Don't build your castle in other people's kingdoms • Hacker News Board

Re: Don't build your castle in other people's kingdoms

#111

Post by teddyh » Thu, Nov 04, 2021, 9:09 PM UTC

riidom wrote: post #13

teddyh wrote: post #6

One thing about domain names: Choose both a registrar and top-level domain for your domain name carefully. Neither your registrar nor your chosen TLD registry should be in the habit of suspending domains at the drop of a hat, or be at risk of going out of business suddenly. For TLDs, I have said before¹ that if you mostly trust your local government, your national ccTLD should suffice. In fact, it should be your defa…

And to add to that (personal experience): If you buy everything from one hand at once (server space, domain, associated email), make sure you have a way of changing provider and keep domain and emails without it being a hassle.

Word. If you have regular backups, and if some downtime is not really a problem, it might be fine to use web server hosting, e-mail (and in extreme cases even DNS hosting), from some fly-by-night el cheapo provider. But your domain name registrar? Pick them carefully, don’t skimp, and make sure they have good support. Because when things go pear-shaped, you really want to be able to actually talk to someone to change your web server or e-mail DNS records (or even DNS servers) to somewhere else.

Full disclosure: I work at such a registrar. No, you’re probably not in our target market.

Re: Don't build your castle in other people's kingdoms

#112

Post by allemagne » Thu, Nov 04, 2021, 9:20 PM UTC

Karrot_Kream wrote: post #105

allemagne wrote: post #100

Earlier quoted context omitted.

Calling this article a "libertarian fantasy" seems unfair. The author does explicitly recommend entertaining new platforms, but being prepared with a backup plan for when these platforms are no longer viable for whatever reason. Where does it end? They call out these options as presumably being "safe enough": - A website on a domain you own - A mailing list - Own and license your Intellectual Property - Sell your mer…

> Where does it end? They call out these options as presumably being "safe enough": Is there any data around why these are safer than the alternatives? Because the article is just a classic case of nerd philosophy; there's _no_ justification of the actual risks involves in these platforms aside from vague anecdotes and analogies, then using these vague anecdotes to drive recommendations.

I don't know, I think you can go ahead and use your best judgment on whether it's riskier for an indie game studio to go all-in on marketing through a social media juggernaut or trying to steer people to channels like a mailing list that they control. This is not a research paper.

I'm sure there's lots to criticize in "nerd philosophy" but an article that boils down to "avoid vendor lock-in" with a cute analogy is just a weird target to choose. If you can't help but read a certain ideological slant onto it and can't possibly understand it separate from that reading, I think that says more about you.

Re: Don't build your castle in other people's kingdoms

#113

Post by okhuman » Thu, Nov 04, 2021, 9:33 PM UTC

josephcsible wrote: post #79

okhuman wrote: post #49

This message applies to your repo on Github too.

Not nearly to the same extent, though. It's easy to move your repo over to GitLab or Savannah or something if need be.

it's not git thats valuable, it's all the social interactions around it - your issues, discussions, actions would be very hard to move. Github is only releasing more vendor lock in features as time goes on.

Re: Don't build your castle in other people's kingdoms

#114

Post by marc_io » Thu, Nov 04, 2021, 11:00 PM UTC

9dev wrote: post #51

antonzabirko wrote: post #37

A lot of people need to see this. If you don't control something, you will be disappointed eventually. Philosophically this can even be extended to your mortality.

Ah, but this brushes an interesting, yet inevitable fact: there is no such thing as control. We tend to imagine to have control to various degrees, as that makes it easier to cope with the chaos around us. Still, we have to accept the fact that we’re all going to die, eventually, for example. Many people never do that; they simply pretend they’re immortal, until one day something bad happens and shakes them awake. I…

True, but we always have some sort of agency over this.

Re: Don't build your castle in other people's kingdoms

#115

Post by marc_io » Thu, Nov 04, 2021, 11:13 PM UTC

Karrot_Kream wrote: post #105

allemagne wrote: post #100

Earlier quoted context omitted.

Calling this article a "libertarian fantasy" seems unfair. The author does explicitly recommend entertaining new platforms, but being prepared with a backup plan for when these platforms are no longer viable for whatever reason. Where does it end? They call out these options as presumably being "safe enough": - A website on a domain you own - A mailing list - Own and license your Intellectual Property - Sell your mer…

> Where does it end? They call out these options as presumably being "safe enough": Is there any data around why these are safer than the alternatives? Because the article is just a classic case of nerd philosophy; there's _no_ justification of the actual risks involves in these platforms aside from vague anecdotes and analogies, then using these vague anecdotes to drive recommendations.

I've seen many examples here in HN about people who have lost access to one of these kingdoms for no good reason or as a result of some weird algorithmic decision.

Re: Don't build your castle in other people's kingdoms

#116

Post by sneak » Fri, Nov 05, 2021, 2:36 AM UTC

rambambram wrote: post #30

sneak wrote: post #12

Earlier quoted context omitted.

Sadly you can't build apps that can send notifications on the most popular phone on the planet without subjecting yourself to arbitrary censorship by Apple and a 30% cut of revenues. You also can't build apps that use decentralized backends that receive notifications on that platform. All notifications have to come from the centralized app developer. Sharecropping is beginning to be the only way.

I did hold that belief for some time, but I solved it with the most reliable notification-method I came across thus far, and that's plain simple email from my own domain. I understand this is not possible for specific apps that need smartphone notifications to function properly, but when you think about it: a lot of (web)apps don't need that. In general, one might need less notifications than one thinks.

As a vendor, notifications sell more. As someone who sells things, you need to outsell your competition. Therefore, as a vendor, you need more notifications than you think.

Re: Don't build your castle in other people's kingdoms

#117

Post by LeifCarrotson » Fri, Nov 05, 2021, 2:39 AM UTC

frosted-flakes wrote: post #65

LeifCarrotson wrote: post #22

Earlier quoted context omitted.

I'd argue that on today's Internet, if you don't have a .com domain, you do have an exotic/fancy TLD and therefore need to have the .com domain live and redirecting to your exotic TLD.

Your local ccTLD is the exception.

.us? That sounds like some .gov phishing scam or del.icio.us domain name hack. Sites using .us are the exception.

Re: Don't build your castle in other people's kingdoms

#118

Post by frosted-flakes » Fri, Nov 05, 2021, 3:55 AM UTC

LeifCarrotson wrote: post #117

frosted-flakes wrote: post #65

Earlier quoted context omitted.

Your local ccTLD is the exception.

.us? That sounds like some .gov phishing scam or del.icio.us domain name hack. Sites using .us are the exception.

Except .us because the USA has so many other common TLDs commonly associated with it. In Canada, .ca is the norm, and I tend to associate .com with $USD and expensive shipping.

Re: Don't build your castle in other people's kingdoms

#119

Post by Karrot_Kream » Fri, Nov 05, 2021, 5:58 AM UTC

marc_io wrote: post #115

Karrot_Kream wrote: post #105

Earlier quoted context omitted.

> Where does it end? They call out these options as presumably being "safe enough": Is there any data around why these are safer than the alternatives? Because the article is just a classic case of nerd philosophy; there's _no_ justification of the actual risks involves in these platforms aside from vague anecdotes and analogies, then using these vague anecdotes to drive recommendations.

I've seen many examples here in HN about people who have lost access to one of these kingdoms for no good reason or as a result of some weird algorithmic decision.

I've seen people lose their data on their own blog, forget to renew their domain name and have it takes over by squatters demanding ransom, had their machines hosting Mastadon die on them leaving users high-and-dry, exposing keys to the world and getting their database stolen/dropped... The list goes on. It's not as simple as "kingdom bad, self-reliance good". Hence asking for a cost-benefit analysis.

Re: Don't build your castle in other people's kingdoms

#120

Post by teddyh » Fri, Nov 05, 2021, 12:16 PM UTC

LeifCarrotson wrote: post #117

frosted-flakes wrote: post #65

Earlier quoted context omitted.

Your local ccTLD is the exception.

.us? That sounds like some .gov phishing scam or del.icio.us domain name hack. Sites using .us are the exception.

Your local ccTLD might not be .us. Not everyone lives in the US.