6 ms·

I'd like to see a project like this which is self-hosted, where the process of buying a domain, acquiring hosting, etc. is made easy for non-technical users. Th

by _9xrb 4y ago

I'd like to see a project like this which is self-hosted, where the process of buying a domain, acquiring hosting, etc. is made easy for non-technical users. The problem with sites like about.me is that eventually the maintainer gets acquired, gets bored, gets hit by a bus, etc. With a self-hosted site, it works as long as you keep paying Namecheap and AWS, for example. This arrangement also separates the core project owners from responsibility for whatever users slap on their page down the line. I imagine users would go to one site (or app) to get it all running, and that site would hook them up with hosting and a domain name behind the scenes, maybe get them set up with a static page through a WYSIWYG editor, then boom done. Maybe give users a standalone app with their hosting keys and credentials that lets them update the content on their site directly (again through WYSIWYG, like Adobe Dreamwaver but easier), without going back to the initial setup site. A platform or service that handholds users through initial setup and billing, and then steps away once they're up and running, seems to me like the right mix of centralized/devolved control and responsbility. I realize what I'm describing sounds a lot like Wordpress or Dreamwaver, but I think there's a big opportunity to slim down these offerings into a simple, secure, easy-to-use package. I guess it depends on your choice of target audience.

philip1209 4y ago
Email sending is surprisingly hard to get right. That's where a lot of the value for a hosted product comes from. With Postcard - I've had to add various checks and protections to prevent bot abuse (and to manage sender reputation). It would be really hard to be reactive like that on a self-hosted product.
- deepdriver 4y ago
  Is it more or less the same problem as running a GNU Mailman listserv in 2022? That is to say, a fairly hard problem of staying on top of security patches, avoiding getting marked as spam, and avoiding getting put on the big email provider's badlists. In theory, the problem seems simple: replace a manually-managed BCC list with an alias to said list, so it's less hassle for the user/sender. Maybe the answer is an email client plug-in that just autofills BCC to-addresses? Assumption would be the user's sending this email from a known provider like GMail or Fastmail. I imagine it would be easier for the completely self-hosted email system if signup gave receivers a from-address they could whitelist, if their email provider allows it. Maybe you already do this. Edit: I'm also assuming email is sent one-way to whoever signs up, maybe with a link to the dynamic chat server/app/platform of the moment. Slightly different from a GNU Mailman list, which enables back-and-forth conversation.
  - philip1209 4y ago
    Yes, a little bit. I'm working on a different project that's more of a refreshed Google Groups / GNU Mailman listserv, and that has even more complexity for allow-lists: https://booklet.community https://booklet.community > I imagine it would be easier for the completely self-hosted email system if signup gave receivers a from-address they could whitelist, if their email provider allows it. Maybe you already do this. Yeah, that could be possible. I'm not opposed to releasing a self-hosted version of Postcard - but it's been useful to not have that constraint from the beginning. For instance, email sending happens in an async queue backed by Redis - and that might just be unnecessarily complex for a self-hosted tool!
x-complexity 4y ago
> I'd like to see a project like this which is self-hosted, where the process of buying a domain, acquiring hosting, etc. is made easy for non-technical users. The problem with sites like about.me is that eventually the maintainer gets acquired, gets bored, gets hit by a bus, etc. With a self-hosted site, it works as long as you keep paying Namecheap and AWS, for example. This arrangement also separates the core project owners from responsibility for whatever users slap on their page down the line. Right from the get go, long-term maintainability of such a platform is very hard, as there are now a few dependencies that you're reliant upon for your self-hosted site: - Domain registrar - Infrastructure provider (IaaS) - DNS record management Trying to make the providers swappable will be a challenge, as APIs are not consistent across the ecosystem. This is compounded when trying to make the solution turn-key, as now there needs to be a developmental effort behind the project to keep it up to date, & to add in new options to swap the providers out. > I realize what I'm describing sounds a lot like Wordpress or Dreamwaver, but I think there's a big opportunity to slim down these offerings into a simple, secure, easy-to-use package. The pennies that can be picked up in this specific niche is too small to bother: The ones that just want a website without any frills will stick to WordPress & their providers, while those wanting extreme configurations & customizability will just run their own language/framework on the infra of their own choosing. The niche mentioned sits in between the two, with the extreme ends increasing their grasps: WordPress plugins from the easy end, and web frameworks from the other end (Vue, React, Svelte, Django, etc.). **This kind of thinking, however, misses the forest in favor of nostalgia: The vast majority of modern Internet audiences want fast & easy discoverability/search, along with easy public sharing of posts, & account follows. Modern social media is what results from the inevitable result of catering to the market's demands. Such market wants also inevitably creates gravitational pulls towards platforms with large audiences, creating a feedback loop that's hard to break.** Right now, there's a weak-but-present desire to decouple from Big Social Media (Mastadon, Lemmy, Friendica), but the wants of the modern audience run against the ideals of self-hosted social media taking off: Their customizability is their own Achilles Heel, with the technical & financial limitations of scaling self-hosted services making it near impossible to compete with the big guys.

by _9xrb 4y ago

philip1209 4y ago

Email sending is surprisingly hard to get right. That's where a lot of the value for a hosted product comes from. With Postcard - I've had to add various checks and protections to prevent bot abuse (and to manage sender reputation). It would be really hard to be reactive like that on a self-hosted product.

deepdriver 4y ago
Is it more or less the same problem as running a GNU Mailman listserv in 2022? That is to say, a fairly hard problem of staying on top of security patches, avoiding getting marked as spam, and avoiding getting put on the big email provider's badlists. In theory, the problem seems simple: replace a manually-managed BCC list with an alias to said list, so it's less hassle for the user/sender. Maybe the answer is an email client plug-in that just autofills BCC to-addresses? Assumption would be the user's sending this email from a known provider like GMail or Fastmail. I imagine it would be easier for the completely self-hosted email system if signup gave receivers a from-address they could whitelist, if their email provider allows it. Maybe you already do this. Edit: I'm also assuming email is sent one-way to whoever signs up, maybe with a link to the dynamic chat server/app/platform of the moment. Slightly different from a GNU Mailman list, which enables back-and-forth conversation.
- philip1209 4y ago
  Yes, a little bit. I'm working on a different project that's more of a refreshed Google Groups / GNU Mailman listserv, and that has even more complexity for allow-lists: https://booklet.community https://booklet.community > I imagine it would be easier for the completely self-hosted email system if signup gave receivers a from-address they could whitelist, if their email provider allows it. Maybe you already do this. Yeah, that could be possible. I'm not opposed to releasing a self-hosted version of Postcard - but it's been useful to not have that constraint from the beginning. For instance, email sending happens in an async queue backed by Redis - and that might just be unnecessarily complex for a self-hosted tool!

x-complexity 4y ago

> I'd like to see a project like this which is self-hosted, where the process of buying a domain, acquiring hosting, etc. is made easy for non-technical users. The problem with sites like about.me is that eventually the maintainer gets acquired, gets bored, gets hit by a bus, etc. With a self-hosted site, it works as long as you keep paying Namecheap and AWS, for example. This arrangement also separates the core project owners from responsibility for whatever users slap on their page down the line. Right from the get go, long-term maintainability of such a platform is very hard, as there are now a few dependencies that you're reliant upon for your self-hosted site: - Domain registrar - Infrastructure provider (IaaS) - DNS record management Trying to make the providers swappable will be a challenge, as APIs are not consistent across the ecosystem. This is compounded when trying to make the solution turn-key, as now there needs to be a developmental effort behind the project to keep it up to date, & to add in new options to swap the providers out. > I realize what I'm describing sounds a lot like Wordpress or Dreamwaver, but I think there's a big opportunity to slim down these offerings into a simple, secure, easy-to-use package. The pennies that can be picked up in this specific niche is too small to bother: The ones that just want a website without any frills will stick to WordPress & their providers, while those wanting extreme configurations & customizability will just run their own language/framework on the infra of their own choosing. The niche mentioned sits in between the two, with the extreme ends increasing their grasps: WordPress plugins from the easy end, and web frameworks from the other end (Vue, React, Svelte, Django, etc.). **This kind of thinking, however, misses the forest in favor of nostalgia: The vast majority of modern Internet audiences want fast & easy discoverability/search, along with easy public sharing of posts, & account follows. Modern social media is what results from the inevitable result of catering to the market's demands. Such market wants also inevitably creates gravitational pulls towards platforms with large audiences, creating a feedback loop that's hard to break.** Right now, there's a weak-but-present desire to decouple from Big Social Media (Mastadon, Lemmy, Friendica), but the wants of the modern audience run against the ideals of self-hosted social media taking off: Their customizability is their own Achilles Heel, with the technical & financial limitations of scaling self-hosted services making it near impossible to compete with the big guys.