The trending page is pretty much the only reason I go to GitHub these days.
Weirdly, I've been low-key worried about it ever since they added the "Spoken Language" dropdown, which AFAICT is the only real change from the last few years.
It's not that I think being able to search by spoken language is bad or anything, it just screams of a middle manager trying to "solve" the important problem of a few people being confused about how the "language" filter referred to the programming language, and the solution was "let's add another button!"
The language filter was actually a suggestion I made (I’m probably not the only one but I sent them feedback emails and everything).
The reason is that at some point a few years ago, the trending list was filled with Chinese repositories which makes it useless if you don’t speak the language. The language filter fixed it for me and made the page useful again.
I mean that sounds like one solution - especially if you cannot work with Chinese nationals for reasons. But in the spirit of open source and access to other people's code, an alternative solution would be for Github to integrate with a translation service and translate descriptions.
> an alternative solution would be for Github to integrate with a translation service and translate descriptions
Please don't do this - plenty of service auto-translate descriptions and they drive me crazy. If the description is in a language I don't speak then at least I know that I can skip it. But if it's auto-translated then I'll click on the repo, only then realize that I don't understand it, go back and try again with the next repo.
The only situation in which I would welcome such a change would be if GitHub translated comments, function names, variable names, READMEs and all documentation (including screenshots and reference guides not hosted on GitHub) AND THEN translated my contributions back from my language to the original one.
> especially if you cannot work with Chinese nationals for reasons
This comes off as inflammatory. It sounds like you're misrepresenting GP's desire to filter out a language they cannot read as a kind of bigotry against the speakers of that language.
That may explain the apparent disapproval from the community.
As for auto-translation, there are plenty of browser extensions and other solutions if people want that.
That's not how it read to me, as bigotry. It read more as, if you work in some sensitive role such as in the government, and you cannot work with Chinese nationals due to that, then the filter makes sense.
I also do agree with the filter itself purely because I ran into the same problem too, lots of Chinese repositories with poor if any English READMEs.
You need to look up the definitions of "cannot" and "Chinese nationals". The first means you don't get a choice. The second is citizenship - not race, ethnicity, or language. How you misread that as choosing not to work with Chinese people is beyond me. Seems like you're the one looking to be inflammatory.
Linkdink, things are a bit more cordial around here than you might be used to from other forums. Mine was just a friendly note to someone who seemed to have incurred the displeasure of the community, fair or not, suggesting why that might be so.
There’s a need for language filter for another reason: number imbalance. The trending page used to be full of Chinese projects with an absurd number of stars, so no non Chinese project could be featured on the trending page.
At some point I actually tried to read some of these projects and they often were complete garbage (empty repositories or disgusting code) or completely specific to China. The useful projects tended to have an English readme anyway.
This honestly made me realize there’s kind of a fracture in open source between China and the west: Chinese don’t speak English, and the west doesn’t speak Chinese. It’s almost like they have parallel libraries for everything we routinely use. Which makes sense because we rarely provide Chinese documentation on our projects.
Just a nitpick: I'd like to see the repo org or author in the repo listings. The top 20 is currently listing 3 different stable-diffusion repos with no obvious differences (except for the ranking scores).
I have run across several projects I wanted to use but whose main documentation was not in a language I spoke well enough. It would be nice to only show repositories documented in languages I know well enough to interpret reasonable demands.
Totally agree! Documented language is a great filter feature. I want more filters, not fewer.
The fact that this is the only filter they’ve added in years, and it is next to a different “language” drop-down doesn’t suggest that they’re really interested in making filtering actually helpful — but that they’re trying to solve one specific label being a source of confusion by adding another button.
I check this page everyday. A lot of my starred repos come from this page.
I think this hints at part of the problem with Github as a social hub for tech. I do the same thing, and star things I've seen in the Trending page. Very often I never actually go back to those things though. That star is essentially a signal to the repo owner saying "I'm happy this exists." The problem is that there are also people who star things to say "I love this" or "I am using this" or simply "I'm bookmarking this."
That's fine, of course, but then repo owners then start saying things like "My library has 5000 stars, it's clearly the best library for X!" and "My library got to 20000 stars much more quickly than Y and Z, so people must think it's better than Y and Z!"
Github probably should turn off the social features unless they can think up a way of discouraging people using the social aspect as a measure of quality or popularity when they're really not those things.
Hm, I use the star as "interesting; maybe good". It's a very weak signal that I didn't think something was awful and wanted to be able to find it again. I wouldn't assume that a repository is amazing just because it gets 10+k stars -- this only means that it became viral.
GitHub should reduce starring to just be a bookmarking feature, that is, not show the numbers of stars a repo has. Showing it is what makes it an ambiguous social feature.
I see GitHub stars in orders of magnitude (1-9, 10-99, 100-...). If interested, I also look at the other public metrics: # of branches, forks, open and closed PRs, time since last update. I have also used <10 star repos in high traffic prod env with no issues. I think you're right and I'm fine with 1 star = 1 person took interest for at least a moment and liked what they saw.
For me, the explore page ALWAYS says I'm interested in the topic "vorbis audio" because of the 'topics that i follow' (I follow no topics), so that's a nice bug. 75% of the repos there never change for me.
The trending page otoh is nice and fresh, though I wish they'd change it to prevent already popular repos like typescript from "trending" (why are people starring this? ridiculous.)
Just curious, but why star them at all? I've starred 11 repos in my many years on github, to bookmark some specific projects. But others in my feed star projects daily and have over a thousand projects in their star list. I'm curious to why. Will you ever check them out again? Isn't the list now so long it's hard to find things in it?
While maybe it's not very useful for your own bookmarking, there are other useful effects:
- Saying "I think it's neat" internet points is kindof a nice kudos. It's relatively pleasing to see some stars on your repositories.
- It's an indication of popularity/usage. If you have to rely on some dependency, "choose the one with the most stars" will probably give you a good result for next to no effort.
- In terms of social networking: I can then see what repositories others have starred.
I use it as both a bookmark so I can find a repo again in future, but also as a kind of "I think this project is great" thumbs up. It's pretty clear that that's a common reason for stars.
Agreed, it's also how I keep my finger on the pulse of languages I enjoy that I'm not professionally using at the current moment and also to keep up with niche domain specific languages as to where and how they are being used. Boo!
Unsurprising, in my usage of GitHub, I'm usually completely disinterested in what others are doing. Maybe they could occasionally feature really interesting hand-picked projects.
From [1], 1a "not having the mind or feelings engaged" certainly sounds like GP's meaning and is the same definition given in [2] "not having the mind or feelings engaged".
> Due to low usage of Trending Repositories and Trending Developers, this tab will no longer be available
I've never used this tab. (Just opened it for the first time.)
It's cool, and I'm disappointed to see it go.
BUT: We should spend a minute thinking about the counter argument. Little and unused features can add cruft to a codebase, and significantly impede development. It's possible that hooks for corner cases and features can be so complicated that working around them becomes more burdensome than supporting the important use cases.
Case in point: When I was lead on Syncplicity, we had a corner case where a given folder on disk could be in multiple synchronized folders in the cloud. (Long story.) The problem was this corner case made many use cases, in the desktop client, significantly more complicated and bug-prone than needed. (What happens if you have different permissions to the same file?)
There was a point where I put my foot down (as lead) and said we couldn't support the corner case anymore. This lead to almost a year of two branches of the product, and daily merge conflicts between both branches. I had to put my foot down again (as lead) and push that we had to ship the version of the product that didn't have the corner case and stop working on two branches.
This makes me wonder if Github needs to cut back on corner cases and other little-used features in order to fix some maintainability issues?
The main issue is the conflating of usage and value. The persona/user group that consumes the Trending pages are likely users who themselves contribute to open source, and thus are part of a flywheel that drives more code contribution overall. (e.g. how 1% of users post on reddit vs 99% consume) Not all usage is made equally.
This scenario reminds me of a funny story posted on HN a while ago: https://memex.marginalia.nu/log/48-i-have-no-capslock.gmi (TL;DR; if you take this "usage based" feature removal logic to it's conclusion, a lot of important keys would be removed from computers/keyboards).
Sounds like a fun weekend project to replicate it?
The GitHub search API seems to provide everything that is needed. It returns 30 repos per query and allows 30 * 60 * 24 = 43200 queries per day. So one could track the top 43200 * 30 = 1296000 repos.
That is enough to track all repos with more then 20 stars:
I spent the afternoon building something similar to the trending page (starting with tracking github actions, will add more features over the weekend): https://trendy.dev/
GitHub could never quite decide whether they want to be a social network or not. The "Trending" page (which has been useless for a long time due to the ranking algorithm rewrite several years ago) is just one example of this indecision. The way users can vote on issues with "thumbs up", "heart", "rocket", and "fanfare" emojis (instead of simply "up" and "down") is another, and the ridiculous "achievements" on user pages is yet another. They've also added and removed the "Social Coding" label several times from their front page if memory serves correctly. It's like the whole company is having a neverending identity crisis.
I think criticizing them is fair. GitHub is a well established code repository which does something and does it well.
Microsoft endlessly tinkering with the product shows a lack of stability in it's lifespan, what if you adopt a new feature they add only for it to be removed in the next cycle. But they paid 7.5 billion dollars for the website, and need to justify the salaries of dozens of VPs of marketing and PMs, so here we are.
That is helpful at times. There are some developers whose opinions I know and respect, and it's helpful to see what projects are catching their attention.
The social network aspect was a means to an end ("get enterprise customers"), but the needs of enterprise customers differ from the needs of open source communities. Hence the "indecision"
> the needs of enterprise customers differ from the needs of open source communities
Indeed, but GitHub's "social" features cater neither to enterprise customers nor to open source communities, but to a (hypothetical?) group of people that want a blend of software engineering and Instagram-style popularity chasing, with a bit of cartoon cuteness sprinkled on top.
I honestly found the introduction of the "achievements" section on user pages to be borderline insulting. I'm building cool stuff, not playing an RPG. Who does GitHub think their userbase is?
Stars at some point served as "social proof" of something (eyeballs, users, interest), but Goodhart's law struck.
At least three investor groups are explicitly looking at lists of the most starred / forked projects and throwing money at those with "Instagram-style popularity", so we've reached the stage where people can trade in the "internet points" for real money.
ironic, as a person who was very warm to GitHub previously and after attempting to coldly assess their offering for enterprise: I found it to be quite lacking compared to Gitlab.
I can't imagine choosing GitHub except for cool points or if you're working very closely with code that isn't yours primarily.
For example: Things like SAML and SCIM are very important to enterprise customers and it feels extremely jank on GitHub.
We're on Azure DevOps (despite deploying to AWS) and it's been weirdly good. I'm a little sad that Microsoft's focus is shifting to GitHub. We're evaluating GitLab as another possible all-in-one solution.
The worst thing on GitHub are those “rating” badges in profiles that always say A+ regardless of the profile quality. Just having one of those is a decent screening filter itself.
> While people have been able to include emoji in responses for a long time, using them as reactions resulted in a lot of noise. In many cases, especially on popular projects, the result is a long thread full of emoji and not much content, which makes it difficult to have a discussion. With reactions, you can now reduce the noise in these threads.
I'm well aware of why reactions were added, but the way it was done is asinine. Instead of adding two, unambiguous ways to express support or opposition ("+1"/"-1"), GitHub chose to add three or four variations of each ("heart", "hooray", "rocket" etc.).
Now some people, to express support, react with "+1". Others react with "heart" or "rocket". How do you find the issues with the most popular support behind them? Do you sort by "+1" or by "heart"? On some issue trackers, one reaction type is preferred over another. On other trackers, reactions are mixed, and you find issues with 5 "+1" reactions and 40 "heart" reactions, and issues where it's the other way round. This is so hilariously bad that it almost feels like deliberate sabotage for laughs.
Imagine if HN, instead of upvotes and downvotes, had eight different ways to "react" to posts and comments, each with their own subtly different connotations. That's what GitHub implemented. It's something I would expect from a social network where people share photos of their cats, not from a system where software deficiencies are being discussed.
I think the social networking aspect of GitHub is one of its worst features. It carries with it all the negatives while also being required to use at work.
for me personally the only thing I evaluate this type of services against is:
- is it free for my personal projects?
- how many interesting libraries I may use are there?
- does it offer a free tier of pipelines and things like that?
- can I make private projects for free?
After the ranking algorithm change to the Trending page, I found I basically stopped using it (I lived on that page before the change), I found I'm more likely to find cool projects/libraries/etc here.
The rewrite was horrible. I remember before how you could find interesting projects from random developers with growing communities. Now it's side-project X from reknowned developer N, or the newly created repo from some known tech firm that somehow has 100 stars already. All the actual fresh meat got dried out. I lost interest after the rewrite.
Same here. For the first time ever, I clicked through the explore section today and found nothing of interest there. They could remove the whole explore section and I would never have noticed.
The same. Although, you would think GitHub had the resources to keep maintaining this page even for just public image purposes among the power users / enthusiast who maybe care more. Seems like they are tuning town the social media aspect in their marketing too.
That's a shame. I admit that I hardly ever look at this page but when I do rediscover it every few months I always stumble into a load of great projects.
most things where organic growth can shine through seem to get nerfed and/or closed down. they made it worse years ago, and will now close it even though it's a great view into what's happening on github, great for fosterting what little spontaneous community/interaction exists on github, and enjoyed by many people.
but as others have alluded to, this will send some PM's kid(s) to college or help with the mortgage on their 2nd or 3rd property and should be a reminder that it's not great to rely on a company (especially a big co. with a wild track record like microsoft) for services, especially where important things like community or source code management are concerned
The social aspect of GitHub is the only reason I'm considering it over GitLab for a new open source project. Reducing discovery seems like it could be harmful - they should really provide readers rough comparative traffic to better justify this..
I've been a big fan of the trending page, and have found many cool projects from there, as well as been very happy when my project has made it to the list, which usually brings on a big wave of new users!
As someone who's building apps and actions to run on GitHub, I've always found the discoverability of apps to be lacking. The marketplace is nothing but a big list of things, in seemingly random order. Compare this to the Slack App "store" which has a much better discoverability and feel to it.
I which it had reviews, trends, new highlights etc, to make it easier to discover new apps, and for app developers like myself to get a chance to attract more users.
Just started building something, let me present: trendy.dev! For now it's just a list of actions, sorted by stars. I'm planning to add a few more features, to keep track of currently popular actions, and of course, repositories that are currently getting a lot of traction!
Just because some feature has low usage does not mean you should deprecate it. The whole point of GitHub is to provide discoverability to what others are doing. If anything they should improve upon the feature.
GitHub calls itself social coding, but trending fits in poorly with that. When I hit a problem that I think might be solved by open source, telling me that the TypeScript repo is trending gives me exactly zero information.
If we're honest, the Trending tab often seems to mean, "these projects went viral on HN/social media." And I'm not sure why they deserve extra attention, but it follows the general trend of social media being obsessed with perpetuating the Matthew Effect.
Shame, having my repos show up there felt really validating - and judging from the amount of people that ended up being interested, it must've had some kind of userbase.
Since there are less and less ways to discover cool repos, recently I’ve made an open source extension to get recommendations on GitHub. You can check it out https://indexstorm.com/git-rec
Just installed it and am already loving it! I tested it on a repo for which I thought I knew all the available alternatives but was surprised to find multiple new ones!
And I am saying as a person who has always hated recommender systems at shopping sites/social media.
I am fine with company killing apps or features that don't make money. What i don't like is that company uses corporate talk to hide the real motive to kill popular products. This has happened to CentOS after ibm taking over redhat, now we are witnessing this.
I think they want you to subscribe to email updates to explore new interesting repositories. Probably they are going to use it for marketing too. That might be the reason of the deprication of the page.
That is really weird. It's clearly a very popular feature. Part of me thinks that their metrics are broken for some reason and that this has been incorrectly flagged as low usage
There's a lot of "hidden" stuff on Github. I'm not really sure why, the default page when you're logged in and go to github.com is pretty much useless. They could put some of the functionality there.
Maybe you can change it, I don't can't figure it out. Why is "Following" the default, I never need to know that.
I never used the "Trending" but it could be one of these cases where companies are measuring something, decides it's not being used, but fail to consider why that might be. Personally I don't see it being particularly relevant, but many seem to enjoy it.
To navigate to the "tab", which is just a page really. Click on "Explore" and then it's in the middle of the page, under the black menu. Or go here: https://github.com/trending
huh!?? don't they have amplitude to check how many users click that page? I click on that every day to see interesting projects when I'm bored and I have starred so many cool projects from there..
the product people is insane, it's an insane decision
> Seems like a totally reasonable product decision from my seat.
The same probably applied to shutting down Google Reader. But it's still brought up a decade later, because it hit the "right" group of users. These changes spend a lot of the goodwill built by Github.
Seems like a weird decision not to expose this in some capacity.
It encourages external crawling and aggregation of content hosted on their site. Seeing that their search feature already produces abysmal results, it only further obfuscates public repositories.
I spend so much time exploring repos related to work or personal projects. I literally do not have time at the end of my day where I think to myself, "you know what, I really want to find some more repos to star and never look at again."
Weirdly, I've been low-key worried about it ever since they added the "Spoken Language" dropdown, which AFAICT is the only real change from the last few years.
It's not that I think being able to search by spoken language is bad or anything, it just screams of a middle manager trying to "solve" the important problem of a few people being confused about how the "language" filter referred to the programming language, and the solution was "let's add another button!"