Hacker News Reader

Comment by azinman2 | Hacker News Reader

I just don't see a big market for hosting FOSS apps. There are few apps, if any, that are better than a commercial equivalent already on web/iOS/android with support teams, marketing, etc. Ordinary people don't have much qualms about using a commercial service despite all the privacy warnings in the world: just look at the 1B+ users on Facebook.

So if we're left with a small crowd who does care, they're also largely the same crowd who feels comfortable getting a DO droplet and apt-get installing whatever app.

I fee bad for the sandstorm guys, seems like they put a lot of energy into it, but they approached it as an engineering challenge rather than from a market research "build what people want" challenge.

View on HN

It's definitely true that we've been more focused on technology than market research, and that this is a huge factor in why our business didn't work.

However, I don't believe you can really make revolutionary changes based on the "lean startup", "do lots of market research and test everything with metrics" strategy. It's absolutely a great way to make incremental improvements -- even big ones -- but not paradigm shifts.

Sandstorm's vision is a long-term one, and it actually isn't primarily focused specifically on self-hosting, privacy, or even FOSS, but rather on creating infrastructure that allows decentralized software to stand on equal or greater footing compared to centralized services. There is a lot of work that needs to be done for this to function, and you can't justify it by saying "look, these customers asked for it" -- you justify it by laying out the vision and saying: "Look, there will be these clear enormous advantages if this works."

For reference, here's our technology manifesto: https://sandstorm.io/how-it-works

This is always a tough sell, because people rarely agree on hypothetical outcomes that can't be measured in advance. And if it were clear, someone would be doing it already. So, I don't expect you to agree. But I'm going to keep working on it.

"allows decentralized software to stand on equal or greater footing compared to centralized services."

While that's great from a CS/FOSS/EFF/hacker perspective, the question is what's necessary for such software to be on equal ground in the eyes of ordinary users? My guess is that the decentralized/centralized split isn't (yet?) it, but rather the UX and functionality. Few open source end-user apps are entirely original and cutting edge; most are poor knockoffs of commercial products or are failed commercial products that got open sourced.

To me that's why sandstorm didn't make much sense. I applaud your efforts, I really don't want to rain on your parade -- I poured my sweat and tears into a startup that failed as well so I get it -- I'm just reacting to what seemed like not-honest-enough reasons for failure on the website. It's really important to know what didn't go right for next time lest you make the same mistakes again.

It sounds like you're assuming that Sandstorm is a platform strictly for open source software, but that was actually not our intent. In the ten-year vision, there is a thriving ecosystem of both open source and proprietary / paid applications that build on the platform because it provides many advantages both to developers and to users, allowing for better-quality apps.

We've used open source apps to seed the app market, because we can do that without the upstream developers' help. We also are big fans of open source ourselves, obviously, and I feel open source is especially disadvantaged in a SaaS world, so Sandstorm will make it more competitive. But in the end what I really want is high-quality decentralized software in general.

> However, I don't believe you can really make revolutionary changes based on the "lean startup".

You know more than me, but isn't of the key ideas of the Lean Startup to validate your assumptions, which everyone has? Write them down and validate them ASAP to de-risk your project. For Sandstorm some could be:

- Developers are willing to bet on an unproven company/project/platform.

- Users are willing to create Sandstorm accounts.

- Developers are willing to build on top of Sandstorm rather than owning all the data and customer relationship.

But +1 for "people rarely agree on hypothetical outcomes that can't be measured in advance." Well said.

Open source succeeds not only when it's open, but when that openness also just makes it better than using the closed alternative. For example, I was able to send my friends a single Sandstorm sharing link that contained the equivalent of a bunch of Google Docs (Etherpad) and Trello boards (Wekan) in one easy location. Because Google Docs and Trello are closed source, separate platforms, I can't do that. But I can with Sandstorm.

A lot of what Sandstorm is aiming to do is still in development feature-wise, and then people have to build apps on top of that. So I'd say this is a long haul destination here. But the key point is: Open won't win because it's more private, or more free. It'll win when it's better.

One of my key worries with the current walled garden approach is exactly what you describe. Value can't be easily reached by connecting products together. Every major company seems to be working hard to keep there customers completely inside there ecosystem. It makes sense for them and I certainly don't blame them, but it does seem to represent some level of failure for the current system.

I think Android is a great example of how providers can realize that a common, unified platform has huge advantages to their business. Of course, I'd argue the largest problem with Android is the control Google exerts over it. But, as far as the manufacturers go, it's a great example of seeing the benefits of a common platform.

My hope is someday everyone will use Sandstorm (I'm an optimist), and maybe you'll use Sandstorm on Google's servers or Apple's servers or whatever but you can all access the same apps everywhere, whether you're in someone else's cloud or hosting your own.

Universities would seem like a good target market to me. There's privacy requirements (student grades shouldn't leave the premises!), there's a need for fine-grained access control, there's a need for easy to use file sharing, calendaring, and for hosting faculty- or student-developed apps in an environment where it's hard to impose a single OS or enterprise-y software.

Universities also typically have multi-million dollar IT budgets and larger institutes have dedicated IT staff with >5 employees that manage the email-servers, computer pools etc. It is not a stretch for them to also offer other services like a mattermost installation or a gitlab installation and indeed that happens.

I agree. In fact, some of our biggest fans seem to be "ed tech" users, especially outside of the US.

They're notoriously hard to make money from, though. Luckily that's not a concern for us anymore. :)

That's certainly fair, but at the same time as somebody who does care about this kind of stuff I deeply wish we didn't live in this world of massive centralized walled gardens.

I hadn't heard about this project before, but for a long time I've been thinking it would be nice to have some form of sand-boxed, probably Node based local cloud system.

I think you under-estimate the difficulty to install services locally. Sure popular packages are usually easy, but things with even a little bit less support can take days to get right.

> I think you under-estimate the difficulty to install services locally. Sure popular packages are usually easy, but things with even a little bit less support can take days to get right.

This. Generally it's people who don't self-host that usually claim that web apps are just a 'apt-get install' away.

Ordinary people don't have much qualms about using a commercial service despite all the privacy warnings in the world: just look at the 1B+ users on Facebook.

That's changing, more people are keep their mouths shut more often on social media because they know their data is being scooped up.

I think it's a matter of educating more customers; they have no idea that it's even possible to host their own Google or Facebook in some instances.

Maybe it isn't a big market but the market does exist and it does require more customer education and awareness raising. It's a harder sell than enterprise sales.

Reply on news.ycombinator.com