologies. As part of that the visual experience has changed on some elements, this was not the goal of the rebuild (which was load speed, reliability and making easier to build in future). But as part of the rebuild we have to choose how it looks, and because it’s all coded from scratch it’s no more effort to change it slightly in places where we thought there could be improvements.
To make the rebuilds as fast as we can there are certain elements we may have chosen not to rebuild if they’re not used by many people or serve our key goal of helping people learn languages. Sometimes we get that balance wrong and underestimate how important an element is for users so don’t have it in the rebuild initially, and we appreciate the feedback on this forum and the surveys that explain why something is valuable top you. I hope from the web beta learning sessions and dashboard you’ve seen how many suggestions we have taken on board.


Regarding the typing tests, we have quite good evidence for users on web having a stronger preference typing tests compared to mobile users. In app analytics lets us know how many users disable typing in settings, how many use the keyboard button, how many typing tests are skipped etc. And we get many messages telling us how much a user hates typing tests for being ‘boring’ so we try and balance this against the more serious users who appreciate the value of typing.
Of course this is just the ‘average’ user of web and mobile apps, and individuals have different preferences for which we want to allow greater choice in the experience where practical. The rebuilt learning sessions will make such things easier to build.
And thank you for your suggestion on speed review for difficult words. Relatedly I hope you’ve found the ‘mark as’ functionality on the level details screen we added a couple of weeks ago easier to mark multiple words as difficult or to mark them as already known (ignore)?


It’s actually funny that you mention this. I used to use the keyboard all the time. That is, until someone of you guys decided to get rid of storing the “last used means to enter letters”. Back then, I only (on Android) tapped the keyboard icon once which brought up the keyboard and entered my text. When the next question was presented, the keyboard came up automatically. This is no longer the case, which means I have to tap the keyboard button for each and every question. Considering that it takes 1-2s until the keyboard appears, it’s quicker to just type using the set of letters provided.

I wrote about that several times, I even offered to pay (at least for a part of) the development costs! The only means for users to work on DWs is to repeat them several times each in groups of four. I’d rather (speed-) review through all of them until they stick. It’s a personal preference, of course, but I do think “repetition is king”, as often as possible.

Yes, I did of course. It do think it’s actually quite confusing that I can now flag words as “learned” on the level view by tapping the button on the top right, whereas it is “ignoring” words everywhere else. I’d say that “ignore” is the far better denomination of the feature!

In Android app when you have to compose a sentence you have added a glitter/glowing for the chosen elements. I’ve described it my post.

I don’t see what it has to do with “new technologies” and “improvements”. It’s so senseless and distracting I stopped updating your app.

Why don’t you do it properly in the first place? Users are doing half of your work.

Yes, it’s confusing. You could expect that learned words will appear in review sessions, but you don’t have to learn them.

1 Like

Can I check what the translation is in your native tongues, in English we were going with ‘mark as known’, rather than ‘mark as learnt’, is not that coming through in the translations? The intention being if you ‘know it’ then you wouldn’t want to review it. But if it’s saying mark as ‘learnt’ then I can understand the confusion of you expecting it to fully grow a word rather than ignore.
We’re moving away from ‘ignore’ as our testing showed people found it intimidating whereas ‘mark as known’ people were more confident in using it and we use that terminology already on iOS so wanted to use it on Android and then on web. More widely we’re trying to make it easier to mark as difficult and known, of which adding to level details was one step.
We’re also going to make the way to mark as ignore and difficult from the presentation cards easier, rather than the fiddly triple dot menu.

My point on new technologies and visual design was that if we have to recode the entire learning sessions, then we can implement whatever design we want, without going to incremental effort. In an oversimplified example if there’s a box on a page that used to be blue but we think it would be better red then it’s just as easy to make it red when recoding it. There is no extra effort going into sparkles.

To your question on alphas and using feedback, an alpha in the software industry by definition is always limited, it’s a good way to find out what people value on top of the core experience, rather than spending time building something you don’t know if anyone will use. That’s why the alphas are opt in and for a limited subset of users.

Olaf to your point on keyboard persistence this has been fixed in the rebuilt learning sessions so when they go live i hope you enjoy. Sorry we can’t get it done in the current ones.


It was in Russian, I don’t remember how exactly it was translated, but my thought was that it’s not the Ignore feature, but something new.

Oh my God… Who is intimidated? The same testing bunch of folks who approved Ziggy? How can anyone be intimidated by a short description of a feature? It’s so absurd. “Ignore” is a fine and straightforward word.

Honestly, it’s good that you are trying to explain your decisions, but it all looks as if Memrise as a team has no ideas how to really make this product better. As you can see, you have already created confusion with this small detail, and what about big things?

Is this app for learning languages or for enjoying sparkles and glitter? Do you need to see visial effects while learning? Is it distracting and straining your eyes and attention or something helpful? To put it simply, do you need a hammer with sequins? To me it looks redundant.

Yep, I understand, but mostly users were mentioning things which are lying on the surface. You could have thought about them without anybody’s help. Just an opinion.

I am on German, there it is “Bekannt” instead of “ignorieren” (“known” instead of “ignore”).

That is completely incorrect. Why would you even care to think about changing “ignore” (which is the perfect term whereas “learned” can mean several things, particulary “I already know it but I still want to review it”) and then only change it in one place? That is so typical, don’t you see this yourself?

Great, thanks!

I seem to have been under the wrong impression that this would be fixed on Android. IOW, I was thinking the keyboard would actually be toggled again by tapping the keyboard-button rather than having to be activated after each question. But nothing changed in the latest Android version.

Re-reading this now I notice that you probably meant something else.
So I now I ask myself, what are “the rebuilt learning sessions”? Does this even have to do with the Android app?

Hi Olaf, as I mentioned somewhere above, we are completely rebuilding all the learning sessions in a new technology on all apps, web is pretty much done as you can see in the beta and Android is underway. They’re offering big improvements in loading speed and being in a modern framework will make future changes much easier. It will be a few weeks until the Android ones are released to all users I’m afraid


Why Memrise is doing so many things on the quiet? You could have made release notes with description of this renaming, illuminate users on the changes. I see that other developers are happy to make announcements about new features, but you prefer to do it in the dark. Why? There’s nothing to tell about?

Without typing there is no reason to use memrise