News for hackers. Stuff that matters.
[dead]
Re:Considering Sourcehut by junon on Monday July 25, 2022 @04:27AM (#32221839)
Because it's inconvenient. I went down the sourcehut road. Realized people don't need that stuff to develop, and it's way more annoying to work with when you just want to get something done and you don't _care_ about the "truly decentralized" parts.

To be clear, I adore sourcehut and it felt like a good way to develop. Certainly felt more like a hacker doing it. But I had the exact concerns as the author. I needed contributors, or the potential for that.

Re:Considering Sourcehut by Arnavion on Monday July 25, 2022 @04:38AM (#32221926)
Right, calling contributors "hostile" and putting "hate" in their mouths is sure to win them over to your cause...
Re:Considering Sourcehut by bayindirh on Monday July 25, 2022 @04:28AM (#32221842)
Because many people are afraid of what is not familiar, that's it.

It's not hostility, but a rather a case of shying away from reading the docs and changing the way they work.

From my experience not everyone is conductive to try and adapt to new ways, and many people do not use git CLI or SSH. They just use GitHub's gh CLI command.

Re:Considering Sourcehut by WJW on Monday July 25, 2022 @04:41AM (#32221943)
This needlessly dismisses people who have tried both and still prefer not to use the one with more hassle.

Believe it or not, there are people out there who don't care about decentralization and latency-tolerant store-forward networking for their own sake. Getting things done is way more important to them and the user experience of git send-email is simply not as good as the centralized version.

Re:Considering Sourcehut by bayindirh on Monday July 25, 2022 @04:46AM (#32221973)
People can have preferences, I deeply respect that.

On the other hand, I put my code out there (anywhere), and people are free to decide whether they contribute to it or not.

This is the platform I use, and this what the platform provides. If they want to contribute, and don't prefer the mail flow, they can reach me, and we can discuss it.

This is a two way street. Nobody is forced to contribute to anything, and I have no obligation to use services which people prefer. No?

Re:Considering Sourcehut by drKarl on Monday July 25, 2022 @04:37AM (#32221916)
It's not black or white, I love CLI and I use git CLI and SSH, I use tmux, vim etc, but to diff two branches or perform a code review I prefer my IDE (IntelliJ Idea) or a Pull/Merge Request on GitHub, Gitlab or similar
Re:Considering Sourcehut by bayindirh on Monday July 25, 2022 @04:41AM (#32221940)
Of course. I'm in the same boat with you, but in the younger generation I generally see a trend.

GitHub creates this intellectual soft-jail, making developers think that the only way to work with git is gh or GitHub itself. If Kernel can cooperate with e-mail, everybody can.

OTOH, Source Hut is considering a PR-Like subsystem IIRC, and they use mail-based collaboration primarily, because it allows cooperation without opening an account with Source Hut.

This is an important feature, and line of thought, if you ask me.

Re:Considering Sourcehut by anotherhue on Monday July 25, 2022 @04:31AM (#32221863)
I agree, I see many impedances (sorry) in command line usage. I think the percentage of people who can use one declines every year, there's simply no competing with the advertising budgets of flashier systems.
Re:Considering Sourcehut by lucideer on Monday July 25, 2022 @04:31AM (#32221862)
Because:

- UX is important to the average software developer

- UX within popular email clients targeting the task of code review is non-existent

- UX within IDEs for code review is better but: (a) send-email integrations are less polished/mature than they could be (largely due to the dominance of PR/MR flows) and (b) IDE code-review UX is still, despite everything, lagging Web UI code review UX

TL;DR: open distributed specs are what we need ideally, but you can't escape the fact that they'll always need much more work to overcome feature inconsistency and niche uses

---

I'm really rooting for sourcehut in this battle as I think if it got some adoption we could start seeing improvements to the above problems (they're all surmountable), but I think the biggest thing that's going to help that adoption is sourcehut adding a (good) web UI for their flow, per https://todo.sr.ht/~sircmpwn/lists.sr.ht/111

Re:Considering Sourcehut by girvo on Monday July 25, 2022 @05:02AM (#32222088)
> (b) IDE code-review UX is still, despite everything, lagging Web UI code review UX

This can't be overstated. Visual Studio Code seems to be making moves in this direction at least, but then again it's GitHub specific I believe?

Re:Considering Sourcehut by gspr on Monday July 25, 2022 @05:05AM (#32222100)
> Why are so many of your great contributors so hostile to the only part of the Internet that is still truly decentralized?

> Why do they hate store-and-forward, latency-tolerant networks?

I do like email-based development. And I pay for Sourcehut. But I also understand why others don't like email-based interfaces. And I too sometimes get annoyed by Debian's BTS taking minutes to act on (email-based) changes to bug reports. Minutes where I'm just idly sitting around waiting so as not to have to context switch. I abhor GH and friends, but oh man sometimes an instantaneous interface is important.

Re:Considering Sourcehut by Tmpod on Monday July 25, 2022 @06:04AM (#32222456)
> but oh man sometimes an instantaneous interface is important.

That can indeed be a flow breaker, however, in my experience, it's not necessarily an email issue, but Debian's in this case, since sending stuff to lists.sr.ht is pretty snappy (like a couple of seconds after the client sends it). It isn't as fast and instantaneous as your typical web UI yet it's not a flow breaker either, imo, though YMMV.

Re:Considering Sourcehut by gspr on Monday July 25, 2022 @09:09AM (#32223748)
But that's the point: the original poster pointed out the great value in store-and-forward, latency-tolerant systems like email. So while you may indeed be lucky and have low-latency interactions through email in certain situations, you've then added extra assumptions and requirements to the aforementioned latency-tolerant robust system. Wanting fast ("instant") response times from a system based on email seems like a folly.