Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

x/pkgsite: "popular packages" algorithm might be off #37703

empijei opened this issue Mar 5, 2020 · 4 comments

x/pkgsite: "popular packages" algorithm might be off #37703

empijei opened this issue Mar 5, 2020 · 4 comments


Copy link

@empijei empijei commented Mar 5, 2020

What is the URL of the page with the issue?

What is your user agent?

non relevant



What did you do?

Visited the page and took a look at the popular packages.

What did you expect to see?

A list of the popular packages sorted by some criteria, may it be dependents count or amount of times it was looked for on the site. I would also expect this to take into account the update frequency of the package to not advertise stale packages.

What did you see instead?

Some popular packages in a surprising order.

The current order is

  • Sourcegraph
  • Sourcegraph
  • Sourcegraph
  • Sourcegraph
  • Sourcegraph
  • Sourcegraph
  • Sourcegraph
  • Sourcegraph
  • net/http
  • encoding/json

This is unexpected for many reasons:

  • How is cobra more popular than json?
  • Why is the stdlib mixed with external packages?
  • glog has not been updated in 5 years and I think it to be quite self explanatory so I don't think it is here because it is commonly looked for in the docs. If the reason it is here is that It has 43k dependents why is it not above viper which only has 9.7k?
  • Testify has 9k dependents, why is it not here while echo is?

Could you please shed some light on the algorithm behind this list? Will the list auto update if the lookups or dependents amount change?

Is the currently displayed list working as intended?

Thanks in advance.

@gopherbot gopherbot added this to the Unreleased milestone Mar 5, 2020
@gopherbot gopherbot added the pkgsite label Mar 5, 2020
@julieqiu julieqiu changed the title The "popular packages" algorithm might be off. "popular packages" algorithm might be off Mar 5, 2020
Copy link

@mvdan mvdan commented Mar 6, 2020

I am similarly confused by "featured packages". Who is featuring them? Is the list generated? Is the list changed or updated periodically?

What worries me the most about these short lists is that they encourage blindly following advice on what's popular and featured. If I am new to Go and visit that page, I might come to the conclusion that logrus is the most popular and idiomatic way to do logging, or that gin is the best way to write http servers. When, in practice, popularity is very hard to measure and shouldn't be front and center when choosing what module to use.

Copy link

@mholt mholt commented Mar 6, 2020

My $0.02: I think both lists are not helpful. I think they should just be removed. Or better yet, replaced with a few tips about how to search effectively, like example search queries. Some examples could include:

  • A straightforward package path
  • With version
  • Searching forks?
  • License filter
Copy link

@julieqiu julieqiu commented Mar 10, 2020

Thanks for the feedback! @mholt - that suggestion makes a lot of sense.

We are currently working with our product and UX teams on updates to the homepage.

/cc @spf13 @fflewddur

@julieqiu julieqiu added the UX label Mar 10, 2020
@julieqiu julieqiu changed the title "popular packages" algorithm might be off x/pkgsite: "popular packages" algorithm might be off Jun 15, 2020
@jamalc jamalc self-assigned this Jul 29, 2020
Copy link

@jamalc jamalc commented Jul 29, 2020

We launched a redesigned homepage today and removed the popular packages section so I'll close this issue.

@jamalc jamalc closed this Jul 29, 2020
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
None yet
Linked pull requests

Successfully merging a pull request may close this issue.

None yet
6 participants