Live data from Hacker News

Local-first software (2019)

inkandswitch.com

81–90 of 136 posts

Re: Local-first software (2019)

#81
post #56

I think there's probably a market for a personal cloud, which probably sounds dumber than personal computer did in the 80s. What I mean by that is a computer somewhere in the garage, like a furnace, with enough compute and storage to drive all the devices and appliances in a house. In this model, devices do not have CPUs or memory, only input/output and a network chip. The way this would work is for the computer in t…

I would love a hardware/software solution that makes it easily to backup my data (ideally with integrations to Google Takeout, Facebook, etc.). Perhaps it exists already?

Edit: of course local-first does not mean merely "backup", but instead the (redundant) hardware serves as a primary data store. I would welcome that as well!

Re: Local-first software (2019)

#82
post #69
post #62

Earlier quoted context omitted.

Why not?

Short answer: generally it costs more money for worse results. Longer answer: doing things in-house that are outside your core competencies and/or value creation model is a poor use of scarce resources (both capital and human—predominantly management bandwidth) and increases risk carried. To give a concrete example: imagine you need to host your source code repository. You can pay for something like Bitbucket for $6/…

At my last job, we hosted our own git repo with gitolite. That doesn't give you a ticketing system of course, but it was fairly easy to administer. And it sat on a miscelanous jobs server we already needed and didn't use up too much resources.

Remote access followed production norms, so no extra work there (other than a lot more people need access to git than other production servers). Maybe a few hours one time to lock down permissions for git etc, probably less fuss than getting SSO setup for a git SaaS.

Re: Local-first software (2019)

#83
post #41

Earlier quoted context omitted.

Many far smarter people have said this before, far more eloquently than I can, but in short: Cloud Computing (and SASS even more so) is little more than just another attempt to recreate access/info monopolies, essentially the same profit proposition as existed with closed source software, while pretending to be one of the cool kids and use politically more acceptable (but in this context rather meaningless) terms lik…

>Cloud Computing (and SASS even more so) is little more than just another attempt to recreate access/info monopolies, [...] , everyone will blame anything and everything but their own willful ignorance, For some reason, "cloud computing" has become a bogeyman and therefore corporations paying for it are clueless "sheeple". To help prevent the phrase "cloud computing" from distorting our thinking, we have to remember…

That's a very good point, thanks for bringing it up, with all the examples to back it up.

I agree that "cloud" is becoming a thought-terminating cliché (and I myself fall prey to reflexive reactions to that word all too often). I think the core issue is something different - it's about relationships. It's a point of view I've been exploring recently, and I find it to be illuminating. Some thoughts:

1. Individuals do not handle relationships as well as companies. That's the source of "subscription fatigue" - you end up having to mentally keep track of all the third parties you may or may not owe money to, that may or may not do something that impacts your life down the line. Companies have dedicated people and processes for handling this, most individuals do not (rich people can hire help).

2. Individuals do not necessarily want all these relationships. If I buy a toaster, I don't want any relationship with its maker[0] - for the same reason that when I buy bread from a grocery store, I don't want to enter into relationship with the seller, or the baker, or the farmer that provided the flour[1].

3. Power imbalance in relationships matter. Individuals are almost always irrelevant to the service provider, so for a regular person it's better to not have any relationship at all. For companies, it depends. A corporation like Microsoft or Google can be rather certain that Salesforce isn't going to pull a fast one over them, due to relative sizes and the amount of money changing hands. Smaller companies fall somewhere on the spectrum between individuals and megacorps.

4. The above colors risk calculation. If my relationship with the service provider is closer to that of peers, I can feel safer about depositing data with them or making myself dependent on their service, because they're incentivized to provide a good service to me, and I can hurt them if they don't (say, by switching to a competitor). Companies do the same calculus. A small studio is best to think twice about depending on third-party services for anything that may outlive such services. A large company can derisk the relationship through a contract.

5. Part of the objection I and many others have to SaaS is that consumer-oriented SaaS (including "stuff as a service", aka. turning products into services) tends to be exploitative by design. You wouldn't consider a friend a person that abuses you mentally or tries to cheat you out of your money, so why enter into a relationship with a company that tries the same? Except when you have no other choice, which is why everything becoming a service is a worrying trend.

--

[0] - Beyond the one managed by consumer protection laws. But this kind of relationship is something you do not need to mentally keep track of - it doesn't come with any unexpected consequences, and the terms of relationship are shared (they're part of the law).

[1] - Again, beyond the implicit ones fully handled by consumer protection and health&safety laws.

Re: Local-first software (2019)

#84
post #37

Earlier quoted context omitted.

The same way people federated email forever? Local servers that contact upstream servers?

My question is not "how does federation work?" It is "in what situation is federation useful for video conferencing in a way that a single open source server is not?"

Depending on the network topology and number of callers, you could get some benefit from a multiple server model. Large meetings across offices could make use of a server at each office to streamline bandwidth. Not a good fit for today's reality of everyone from home or your scenario of 10 callers. Could still be useful for EU vs US callers to connect locally and only one stream across the atlantic. Multiparty e2e calling is already challenging, and adding multi-hop routing might make it more so, though.

Re: Local-first software (2019)

#85
post #56

I think there's probably a market for a personal cloud, which probably sounds dumber than personal computer did in the 80s. What I mean by that is a computer somewhere in the garage, like a furnace, with enough compute and storage to drive all the devices and appliances in a house. In this model, devices do not have CPUs or memory, only input/output and a network chip. The way this would work is for the computer in t…

I wonder if there would be market for community clouds, or neighborhood computes? Imagine that a new apartment building comes bundled with a server room in the basement. Every dweller gets compute/storage there. This could serve as edge cache for services like Netflix/YouTube, as well as for the ecosystem you describe.

I once imagined that homomorphic encryption would allow people to store data in their personal/neighborhood clouds and have third party SaaS code operate on that data locally. But I've recently been made to understand that homomorphic encryption would also allow companies to fully close off any access to data beyond what a program/service wants to give out, and unfortunately I get the feeling that the market will prefer the latter over the former.

Re: Local-first software (2019)

#86
post #80
post #75

Earlier quoted context omitted.

>The key difference with this current cloud embrace is that big billion dollar businesses are making their core business dependent on Amazon/Google/Microsoft cloud services. There's still some history missing there. Before AWS/GCP/Azure, billion dollar businesses were making their _core_ business dependent on datacenters owned by HP/IBM/Sungard/Origin/Exodus/etc. E.g. Google's early search engine (their core business…

Exodus and similar colocations aren't really the same thing. They're selling you rack space with connectivity, you bring the servers. If you outgrow that, or they go out of business, you 'only' need to get connectivity to a new facility and drive your servers across town. If you outgrown IBM/AWS/etc, you need to get connectivity and servers and run any services you were relying on. That's a lot bigger dependency.

While you can bring your own servers most of the time you rent those servers from them. You do this so that you can get fast turnaround on warranty replacements or repairs. You can use on-site hardware support and you get a lot of other benefits. When you think about it it's a lot like AWS except AWS offers those benefits even better than the old school places.

You can have your own servers but it's at the cost of slower turn around on hardware issues.

Re: Local-first software (2019)

#88
post #76

Earlier quoted context omitted.

So the end result was that it wasn’t effective. Imagine that, ineffective government regulations with regards to tech. Who would have thought?

So the end result should be that we dont try?

More regulation actually increases the barrier of entry to make it even harder for new companies. Well financed established companies are well able to comply. There is a reason that FB always begs for regulation.

Re: Local-first software (2019)

#89
post #82
post #69

Earlier quoted context omitted.

Short answer: generally it costs more money for worse results. Longer answer: doing things in-house that are outside your core competencies and/or value creation model is a poor use of scarce resources (both capital and human—predominantly management bandwidth) and increases risk carried. To give a concrete example: imagine you need to host your source code repository. You can pay for something like Bitbucket for $6/…

At my last job, we hosted our own git repo with gitolite. That doesn't give you a ticketing system of course, but it was fairly easy to administer. And it sat on a miscelanous jobs server we already needed and didn't use up too much resources. Remote access followed production norms, so no extra work there (other than a lot more people need access to git than other production servers). Maybe a few hours one time to l…

Now repeat that for your mail server, your expense tracking system, your HR system, your project manager system, various collaboration software....

And what happens when that one server goes down or becomes overloaded?

Post reply on HN