Improved Community Solutions

Over the last 24 hours we’ve been pushing out quite a big update to community solutions, which we’ve been working on for a long time!

The main four new features are:

  • We now de-duplicate solutions, so even if 1,000 people upload similar solutions, we’ll only show one of them in the list. When you click on a solution, you can see extra stats.
  • You can now search by code!
  • You can now order by a few more things (with more time come)
  • For C#, you can now filter by programming concepts - Does this use recursion? Is it functional or OOP? Does it use a switch statement? etc. We’ll be rolling this out to all other tracks over the next few weeks - for now C# is our playground.

To dig in, you can click around and explore, but you might like to start with:

There are a few little bugs we’re working through, but if you see anything weird, please open a new thread with some details (let’s keep this thread for celebrating, so we can better track individual issues :slight_smile:)

1 Like

Congratulations on the great work ! :clap::clap::clap::tada:

Is sorting by stars in the projects? I often discover interesting solutions (showing out-of-the-box thinking) thanks to the stars. :sparkles:

Would it be possible to see the stats (Submissions, First submitted, Highest rep user) for personal submissions too? :smile:

1 Like

Thanks! :blue_heart:

I’m not sure really. I feel like the other sort options will probably aid discovery more than stars. And we’d have to work out if starts are cumulative across all identical solutions, or just for the top solution, etc. So I’d say I’m luke-warm on stars, but happy to be talked around :slight_smile:

By personal submission - I guess you mean a non-published solution? If you publish your solution and view the published version, you’ll already see those stats (or at least should do!) But for non-published ones, you won’t. It’s possible for us to do that, but I’m not sure if maybe it’s nice to keep that as a benefit of sharing your code with others. What do you think?

Indeed, I did not think of it from the de-duplicate point of view. I didn’t change my mindset yet! :sweat_smile:
I can see it brings on questions. And the new options are more powerful to search for solutions. But I still very much appreciate having a human-based sort through collective intelligence. I think it’s what makes the strength of Stack Overflow, for example, or even LeetCode solutions. And, in my opinion, being human centered is one of the great strength of Exercism.
Maybe it could be a subject for the next community call? :blush:

Oh yes, I do! :astonished: That makes perfect sense, actually.
I didn’t think of that because I never open my published solution page.
I totally agree on the published vs non-published part.

1 Like

I don’t how to find my published solution in the new Community Solutions configuration. I’d like to compare my statistics on a particular problem that I solved in 2 languages, but the soltion was published before the format change and I didn’t see any of this. Now, I can no longer search for the solution by user name and if such access exists for me I don’t know where to find t

You can access your published solution via your main solution page. Look on the RHS where it says “Your published solution”.

I’m not actually sure we wanted to remove the “search by user” option - we’ll probably add that back, but I’d still recommend using this back regardless.

Thank you. I should have seen this.
I hope to see the search by name option restored. There are a few people, former mentors and such, whose work I’d really like to see. I hope also, that you restore the total nummber of submissions–that gives an interesting comparison of different exercises in different languages.

Great work Jeremy and team!!! The technique filter is just amazing, my personal favourite!

I was about to ask whether the technique tag was determined via parsing or some feature like Reflection (because I only noticed it in C#), but from the looks of it various tags can be supported by any Analyzer? If so, NEAT!

Thanks! :blue_heart:

Yes. And beyond this, we’re planning on using machine learning to apply this to all the tracks. So far, it looks like it’s pretty feasible.