I’m not super opposed, but Drew can be pretty over the top ideological/irritable for my tastes (at least that’s my perception from reading his blog). I’m not sure I want to invest in his platform if he could just boot me over an Internet disagreement or something. Yeah, I can back up my stuff, but at a certain point I may as well self-host.
I have strong principles and I stand steadfast by them. However, one of those principles is tolerance for different views. I want SourceHut to be available for everyone, even if we don't see eye-to-eye. You won't be booted off if you and I happen to end up in an argument somewhere online.
I've made myself personally available to users, as a person they can directly reach about problems and concerns with the services, rather than a support ticket answered by a random lackey. I think this is very valuable. Software is made by people. One of the risks that comes with this, however, is that people can start to view SourceHut and my private identity tangled together. Rest assured, they are separate, and I treat SourceHut with a great deal of deliberate professionalism and reverence for our users.
Nicely done.
About banning... thinking out loud...
I wonder if a public ban log would reduce food fights. Nothing too fancy. Maybe just date, account name, cite terms violated, optionally cite evidence, admin actions taken. No need for comments, discussion, whatever.
Then whenever the peanut gallery starts concern trolling, you can just link to the ban log entry. Instead of explaining yourself, yet again.
SourceHut is not a social network and does not have particularly arduous moderation needs. It's vanishingly rare for us to ban someone at all for this reason. We additionally provide banned users with a complete data export and tools to re-import that data into other SourceHut instances or instances using compatible software, so they are not left without options.
If someone feels slighted by moderation on SourceHut, I would invite them to speak publicly about it. I'm prepared to justify our actions in the court of public opinion with consent from the maligned.
He basically ignores anyone who has or currently does work at any company he deems unethical (read: Google/Amazon/Facebook/etc). You're forever tainted in his eyes, and your opinions are always only useful if someone else agrees with you that isn't tainted.
Which really sucks because his entire software stack depends on the work (and opinions) of the people that he discards: Go, Python, and Git for a few.
He's not unique in this viewpoint. Quoting jwz: "As I've said before, if you work for Facebook, you should quit; it's the only morally defensible thing for you to do."
> his entire software stack depends ...
That sounds like the Mr. Gotcha meme: https://thenib.com/mister-gotcha/ . Again referring to jwz, his nightclub is on Facebook and Instagram because it's not economically viable to avoid them. ... Gotcha!
Similarly, it's hard to avoid all software which doesn't have influence from the big tech companies. I used https://github.com/swenson/sort ('Sorting routine implementations in "template" C"), and even this little indy package has contributors from Google Inc and the main developer worked at Google for a year.
DeVault's working on a new programming language, Hare. You could view that as having the long-term unstated goal of escaping unethical companies.
The larger point still stands though. Instead of consumers simply supporting the biggest players in the market (because of convenience), it would be nice to give smaller players (who are not shady) a chance?
There's a wide gap between being irritable and being unprofessional. Nothing I've seen so far in regards to how SourceHut is being managed, indicate that you'd have such problems. (I'm saying this as a long time user of SoutrceHut and as someone that got into arguments with Drew)
Has that happened?
But the FAANGs do that to people all the time.
[0] (I have no idea if Drew DeVault would actually boot someone off his platform for arbitrary reasons, though. I'd hope that's not the case.)
If so, isn't that also "over the top ideological"? After all, we know Apple's opposition to the GPL dates back to the early 1990s when NeXT used gcc for Objective C, and their rejection of GPLv3 caused them to ship ancient versions of bash, recently replaced with zsh, so it's not like they weren't aware of the GPL.
Some people use Linux (the moral equivalent to "self-host") to avoid Apple's ideological stance.
Which, from a business standpoint, makes for a marketing opportunity for a FOSS-aligned source code hosting company. Which takes us from an ideological position to an economics one.
(To quote Bill Hicks: “Oh, you know what Bill’s doing? He’s going for that anti-marketing dollar. That’s a good market. He’s very smart.”)
https://todo.sr.ht/~sircmpwn/hare
The same goes for patches. You can submit changes directly via email, but you can also use our web interface. We're also working on web-based tools for code review built on top of our mailing lists, which will provide a complete experience similar to GitHub's pull requests, but built on top of the decentralized, standardized protocol that git was designed for.
That really seems the best way to dethrone GitHub, as it's moat is mainly carved via a social network.
The git repo is already decentralized, ForgeFed is a great way to integrate and decentralize the social part.
In the future gitea, srht, and many other implementations could interact with each other's users.
SourceHut is already federated: via email. Email is a standardized, federated protocol with thousands of implementations and a global fault-tolerant system already working in production for decades, and it's the approach git was designed to use. I think this is a better solution. You can already use SourceHut's tooling to contribute to projects like Linux, which don't use any forge software at all, thanks to this standardization.
Being able to bridge standalone gitea instances users with repos on your service would be a good step forward for federation.
It's pretty difficult to host a fediverse server, too. I want to make SourceHut easier to self-host post-beta, including dealing with email bits, but I also don't think the dream of federation involves everyone running their own servers. One of the advantages of the federation model is that it strikes a balance between total centralization and total de-centralization in that it allows a talented sysadmin to provide services for a large number of users. There are more ways to federate than ActivityPub.
As far as hosting difficulty goes, once Gitea supports ForgeFed, you'd be able to spin up your own instance as easy as running one program, and no worries about blacklisting.
I agree not everyone needs their own server, which is where shines sr.ht over Gitea, but a community developing a product sometimes wishes to run their own servers for their users. Sure they could run an email server, and users could use that, but it seems roundabout and outdated.
I agree there's more than one way to federate, but the goal is bridging everyone together. Those two aren't mutually exclusive if you support both protocols.
From the sounds of things he does not want the feature.
If you and several other people happen to have a hard requirement for a specific feature that he (or his buddy Simon) don't see fit for, you won't get that feature, even if you volunteer to implement and maintain it. The only thing you're left with is basically to fork SourceHut, host it yourself and maintain your feature all by yourself, dealing with continuously patching a very much still-in-development (and therefore ever changing) software. That is something you're probably not going to do, especially considering SourceHut's architecture and way of doing things.
SourceHut isn't exactly extensible/pluggable and hosting it as a one man show or even a small company becomes a huge PITA, as soon as you diverge from the holy grail that is Drew's way of doing things (Alpine, no containers, no good config management, no easy way to scale things, and the dedication to invest your blood and tears into maintaining this thing).
Hence I really cannot comprehend the current trend that is "let's all dump GitHub for this, and that, and SourceHut". So far, SourceHut really hasn't made an effort to prove itself worthy of the influx of OSS projects. And while I do see Drew commenting here, reassuring folks he won't ban anyone over any internet disagreement, reading the public mailing lists of the SourceHut repos doesn't really show much of a welcoming behavior either. I mean, he's the person behind what has become one of the most popular Gemini servers, and as soon as that was the case, he began threatening client apps to arbitrary block them for doing things that don't align with his values (in this case, [showing a favicon](https://github.com/makeworld-the-better-one/amfora/issues/19...)). And the cabal of elite internet Amish, that have been on SourceHut since its early days and that makes a large portion of the platform, aren't that different either.
I do agree with GitHub being the wrong place for OSS projects, but I don't agree with SourceHut being the right one. At least for as long as it doesn't become obvious that its founder and the community around him has changed and started to genuinely appreciate people for the work they're doing, regardless of their own ideological beliefs.
That being said, the devs of Gitea are in the process of adding AP support, and it has been mentioned on HN already: https://news.ycombinator.com/item?id=31934758
I understand and support his decision, but I will be going with Gitea because I'd prefer AP networking over email. To each their own for sure, but my main suggestion was to support both so the communities aren't fragmented, that would be a bigger threat to Github's network effect. Oh well.
I'll take another look.
The point I am trying to make here is - by default people spend their money with mega corps, because of brand awareness, convenience etc. Instead, it would be nice if the default is indie/small businesses.
In other words instead of "the mega corps aren't working for me, let me look for smaller/indie options instead", it would be nice if we all did "smaller indie options don't fit my requirements, let me go with Microsoft/Apple/etc instead".
Maybe this is too idealistic?
No but they will quit because an ideology most members of most teams was used to justify a decision that affected their work.
Let me start by telling you a story about me trying to buy clothes for the first time:
Back then, I have no idea of branding and the symbolism behind, I don't even have the idea of quality.
After many hours of browsing and comparing, I picked what I liked based on showcase pictures and user reviews, completely brand-blindly.
A week later when they arrived, I quickly discovered that few of them are false advertisments, some of them are in the wrong size and one got damaged during shipping. And for the rest items that were "just right", half of them ripped/deformed after half-year of use.
So, as a normal person would do, I created a table which consist of factors such as durability, price and styles etc to help me increase my "success rate". And after few years of iteration, that table ended up points me to just few brands that happens to be well-known and big.
I eventually settled on those brands with membership cards etc, and now whenever I need new clothes, I open their store page.
What I learned from my experience is that discovery itself generates cost, and the cost is nonrefundable. It is reasonable for an user to settle on a trusted brand. It's not brand loyalty, it's just rational decision making.
For most users, choose GitHub is rational, because it's the most trusted platform out there, and it provides good services (including the infrastructures and socials). You already knew what to expect before you even sign on, that's the level of trust GitHub delivers.
It's really hard to ask users to discover and move to the alternatives once their needs are well-fulfilled by the current one, even when the alternative provides the same level of service.
What makes it even harder is that services such as BitBucket, GitLab, SourceHut, Gitea etc can't provide the same level of service that GitHub offers for free. At this circumstances, the megacorp such as GitHub will always win even when the user is pro-indies by default.
But we are not completely dead in the water here. GitHub succeed because they created something new at the time, something better than what SourceForge offered.
I assume if someone wants to beat GitHub, they need to do the same. Don't just create alternatives, create something new that also does old things better.
Yes, you can say "well, your intended recipients shouldn't use mail providers that do that", but in reality that's not our choice.
I've had Google silently swallow emails before, both sending and receiving. It's not fun.
That was just one example, I've seen corporate Exchange systems do the same. Are they not professional?
If one of the biggest email providers on the planet can silently swallow and not deliver emails, then blanket statements saying emails never fail are just wrong. It's important to understand practical failure modes of the systems we use, and not pretend they're more perfect than they are in reality.
Because small, local is not always better. It's quite common to find coffee shops here in the PNW that make a horrendous espresso. Most roasters here in Seattle are pretty mediocre and only capitalizing on the Seattle reputation. I can rattle off all the "local" and "small-batch" bourbon/gin/vodka distilleries that use sourced alcohol and are over-priced train wrecks powered only by marketing and no skill. At least with Starbucks, I get consistently bad coffee. I'm not risking over paying for coffee that I spit out.
Definitely have been looking at st.ht but if github fulfills my needs better, I'm sticking with that.
That's exactly what happened to me at Starbucks. I had to wait for a small cookie shop to open up in order to buy an overpriced decent espresso. They had an Illy logo which helped a bit.
In Italy and Greece you specifically want to avoid big chains as almost every small shop makes better coffee.
Problem is that, I would say the majority of, small coffee shops in the US makes horrible espresso. Coffee isn't a craft to your average American no matter they claim, and they'll drink any swill. It's a way to get caffeine in you. Dark roast is the best selling roast.
If I go to a small coffee shop and they ask me what I mean when I order a doppio, that's when I start looking across the street for a Starbucks. At least Starbucks baristas are trained on what doppios are, and I'll get a consistently bad one. There's a very good chance that small mom and pop one will make me cigarette ash juice if they don't know what a doppio is. I'll take bad espresso over ash juice any day.
Interesting, does it? Obviously tastes differ, but I'm trying a tin at the moment because I ran out and could get it same day vs. waiting for delivery from my usual roaster subscription - I may as well be drinking any old thing off the supermarket shelf, don't find it superior at all.
Possibly the blend just isn't too my taste, it also doesn't seem to be as fine as I think I would grind it, but I think the main thing is freshness. Despite its vacuum seal it's still months old, vs. I usually receive beans within a few days of roasting and grind on demand.
If you want the only guaranteed probability about the future: its that most predictions about the future are wrong. Everything dies eventually; the things you least expect last longer than you expect them to; or sometimes much shorter. I do feel pretty confident that both Github and sr.ht will be operational tomorrow. But, not 100%; and even less than that next Tuesday; and even less by September.
Actually, Github has had 12 public incidents since May. So... maybe I shouldn't be so bold with my predictions about Github's operational status.
When people say "something is unlikely to happen" what they really mean is: I really hope that thing doesn't happen. Me, personally; I hope both Github and sr.ht continue to see success. As you say; me choosing to pay Github $5/mo won't change anything for the better or worse; but paying that to sr.ht, now that impact is much bigger. If you distil it down to the manifestation of your own hopes: I think the choice of who to support is obvious.
Say what you want about Microsoft, they do more than just search or advertising. If anything you are interested in from Google isn't search or advertising related, don't count on it lasting. The only people surprised by Wave and/or Wunderlist getting canned are people who haven't been realistically looking at who Google is and how they operate. The worst IT contracts I have ever had to administer are with Google - they are a complete joke.
As for "incidents" - lol - who doesn't have them? What's more important is how does the org in question respond to them. You seriously think sr.ht is going to have the same resources as github/Microsoft? Then again it's a dramatically simpler platform so that's a huge advantage as far as that goes - but just look at this thread; as they add more features the chances for incidents goes up.
Github has massive market share and may get expensive but is unlikely to die.
And that would worry people, sure, new owners may not look after it, etc., but there'd be plenty of time and warnings to move off it.
https://sourcehut.org/blog/2022-04-08-2021-financial-report/
SourceHut is profitable and sustainable, thanks in no small part to the fact that we expect users to pay if they have the means. Small companies providing free services with investor subsidies are much more likely to disappear than a company with a service you're paying for which isn't beholden to any investor demands for infinite growth.
Hi Drew
I’d encourage you to start charging then based on your comment above.
Because sourcehut pricing page currently says “payment is currently optional”.
You’ve been in alpha for a few years now and I’m sure many of us would agree it’s totally appropriate for you to start officially charging based on the quality of your service.
IIUC your report, you are paying developers (and yourself) only USD $30,000/year. The typical software developer would make at least three time that much elsewhere. With 3x salary, I'd say SourceHut would no longer be profitable or sustainable. Or am I missing something?
https://sourcehut.org/blog/2021-08-23-work-at-sourcehut/
The sustainability metrics are based on revenue covering the small base compensation plus operating expenses, as well as having a large enough standing balance to weather any unexpected events. We seek to make models which guarantee the longevity of the platform, and prioritize making sure that our cash flow will always provide for operating expenses. If our revenue dropped to zero today, we would be able to run the platform for several years.
Ultimately, everyone involved in SourceHut is there because they believe in free software and believe in the model we've designed to further its cause. Our non-negotiable ethical standards calls for a creative business model to make this thing work.
The work experience and impact we offer is something that is, in my opinion, completely unmatched in the industry. It's not for everyone, but we believe in what we're doing, and it's worth taking a gamble on an unproven business model.
Github is about 1000x more user friendly?
And the big ones don't ever do this? You're probably shielded from the jira ticket backlogs that are thousands deep.
While there's no hard and fast rule, small companies tend to care more, as they're closer to the pain. A loss of a customer isn't just "churn".
We should strive to have at least four to six alternatives for every type of system. It keeps pricing competitive, gives users lots of alternatives, and prevents single players from strong arming behavior.
Right now there are too many monopolies and duopolies in tech.
Remember OVH server room fire? Was there any failure of such magnitude in AWS/Azure/GCP?
It's hardly surprising that low cost options are more prone to failures in many markets.
Why not buy the best product at the price you're willing to pay for it, even from a company with the values you share? Where does company size take precedent?
If you don’t make an effort to try smaller company’s products, you’re going to be biased towards bigger companies by virtue of them having better marketing and network effects. Some conscious preference to smaller companies can help counteract this.
And even if the smaller company has a worse product initially, some people using it anyways can help keep them alive in order to fund innovation. Smaller companies tend to be more willing to experiment or innovate, but that can’t happen if no one is willing to make a long-term bet.
> A very rich, multi-trillion dollar asshole-ish corporation, that feels lonely on it's birthday and goes on a bender buying less rich and asshole-ish companies to successively kill them, to calm it's mood.
Not sure how true that is, but something tells me it might just be the reality.
Having said that, I still buy a lot of things from Amazon because it's much more convenient than the alternatives. It's a habit I've been trying to kick, but I have a ways to go.
Of course a homogeneous world is less interesting. And anti-competitive behaviors are net bad and need to be disincentivized. Big company hate just seems misplaced.
We hold ourselves to a higher standard than most and we have a specific and documented set of goals to accomplish before we declare the alpha complete. However, any potential user who wants to use SourceHut today can find a wealth of information to secure peace of mind in their choice.
In the end it's an ideals tradeoff. Are you going to prioritize the growth of your project ecosystem but give into megacorps, or are you kneecap contributions by choosing esoteric ethical workflows.
Many a would-be contributor has lost interest when a project was entirely dependent on github, too. It's pretty telling when a contributor uses git, but can't contribute to a project that also uses git.
Small businesses usually treat their employees worse than chains do (because "we're all a family"). Also… coffee shops seem to have big opinions of themselves but the farmers do all the work. Baristas just have value unfairly credited to them by the customers because the product is addictive.
Why support Github when there is st.ht? Why support Starbucks when you can support a mom & pop coffee shop? Why buy books from Amazon when you can buy from a Indie book store? Granted, not everyone has the patience, time and money to vote with wallet, but we can try wherever possible. I hope more people ditch Github