I use SourceHut for quite a long time. Migrating my new project
over it.
I am very happy with this completely different approach.
After a while. I feel something missing:
1) A Star/Stargazers system.
I know quite well that this system leads to so many bad things.
Such people comparing software quality by its stars.
Some have asked a similar question[1]. But I use this feature
differently. I click it to show my appreciation to the author. Not
for bookmark things.
Since the beginning, I know that this feature will never be landed
on SourceHut. I am OK with it. I can live without it. I think
there
are many options to say thanks. However...
2) A Traffic to the repository.
This is the feature[2] that I miss a lot. I want to know which of
my app is useful compared to others and its popularity. Without
this, I don't feel I get any feedback from the users. Unless they
open a ticket, which is very rare.
I am always looking for a solution. Putting analytics code in its
readme is not an option. The result will be messy if someone fork
it. Another way is using a software repository that has integrated
statistics. But not all my software is suitable to put there.
---
I post this question to get the thoughts from other SourceHut
users.
1. Why don't you like the star system. Isn't it can be one of the
basic comparison meters of software?
2. How do you know how active/popular is your repository? how do
you
count how many people attracted to it?
---
Sorry for posting this question here. After a long journey. I
decide this is the best place to ask. Thanks.
[1]:
https://lists.sr.ht/~sircmpwn/sr.ht-discuss/%3CCA%2B952WopztsU9WpeTg16GDXvM-TfkgHFpAf7ZkhxoV8sy1D5tg%40mail.gmail.com%3E#%3CCA+952WopztsU9WpeTg16GDXvM-TfkgHFpAf7ZkhxoV8sy1D5tg@mail.gmail.com%3E
[2]:
https://docs.github.com/en/github/visualizing-repository-data-with-graphs/viewing-traffic-to-a-repository
On Fri Jan 22, 2021 at 5:59 PM EST, Azzam S.A wrote:
> Some have asked a similar question[1]. But I use this feature> differently. I click it to show my appreciation to the author. Not> for bookmark things.
Just send the author an email expressing your thanks in a more personal
manner than a 1-bit "starred or not starred" flag.
> This is the feature[2] that I miss a lot. I want to know which of> my app is useful compared to others and its popularity. Without> this, I don't feel I get any feedback from the users. Unless they> open a ticket, which is very rare.
Just like stars != quality, neither does traffic == utility. Focus more
on improving your code and less on seeing if people like it. Maybe just
ask people for feedback?
> I am always looking for a solution. Putting analytics code in its> readme is not an option. The result will be messy if someone fork> it.
We would also ban this behavior as soon as it was brought to our
attention.
> 1. Why don't you like the star system. Isn't it can be one of the> basic comparison meters of software?
No.
> 2. How do you know how active/popular is your repository? how do you> count how many people attracted to it?
You don't.
Drew DeVault <sir@cmpwn.com> writes:
>> I am always looking for a solution. Putting analytics code in its>> readme is not an option. The result will be messy if someone fork>> it.>> We would also ban this behavior as soon as it was brought to our> attention.
What if I add a link in the README.md pointing to my server with a text
along the line of "if you like this project, click here", an old-style
counter? Once a week I commit an update to the README with "X people
showed appreciation for this project". Would you block even this?
I guess you wouldn't block a "Buy me a coffee" button, right? The way I
see it, they're the same thing.
>> 2. How do you know how active/popular is your repository? how do you>> count how many people attracted to it?>> You don't.
Yes, you do. There are tools for that, completely respecting the user's
privacy.
This is a recurring topic, people migrating from other source forges
obviously compare the features.
I don't want to make anyone change their mind but for the sake of
conversation I'll post here an entertaining talk by Raymon Hettinger,
Python core developer, talking about when he discovered that his package
`mathfunc.py` was used at the scientific research center of CERN in
Switzerland.
https://www.youtube.com/watch?v=voXVTjwnn-U&t=410
We all know how the majority of users behave: they show up either to
complain (bug report) or to ask for a new feature, sometimes feeling
entitled to decide on how you should employ your free/non-paid effort.
We all know that only a few people amounting to a rounding error will
write an email showing appreciation.
I guess the point of the question is not to add social features but to
get metrics to gauge interest about your project. Which could be a
positive reinforcement to keep working on it.
On Fri Jan 22, 2021 at 7:26 PM EST, jman wrote:
> What if I add a link in the README.md pointing to my server with a text> along the line of "if you like this project, click here", an old-style> counter? Once a week I commit an update to the README with "X people> showed appreciation for this project". Would you block even this?
This would be acceptable so long as it were opt-in, but I still don't
think you should do this.
> I guess you wouldn't block a "Buy me a coffee" button, right? The way I> see it, they're the same thing.
That's fine. Funding the developers is a separate matter.
> >> 2. How do you know how active/popular is your repository? how do you> >> count how many people attracted to it?> >> > You don't.>> Yes, you do. There are tools for that, completely respecting the user's> privacy.
On the one hand, it has to do with user privacy. On the other hand, it
falls into a category of features we think of as "dopamine dispensers".
We'd rather you forget about popularity and get back to writing the damn
code. Our principal purpose is to be an effective engineering tool.
Marketing your project and gauging its popularity is up to you, if
that's something you really want to do. SourceHut is not "social
coding", go to GitHub if you want that.
> This is a recurring topic, people migrating from other source forges> obviously compare the features.
There are dozens of features that people like from other forges. We
deliberately do not implement man of them. If sr.ht were the same as
every other forge, there'd be no reason to use it.
> I don't want to make anyone change their mind but for the sake of> conversation I'll post here an entertaining talk by Raymon Hettinger,> Python core developer, talking about when he discovered that his package> `mathfunc.py` was used at the scientific research center of CERN in> Switzerland.
I haven't watched this talk but I'm going to wager a guess: he didn't
find this out because they stared his GitHub project.
> I guess the point of the question is not to add social features but to> get metrics to gauge interest about your project. Which could be a> positive reinforcement to keep working on it.
Similar to our thoughts on dispensing dopamine, it is not the role of
SourceHut to dispense motivation, either.
January 22, 2021 11:08 PM, "Drew DeVault" <sir@cmpwn.com> wrote:
>> I am always looking for a solution. Putting analytics code in its>> readme is not an option. The result will be messy if someone fork>> it.> > We would also ban this behavior as soon as it was brought to our> attention.
Great to know analytics and telemetry won't be permitted. Thank you!
I really appreciate how carefully considered SourceHut is. It's easy
to implement features because other forges have them. You just end
up with GitHub but open source. That exists already - Gitea, GitLab.