In another thread it where requests to get like ribbons or similar for achievements. I like that idea but would also like to see a similar thing on the practice page. Now we get very cheerful texts and sounds for time spend on skritter. That is very nice but I would like to see similar things when i hit full 100 marks. e.g 400 characters, 1200 characters etc. I at least celebrate my self for every 100 character i learn since that for me is an strong achievement. It would be nice to have some simple skitter equivalent. It is like the gold stars we had in school in Sweden when we did well. Even though i am well passed that age I still find that encouraging:-) At the university in China they did it the other way around. Nothing when you did well but if you did poorly on a test they really let everyone know that. I even had my test paper once for public display at the entrance gate. Luckily they have changed that part a great deal now. In a way it did force me to study harder but not really with great pleasure.

Sure, but not on the practice page itself. How about the progress page? Leave the practice page free and simple...

Actually, there's already an option kind of like this. It's really irritating. If you turn sound effects on, for every tone/character/definition/reading etc you learn you get a message at the top of the screen saying so. Massive distraction.

+1 for the ribbons and awards idea, but maybe it'd be better by email / Twitter, and on the progress page. I think this is already in the pipeline though, and I'm happy to wait.

Another thing about the progress page is that the weekly graph gets rescaled to always look about the same. I kind of want it to be more depressing if its been a week where I've only learned 5 rather than 20.

Hmm, maybe a widget to coming last week's learning rate with rate over last n months or something.

totally agree with wispfrog about the rescaling. I figured it was a tactic to not discourage anyone

I still like the idea of having rewards and cool stuff happen on the practice page, but perhaps much less frequently than my first attempt. Will have to think about it later when we pursue a more comprehensive rewards/goals/leaderboards/etc. framework.

The progress page graphs are tough to get right, because one person's 5 may be another person's 20 may be 穆尔's 150. So it's tough to do without looking at past weeks for normalization, which is expensive.

When I want to compare past weeks, I go to the month view, and that helps put it all in perspective. Plus, of course now you can see the actual numbers for each day/week-point on the graph, so that makes comparison gratifying (or not, as the case may be).

Wait a minute, how can looking at past weeks be expensive? Its single numbers that don't change that can be calculated once a week and stored. Compared to everything else its tiny.

So the classic computer science "expensive" means both time (processing) and space (memory). When I say it, I tend to also include our development time. And I should probably include our maintenance time, for fixing bugs with it, if it's something more complicated.

So for this feature, if we did it the easy way, it would be expensive in terms of processing time, and it would take some time to develop and maintain. We would just fetch a few extra data points of the summed and stored progress values. Database fetches take a long time, though, so even that makes the progress page load more slowly.

To make it any faster than that would take a ton of work building an alternate progress storage system, and then we have to keep it in sync with the other one, etc.

Better to figure out a way to minimize the issue without doing custom scaling dependent on extra past data. Maybe we'll try increasing the minimum data range in the graphs.

Perhaps it should be considered at some point to port this Google App thingy to a regular SQL database back-end? I dont know how much is Google App and how easy porting to something else might be, but being limited by Google App might be costly in the long run.

@nick I have to admit, seeing which things are expensive for you guys has really soured me on App Engine's data store.

This with app engine's data store is interesting. They do admit that they do have problems: http://googleappengine.blogspot.com/2010/06/datastore-performance-growing-pains.html I recently visited HP and they told me that Google gets 1000 blade serves per day! There we are talking about growth. The idea with app engine cloud data store is very great but the performance is (sometimes)not as good as in traditional data models. I have done performance comparisons on most of the databases on the market (DBA for some some 70 Tb databases in total) and even though the cloud data stores are very scalable they are not yet up to the performance of traditional databases. Depending on how your define a database.

Nope, it's all worth it. App Engine is awesome.