ruby
Rework or remove top-projects page
(This does not have highest priority.)
The top-projects page a long time(?) ago showed the most highly ranked projects from Rubyforge. At some point the list has been removed, probably because it was completely outdated.
It should be re-thought from scratch whether there should be a top projects page at all, and--if yes--what it's content might be.
At some point I thought about a rather different "Showcased project" block. Each month, there would be a poll somewhere where people would pick their 3 most valuable projects and they would be featured on the website, maybe with an interview of the maintainer(s). Or maybe that would be staff picks, or Matz' picks, or…
Any other crazy idea like this? Please, share.
Sounds really thrilling but also like an extremely high maintenance effort, not to speak of all the translations...
Yes, that's why I called this "a crazy idea" :) Crazy, as in "probably dumb".
But with a team of dedicated maintainers, after we cut it down on the number of available translations, it may turn realistic enough…
I guess it is difficult to fetch, with full automation, a good list of "top projects". GitHub provides a list of trending ruby repos but 1/ that's on GitHub only (wait, Ruby == GitHub nowadays, isn't it? :neckbeard:), and 2/ we have no control on it.
Here's a list of specific "top lists" I can think of right now:
- http://rubyforge.org/top/
- https://github.com/languages/Ruby
- https://www.ruby-toolbox.com
- https://rubygems.org/stats
I actually had created a rake task to scrape the top 30 from rubyforge, so it can't be difficult :) I decided against pushing it, though.
In my opinion a good list needs editing.
I agree with you: a good list needs editing. That's what it is difficult to fetch a good list, not any list :)
Next step would then be: can we agree on what's good editing so as to design an agreeable Top List? I guess… not, actually. It is too subjective and biased. That's why I thought of rather different perspectives, such a highlighting projects from time to time, or simply linking to the various automatic, "trending" lists available out there.
If we're calculating Top Projects based on download count, we wouldn't need to edit the list very often. Rails, Rake, etc would always be most downloaded gems.
I don't think every single most often downloaded gem qualifies as "project". [E.g. activesupport (no. 5 on rubygems.org) ?]
What about using Ruby Toolbox? Not to fetch a Top Projects list to be featured on the sidebar, rather as a link toward a resource with… lists, whatever they are. People do love list, but that doesn't mean we should actually have one on the website. "Looking for popular libraries?" **BAAM, Ruby Toolbox link__ (and maybe other links).
+1 for adding interesting resources such as Ruby Toolbox. Not all gems are projects (devise and devise-encryptable are part of the same project). Not all projects are gems, either. However, how does Ruby Toolbox categorise the most popular projects? I think that "Most active" isn't really an indicator (as new gems are very active but not very polished)
Ruby Toolbox is an under-appreciated resource. I think it's more important to emphasis picking the right tool for the job, instead of just listening the most popular projects.
I replaced the old, broken content with an appeal to contribute ideas (910cbfe8995b8b77f848dfd606423a5d08e1982c), to avoid PRs like the above. Still hoping to have some content soon...
I'm thinking in the direction of a very limited number of handpicked tools/projects (e.g. Rails seems to be a must), and a collection of links to top lists.
Hey all, Ruby Toolbox guy here. While making a change to the libraries page I stumbled upon this discussion. If you'd like some simple API or other means of integration to get data out of the toolbox, I certainly could add that.
As some of you pointed out, the most popular gems are not neccessarily of much interest (see the Top 15 sidebar on https://www.ruby-toolbox.com/) - mostly it's Rails and it's dependencies. Maybe a selection of links to categories from the toolbox would be better. These could be either based on popularity or hand-picked to give a bit of a broader perspective on the ruby ecosystem in general (as the popular-stuff would most likely be rails and it's surroundings). Couple of suggestions:
- Web App Frameworks
- Background Jobs
- Object-Relational Mapping
- Unit Test Frameworks
- Code Metrics
- Ruby Version Management
That would be a basic mix of projects for building actual stuff, code quality and dev environment maintenance.
To add some feedback, I think at this point in Ruby's lifecycle isn't what projects are using Ruby.. but what companies.
In order for Ruby to be taken seriously, other enterprisey companies want to see the big names that are using Ruby and hear their success stories.
What do you think?
Instead of just trending repositories from github api we could use the star count also like here. more starred repos means more chance of being a top project.
The problem with Stars, is that older repositories always seem to have more stars.
I'm stongly opposed to just reproducing "top downloaded" / "top starred" lists from GitHub, RubyGems, whatever.
If we cannot provide good, handpicked, edited content we should drop the page completely.
IMO we should remove the page for now and continue the discussion about alternative contents. It's not well that a link to an unmaintained content is in the top page for a long time.
Thought?
+1, the integration with the Toolbox that I suggested seems to be irrelevant, and in the current state the thing is of no use.
Agree with @ismailarilik, I think this should be closed after https://github.com/ruby/www.ruby-lang.org/pull/608 @hsbt what do you think? :)
