Parley[1] is mainly a vocabulary trainer but can also be used to learn and train other types of information. It is part of the KDE Edu applications[2].

Parley has always been a good application and is quite advanced in its niche. It has 8 different practice modes, such as flashcards, written practice, mixed letters, etc. But it hasn't really lived up to its potential for different reasons and we have now started a pretty ambitious project to make it a modern and attractive application, totally focused on learning. Previously it was a mix of learning and testing and the main UI was more tailored to keeping track of files than the actual learning process.

And this is where you, the designers and artists, come in. We want to make Parley an awesome application in several ways:

• It should be visually pleasing to look at.
  
• It should facilitate the learning by making it easy to see where the student is and what should be done next. We want to create a learning dashboard where you can see which vocabularies s/he is studying and make it alluring to do "just one more session".
• We want to create a library of word lists to study and make it easy to keep track of and manage your library.
• The creation of word lists (the "editor") will still be there but it will be pushed into the background. After all, learning is the main task.

The basic file unit is the "collection" which is a file with several "lessons". A "lesson" is simply a list of words with one or more translations. The student can train the words in the whole collection or select one or more lessons to focus on during the training.

The first sub project that we want to do is the dashboard, which will show the student the collections that have been marked as "active", i.e. which the student is in the process of learning. For each collection we need an icon or maybe a bigger picture that represents the collection. We also need text overlays which show how many percent of the total learning that the student has already finished and how many words are due. Inspiration for how to do this can be found on thaipod101.com. I'd love to include a screenshot but the vocabulary trainer on that site is a premium feature and my test period is already over. I'll try to solve this somehow and be back with pictures.

Are there any takers for this? Parley is a good program already and with this project we can make it awesome. Not to mention move it onto mobile platforms, which is the next frontier for KDE in general and KDE Edu in particular.

Inge Wallin, Amarvir Singh
Parley maintainers

[1] http://edu.kde.org/parley
[2] http://edu.kde.org/

I'm interested in helping you out! Though I have to get to know the application a bit better first.

Very nice; Thank you!

You can install Parley on any Linux distribution pretty easily and play a little with it. You can download collections lessons via the Get Hot New Stuff mechanism to get something to work with.

Just be warned that the "old" Parley is a functional but pretty dry application. Already now in git, it has more learning-oriented features but to see that you have to build it yourself. Actually I think you need to do that to be able to be part of this and naturally we will help you do this. Once the build environment is set up, the process is actually pretty simple.

Let us know what type of help you need.

ingwa wrote:Very nice; Thank you!

You can install Parley on any Linux distribution pretty easily and play a little with it. You can download collections lessons via the Get Hot New Stuff mechanism to get something to work with.

Just be warned that the "old" Parley is a functional but pretty dry application. Already now in git, it has more learning-oriented features but to see that you have to build it yourself. Actually I think you need to do that to be able to be part of this and naturally we will help you do this. Once the build environment is set up, the process is actually pretty simple.

Let us know what type of help you need.

Ôh ok. I'll try to do that tomorrow right away then. I already prepared a few question I have to ask, but I'll check if they're still valid with the version on git.

For communication purposes ... what would be the best way to stay in contact with you people? Should I join your IRC, mailing list or does this forum suffice?

For communication in general I think this forum is enough. But if you want a quicker answer then I recommend email and for real-time communication e.g. about installation issues I recommend #kde-edu on irc.

So I have managed to compile parley from git. It was actually fairly simple with the help of the wiki and

Code: Select all

apt-get build-dep parley

I should remember this command for future use.

My questions and observations

1. Parley goes through all lesson without breaks on default.
I noticed that all lessons are selected by default. Is this intended behaviour or is that a bug? I don't think it makes much sense to divide the collection into several lessons only to teach them all at once.

2. How is a finished lesson defined?
Is there some recommendation on Parley's side whether to repeat the lesson or moving on the the next? Example: The user manages to get 80% of the vocabulary right and is then advised/allowed to start the next lesson.

3. Is there a reminder to go through old vocabulary again?
Does Parley notify the user after a week or so to train the old vocabulary again? Or do the users have to initiate this on their own?

4. Is there a maximum amount of lessons in a collection?
The headlines says it all. I saw "collections" with only one lesson, which holds 1000 vocabs, and Collections with several lessons, each containing only a couple of vocabs.

5. What do these percentages indicate?

Because one can hardly recognise the values - they read 16%, 29%, 14%, 14%, 14%, 14%, 14%, 14%
In the beginning I loaded up the collection and simply always clicked "I knew it" in flash cart mode. Afterwards all the values were 14%. So I wonder what these values represent. My progress? The amount of correct answers?, etc.

6. Can I get some pictures?
I would be really nice if you could give me some pictures of a somewhat fuller library, with collection in higher grades. I'd like to see how Parley currently presents this information. My library is very sparse at the moment and naturally I don't have high graded collections yet.

Very good that it was easy to build Parley. It will be slightly more difficult in the future because then you will have to compile the edu libs too. But for now the build-deps in the system is enough. :)

Here are the answers to your questions:

1. Parley goes through all selected lessons and by default all lessons are selected. This much is true.

We are not satisfied with the terminology that is used now. As you notice, the word "lesson" is only partly appropriate here. To be honest I am not certain of the reasoning behind the terms "collection" and "lesson", but my guess is that the idea is that a pupil uses a book in a language course. This book is divided into chapters or lessons and each lesson has a number of words to study. When the practice starts, the pupil can select the lesson(s) that is appropriate for the study this time and study those lessons. This means that they have to select the right lesson(s) before practice. We will most likely switch the word "lesson" for "unit" instead since that is more generic (what is a sublesson?) and also because it's used by other applications in kde edu such as Artikulate.

One difference between released Parley and Parley in git (which you built) is that the new Parley doesn't go through all words in the selected lessons in one go. Instead the practice is broken up into "sessions" which is a subset of the selected words. The default size of a session is max 20 words out of the full set and of which at most x (5 by default) can be new words, i.e. words that have not been practiced before. The numbers 20 and 5 can be changed if you like in the "Configure Practice" dialog.

But your description makes me believe that you built the new Parley but actually ran the old one. If you do "which parley", does it say "/usr/bin/parley"? If it does, then this is most likely not the parley that you built. To see all parleys that you have installed you can type "which -a parley" and it will give you a list. If your newly built one is not the first in the list, then it's not the one that will be executed by default.

2. A lesson is finished when you have practiced and gotten right every word so many times that you have reached "level 7" on every word. If you open the "Configure Practice" dialog in the "Blocking" tab, you can see how often words will be repeated again if you get them right. If you get them wrong you will see them immediately again in the same session and they will be brought back to level 1. This is according to the "spaced repetition learning" theory which you can read about at https://en.wikipedia.org/wiki/Spaced_repetition. In other words, getting a lesson to 100% takes ~3 months altogether and this is because you need to repeat it during a long time to get a word into your long term memory.

3. Yes. See the answer to question 2. When you start a new session it will only show you the words that are due according to your "blocking" settings and the times that you practiced them last time. But this is not clear to the user in the current implementation and that's one of the most important things to make more clear.

4. No. A collection can have any number of lessons and sublessons to any depth. And a lesson can hold any number of words.

5. See answer to question 2. It's how many % on the way to total learning of all the words in the lesson that you have travelled.

6. Hmm, I'll try to give something to you. But this is not what we are trying to improve first. It's unfortunate that you are presented with this image before you do any training (see why we want to improve it? :) ). What we want to present to the user is an overview of where they are. The picture that you see is only the statistics for *one* collection that should be presented *after* the training not before it. Before training starts, you should be given an overview of all your active collections so you know which one to go into and start training. And I'm thinking of supressing all of this detail into some special detailed statistics that you need to open separately to get at and instead just give the user a single figure by default (like: "You have finished 21% of the practice of this collection" or something).

So, that's the answers. But please make sure that you are actually running the new Parley. I suspect that you are not.

-Inge

ingwa wrote:Very good that it was easy to build Parley. It will be slightly more difficult in the future because then you will have to compile the edu libs too. But for now the build-deps in the system is enough.

Here are the answers to your questions:

1. Parley goes through all selected lessons and by default all lessons are selected. This much is true.

We are not satisfied with the terminology that is used now. As you notice, the word "lesson" is only partly appropriate here. To be honest I am not certain of the reasoning behind the terms "collection" and "lesson", but my guess is that the idea is that a pupil uses a book in a language course. This book is divided into chapters or lessons and each lesson has a number of words to study. When the practice starts, the pupil can select the lesson(s) that is appropriate for the study this time and study those lessons. This means that they have to select the right lesson(s) before practice. We will most likely switch the word "lesson" for "unit" instead since that is more generic (what is a sublesson?) and also because it's used by other applications in kde edu such as Artikulate.

One difference between released Parley and Parley in git (which you built) is that the new Parley doesn't go through all words in the selected lessons in one go. Instead the practice is broken up into "sessions" which is a subset of the selected words. The default size of a session is max 20 words out of the full set and of which at most x (5 by default) can be new words, i.e. words that have not been practiced before. The numbers 20 and 5 can be changed if you like in the "Configure Practice" dialog.

But your description makes me believe that you built the new Parley but actually ran the old one. If you do "which parley", does it say "/usr/bin/parley"? If it does, then this is most likely not the parley that you built. To see all parleys that you have installed you can type "which -a parley" and it will give you a list. If your newly built one is not the first in the list, then it's not the one that will be executed by default.

2. A lesson is finished when you have practiced and gotten right every word so many times that you have reached "level 7" on every word. If you open the "Configure Practice" dialog in the "Blocking" tab, you can see how often words will be repeated again if you get them right. If you get them wrong you will see them immediately again in the same session and they will be brought back to level 1. This is according to the "spaced repetition learning" theory which you can read about at https://en.wikipedia.org/wiki/Spaced_repetition. In other words, getting a lesson to 100% takes ~3 months altogether and this is because you need to repeat it during a long time to get a word into your long term memory.

3. Yes. See the answer to question 2. When you start a new session it will only show you the words that are due according to your "blocking" settings and the times that you practiced them last time. But this is not clear to the user in the current implementation and that's one of the most important things to make more clear.

4. No. A collection can have any number of lessons and sublessons to any depth. And a lesson can hold any number of words.

5. See answer to question 2. It's how many % on the way to total learning of all the words in the lesson that you have travelled.

6. Hmm, I'll try to give something to you. But this is not what we are trying to improve first. It's unfortunate that you are presented with this image before you do any training (see why we want to improve it? ). What we want to present to the user is an overview of where they are. The picture that you see is only the statistics for *one* collection that should be presented *after* the training not before it. Before training starts, you should be given an overview of all your active collections so you know which one to go into and start training. And I'm thinking of supressing all of this detail into some special detailed statistics that you need to open separately to get at and instead just give the user a single figure by default (like: "You have finished 21% of the practice of this collection" or something).

So, that's the answers. But please make sure that you are actually running the new Parley. I suspect that you are not.

-Inge

Thank you Inge for this explanation.

I, indeed, did not run the version from git. The wonders of technology Now it's fixed and the git Parley looks much better.
Before I can really start I have one last question: Can a user loose his "grades" when he suddenly doesn't remember all the words after a week?

First, I don't like the word "grade" at all because it has nothing to do with practice. This is one of the things that I mentioned in the first post, that there is a strange mix of testing and practice in the old Parley.

But the answer is yes. However, that is as it should be. The spaced training algorithm works by exposing the student to each word often in the beginning and then more and more seldom as the student learns the words. But is s/he has forgotten a word it needs to be trained more often again until it has stuck better in the memory.

And this is why I don't like the word "grade". It sounds as if you lose something when you "lose a grade". But in reality you gain something, i.e. you get to concentrate on the words that you have more difficulties with and you can push the words that you already know better a bit into the background.

Parley should help you in your studies and be supportive, not give you a bad feeling when you "lose a grade".

FYI: I have started on the dashboard now so we will soon need some help with the layout and graphics. There's no real rush yet because I need to do some coding on the infrastructure of the program first before I can start on anything that is user visible. But I think it would be good if you started to think about the problem and maybe get some ideas.

ingwa wrote:FYI: I have started on the dashboard now so we will soon need some help with the layout and graphics. There's no real rush yet because I need to do some coding on the infrastructure of the program first before I can start on anything that is user visible. But I think it would be good if you started to think about the problem and maybe get some ideas.

Good. Before I explore this concept in more detail I'd like to know if the Parley team would like such a grid based solution or not:

Personally, I like the card view, because it somewhat give me the impression that each "collection" is something unique that has to be completed, rather than a list one goes through to "arrive" at the bottom.
Anyway, if you the Parley team prefers a list view then I'll try to get a descent mockup that uses a list view.


For all the other readers:
Where are your contributions people?? At least rip my proposal into pieces or something, gosh I feel really alone here

I am not sure which card view you mean when you say "the card view" but I can confirm that I do indeed have a cardview in mind. Just to make sure that you haven't misunderstood what I meant: The dashboard is something completely new, it doesn't exist in the current parley. It will comprise information about many "collections" into one view which will give the student an overview of his total study load, not just statistics on one collection.

And you are totally right that a collection is not a list. You will have an easy time with some words and you will have a harder time with some others. The easy ones will be dealt with quickly and the others will take some time. But there is no order between the cards, at least not in the general case.

ingwa wrote:I am not sure which card view you mean when you say "the card view" but I can confirm that I do indeed have a cardview in mind. Just to make sure that you haven't misunderstood what I meant: The dashboard is something completely new, it doesn't exist in the current parley. It will comprise information about many "collections" into one view which will give the student an overview of his total study load, not just statistics on one collection.

And you are totally right that a collection is not a list. You will have an easy time with some words and you will have a harder time with some others. The easy ones will be dealt with quickly and the others will take some time. But there is no order between the cards, at least not in the general case.

With "cards" I mean to represent the collections as little cards in the dashboard .

Sogatori wrote:For all the other readers:
Where are your contributions people?? At least rip my proposal into pieces or something, gosh I feel really alone here

It could be because not many people here use Parley and are too lazy to compile it from Git

Here's my take on your mockup: I'm not on the Parley team so I can't speak for them, but I like the idea! A dashboard would certainly have to show more than the collections, but how those are depicted both looks nice and makes sense to me to give an overview.

@ingwa
Because I believe we had some communication problems yesterday I threw together a quick picture how I approximately imagine the dashboard to be. The Field on the top will display what the student has achieved so far and give a rough estimate of how much time will remain and what lies ahead.
The student/user can then scroll down to select the collection they want to practise. This all together constitutes the dashboard.
Any objections?

The top part is of course still pretty ugly. I'm not sure if I really want to display that much information all at once. I think it's too much information at once. Anyway, I think it give a good feeling of the direction I want(ed) to take.

Any feedback criticism?