This doesn't sound so great. Unlike phone numbers you can't "port out" an IP address. If you purchase a block (at least 256 IPs) then you can do BYOIP (bring your own IP) with various cloud...
This doesn't sound so great. Unlike phone numbers you can't "port out" an IP address. If you purchase a block (at least 256 IPs) then you can do BYOIP (bring your own IP) with various cloud providers. But that's thousands of dollars per year. This means your "web number" is permanently tied to one cloud provider. Not ideal but not so bad with larger and more stable (but more expensive) providers like AWS. But this could get annoying with smaller and cheaper hosts that might go out of business or prove to be more trouble than they are worth.
Yeah, not to mention that the small web better remain pretty small. If your website gets bigger, it would be seriously challenging to host it off one IP address.
Yeah, not to mention that the small web better remain pretty small. If your website gets bigger, it would be seriously challenging to host it off one IP address.
I'd have to go check what the current status of IPV6 allocations are, but that's the only way it would be feasible. I coordinated my last orgs IPV6 allocation and it wasn't too stingy. So maybe...
I'd have to go check what the current status of IPV6 allocations are, but that's the only way it would be feasible. I coordinated my last orgs IPV6 allocation and it wasn't too stingy. So maybe the small web guy could submit an allocation plan and get a big fat range.
As much as I find the project neat, I don't think he is solving the problem he thinks he's solving. Or if he is, it isn't clear from the little digging I've done. And Honestly, I think it is...
As much as I find the project neat, I don't think he is solving the problem he thinks he's solving. Or if he is, it isn't clear from the little digging I've done.
The idea behind the Small Web is to make it as easy as possible to have your own web site at your own domain.
And
You are allowed to rent space on Megacorp’s servers in exchange for giving up your privacy, freedom of speech, and other human rights.
Non technical folks need a way to host sites to protect themselves from predatory hosts.
This hosting should be trivial to set up and take no time or effort.
It should not rely on traditional network coordination mechanisms to function.
I see a few challenges here:
You can make it easy to setup by wall papering over all the tech complexity. But if you run on your own hardware, you will eventually have issues. Things break in unexpected ways and no software package can completely manage that complexity from a self hoster.
Hosting options like AWS Lightsail do exist that deploy a package to a container host, you manage the configuration through a web interface, but most of it is just taken care of, including backups. With encrypted traffic and ownership of the drive encryption keys I didn't see where the privacy problem is. The downside is you pay for compute.
If you want to self host, configure your border network device with dynamic DNS or tailscale and your traditional domain name will work fine. If this seems like a barrier to entry, it is, see bullet #1.
He's building risk into his platform by relying on two features of two private entities: letsencrypt and Mozilla. I quite like both of them, but there is no guarantee they will continue to operate as is for the next ten years. Will that break the whole ecosystem?
Using mainstream technology, I can move my site from on prem, to AWS in moments, move providers, etc. I realize this takes technical knowledge, but see #1 and...
What is the actual demand signal from non technical folks to run their own site on their own hardware using an address book scheme that limits their interconnection to likely more tech or privacy focused folks?
None of that is to say I don't like the technology or the project. It's neat. But I think he is playing a bit of a mental shell game if he thinks this new architecture solves the problems he thinks he's solving.
Edit: Also, if raw addresses aren't scary then neither are bargain bin gibberish domain names.
The risk component is probably the biggest roadblock to the whole idea. Not to mention, having to introduce a new convention to a large enough group of people that are nontechnical. The "best"...
The risk component is probably the biggest roadblock to the whole idea. Not to mention, having to introduce a new convention to a large enough group of people that are nontechnical. The "best" solution almost never wins in tech, it's always about early adopters, and overwriting 30+ year old conventions will prove difficult.
I don’t think I really get this one… The issues they’re discussing around domain registration do make total sense to me - even though I’m a big believer in the importance of actually owning that...
I don’t think I really get this one…
The issues they’re discussing around domain registration do make total sense to me - even though I’m a big believer in the importance of actually owning that piece of your public identity rather than borrowing it from a company that could pull it away at a whim, I understand that the friction is just too high for that notional extra security to matter to a lot of people.
The idea of maintaining a personal address book rather than only using ICANN’s global one is interesting, and I quite like it as an idea. The comparison between phone numbers and IP addresses is a good one, and makes it clear that IPv4s aren’t actually that intimidating.
But didn’t phone numbers always kind of suck? I always looked at them as a consequence of pre-digital technology rather than an optimal solution, and I don’t think it’s a coincidence that pretty much every web based service (RIP ICQ) lets you chose a username of some kind. We’ve even got what3words because people like to be able to share locations in a nice human readable way rather than just as a link with coordinates embedded.
Like @teaearlgraycold said, IP addresses are actually a lot less portable and more complex to reassign than domains, so you’re not getting something “truly yours” any more than you did with their previous subdomain based solution. If we’re taking full domain registration as the gold standard but taking a step down to allow for convenience, I guess I just see an IP based approach being equal to or a bit worse than a subdomain one in all aspects?
I feel like the author has some fundamental misunderstandings It's the opposite of artificial scarcity, domain names are scarce by definitions. Because for a domain to be useful, it needs to point...
I feel like the author has some fundamental misunderstandings
The commercial domain name system is a perfect example of the type of artificial scarcity capitalism creates and exploits.
It's the opposite of artificial scarcity, domain names are scarce by definitions. Because for a domain to be useful, it needs to point at only one thing. If a domain like round robined between 6 different websites no one would use it.
Domain names are tiny little rows in a database. They cost next-to-nothing to set up and maintain. There’s absolutely no reason why they couldn’t be a public good, paid for from the public purse
They DO cost next to nothing to setup and maintain. They only cost "exorbitant fees" in a situation where someone else already owns it, and therefore has the right to continue owning it unless they are convinced to not own it anymore.
OP's answer to this is for people to memorize IP numbers, but besides the numerous problems, like never being able to change where you're hosting your website, if you accept the premise that your viewers are OK with memorizing random alphanumeric characters to get to your website...
...then domains aren't expensive. It's expensive to get food.com. It's not expensive to get asdifja132008fas.com. It is, in fact, extremely cheap. A quick search on namecheap indicates that asdifja132008fas.com can be had for $11/YEAR.
So what's even the point? Just buy domains that aren't common words. It's fine. Go on horse staple battery generator or something. Or just have some creativity.
That's also ignoring that his address scheme solution is for people to use address books. To attach a friendly name to an address. Just like domains and DNS. He wants everyone to run a small...
That's also ignoring that his address scheme solution is for people to use address books. To attach a friendly name to an address. Just like domains and DNS. He wants everyone to run a small nonpropogating name resolver. Progress! 😃
The author framing the actual scarcity of domain names people want as artificial scarcity for the sake of capitalism seems kind of crazy to me. I would kind of like to have zestier.com, but not...
The author framing the actual scarcity of domain names people want as artificial scarcity for the sake of capitalism seems kind of crazy to me. I would kind of like to have zestier.com, but not enough to pay $4k so surely I'd rather have it for free, right? But it's just a word from the dictionary + .com and it would've long been taken by someone that just registered the whole dictionary if it wasn't for the cost, so it wouldn't even be available in the world of free domains.
Someone did purchase it. It's owned by brandbucket.com. They and a lot of other parked domain businesses just sit on domains they think are likely to be purchased. They go through the dictionary...
Someone did purchase it. It's owned by brandbucket.com. They and a lot of other parked domain businesses just sit on domains they think are likely to be purchased. They go through the dictionary and buy up the domains for $10/year and try to resell for thousands of dollars. After that point you usually just need to pay the $10/year ICANN fee.
IMO parked domain businesses are leeches and it should be a banned practice. Let people buy as many domains as they want, but heavily restrict when they're allowed to sell them.
I was going to say increase the cost nonlinearly above a certain number, but I think you'd just end up with a bunch of corporate subentities each buying their limit, and I doubt ICANN has the...
I was going to say increase the cost nonlinearly above a certain number, but I think you'd just end up with a bunch of corporate subentities each buying their limit, and I doubt ICANN has the resources to fight the legal battles for enforcement, especially internationally.
I wonder if there are other cases of incentive structures that encourage limited ownership? The only one I can think of is homestead tax exemptions, but they are also not enough.
Admittedly I hadn't bothered to check again. The example I was using was from when I last looked years ago. I now see it also is not the price I remember. At least they're having to pay something...
Admittedly I hadn't bothered to check again. The example I was using was from when I last looked years ago. I now see it also is not the price I remember.
At least they're having to pay something to squat it though. It isn't enough enough, but at least they have some overhead for all the time they're sitting on it.
An example that isn't already being sat on may have been better, although I guess it still fits my point that desirable domains are scarce. Sure there are tons of others, but those aren't the one I want and I'm not the only one to have wanted it.
(As someone with a domain under my belt currently) I like it! My first thought was exactly "but what about IPv6" as well, but it makes sense to first introduce people to the topic with the more...
(As someone with a domain under my belt currently) I like it!
My first thought was exactly "but what about IPv6" as well, but it makes sense to first introduce people to the topic with the more friendly-looking IPv4 addresses as long as those are still available, and once the system is in place, you can show an old address as an example first and then introduce the longer + more varied new ones in the same breath.
My only issue is that I'd sometimes like to connect to my own website from another device – but this isn’t really an issue given people also memorize their own plus maybe a few other important phone numbers all the time, too!
(Plus the fact you can just go to your contacts/address book and… look them up.)
Hell, now I’m wondering if I can get an IP address that matches my mobile number…
Edit: the other comments have convinced me – why not just not risk any technical issues and choose longer domain names, if it’s not meant to be rememberable addresses anyways?
I like the idea personally. Just for the purposes of getting nontechnical people to host websites I think it makes sense and is reminiscent of the "old days." Using a link shortening service...
I like the idea personally. Just for the purposes of getting nontechnical people to host websites I think it makes sense and is reminiscent of the "old days." Using a link shortening service someone could maintain a "readable" index of these as well.
This doesn't sound so great. Unlike phone numbers you can't "port out" an IP address. If you purchase a block (at least 256 IPs) then you can do BYOIP (bring your own IP) with various cloud providers. But that's thousands of dollars per year. This means your "web number" is permanently tied to one cloud provider. Not ideal but not so bad with larger and more stable (but more expensive) providers like AWS. But this could get annoying with smaller and cheaper hosts that might go out of business or prove to be more trouble than they are worth.
Yeah, not to mention that the small web better remain pretty small. If your website gets bigger, it would be seriously challenging to host it off one IP address.
Just have everyone on the small web front an F5 load balancer, problem solved!
I'd have to go check what the current status of IPV6 allocations are, but that's the only way it would be feasible. I coordinated my last orgs IPV6 allocation and it wasn't too stingy. So maybe the small web guy could submit an allocation plan and get a big fat range.
Still wouldn't address the other issues with it.
As much as I find the project neat, I don't think he is solving the problem he thinks he's solving. Or if he is, it isn't clear from the little digging I've done.
And
Honestly, I think it is telling that he describes himself as an activist more than a developer. The project seems less about a solution and more of a resistance to... the way the world works?
His general thesis here seems to be:
I see a few challenges here:
You can make it easy to setup by wall papering over all the tech complexity. But if you run on your own hardware, you will eventually have issues. Things break in unexpected ways and no software package can completely manage that complexity from a self hoster.
Hosting options like AWS Lightsail do exist that deploy a package to a container host, you manage the configuration through a web interface, but most of it is just taken care of, including backups. With encrypted traffic and ownership of the drive encryption keys I didn't see where the privacy problem is. The downside is you pay for compute.
If you want to self host, configure your border network device with dynamic DNS or tailscale and your traditional domain name will work fine. If this seems like a barrier to entry, it is, see bullet #1.
He's building risk into his platform by relying on two features of two private entities: letsencrypt and Mozilla. I quite like both of them, but there is no guarantee they will continue to operate as is for the next ten years. Will that break the whole ecosystem?
Using mainstream technology, I can move my site from on prem, to AWS in moments, move providers, etc. I realize this takes technical knowledge, but see #1 and...
What is the actual demand signal from non technical folks to run their own site on their own hardware using an address book scheme that limits their interconnection to likely more tech or privacy focused folks?
None of that is to say I don't like the technology or the project. It's neat. But I think he is playing a bit of a mental shell game if he thinks this new architecture solves the problems he thinks he's solving.
Edit: Also, if raw addresses aren't scary then neither are bargain bin gibberish domain names.
The risk component is probably the biggest roadblock to the whole idea. Not to mention, having to introduce a new convention to a large enough group of people that are nontechnical. The "best" solution almost never wins in tech, it's always about early adopters, and overwriting 30+ year old conventions will prove difficult.
I don’t think I really get this one…
The issues they’re discussing around domain registration do make total sense to me - even though I’m a big believer in the importance of actually owning that piece of your public identity rather than borrowing it from a company that could pull it away at a whim, I understand that the friction is just too high for that notional extra security to matter to a lot of people.
The idea of maintaining a personal address book rather than only using ICANN’s global one is interesting, and I quite like it as an idea. The comparison between phone numbers and IP addresses is a good one, and makes it clear that IPv4s aren’t actually that intimidating.
But didn’t phone numbers always kind of suck? I always looked at them as a consequence of pre-digital technology rather than an optimal solution, and I don’t think it’s a coincidence that pretty much every web based service (RIP ICQ) lets you chose a username of some kind. We’ve even got what3words because people like to be able to share locations in a nice human readable way rather than just as a link with coordinates embedded.
Like @teaearlgraycold said, IP addresses are actually a lot less portable and more complex to reassign than domains, so you’re not getting something “truly yours” any more than you did with their previous subdomain based solution. If we’re taking full domain registration as the gold standard but taking a step down to allow for convenience, I guess I just see an IP based approach being equal to or a bit worse than a subdomain one in all aspects?
I feel like the author has some fundamental misunderstandings
It's the opposite of artificial scarcity, domain names are scarce by definitions. Because for a domain to be useful, it needs to point at only one thing. If a domain like round robined between 6 different websites no one would use it.
They DO cost next to nothing to setup and maintain. They only cost "exorbitant fees" in a situation where someone else already owns it, and therefore has the right to continue owning it unless they are convinced to not own it anymore.
OP's answer to this is for people to memorize IP numbers, but besides the numerous problems, like never being able to change where you're hosting your website, if you accept the premise that your viewers are OK with memorizing random alphanumeric characters to get to your website...
...then domains aren't expensive. It's expensive to get food.com. It's not expensive to get asdifja132008fas.com. It is, in fact, extremely cheap. A quick search on namecheap indicates that
asdifja132008fas.comcan be had for $11/YEAR.So what's even the point? Just buy domains that aren't common words. It's fine. Go on horse staple battery generator or something. Or just have some creativity.
That's also ignoring that his address scheme solution is for people to use address books. To attach a friendly name to an address. Just like domains and DNS. He wants everyone to run a small nonpropogating name resolver. Progress! 😃
real answer is land value tax for domain names to prevent domain parking
The author framing the actual scarcity of domain names people want as artificial scarcity for the sake of capitalism seems kind of crazy to me. I would kind of like to have zestier.com, but not enough to pay $4k so surely I'd rather have it for free, right? But it's just a word from the dictionary + .com and it would've long been taken by someone that just registered the whole dictionary if it wasn't for the cost, so it wouldn't even be available in the world of free domains.
Someone did purchase it. It's owned by brandbucket.com. They and a lot of other parked domain businesses just sit on domains they think are likely to be purchased. They go through the dictionary and buy up the domains for $10/year and try to resell for thousands of dollars. After that point you usually just need to pay the $10/year ICANN fee.
IMO parked domain businesses are leeches and it should be a banned practice. Let people buy as many domains as they want, but heavily restrict when they're allowed to sell them.
I was going to say increase the cost nonlinearly above a certain number, but I think you'd just end up with a bunch of corporate subentities each buying their limit, and I doubt ICANN has the resources to fight the legal battles for enforcement, especially internationally.
I wonder if there are other cases of incentive structures that encourage limited ownership? The only one I can think of is homestead tax exemptions, but they are also not enough.
Admittedly I hadn't bothered to check again. The example I was using was from when I last looked years ago. I now see it also is not the price I remember.
At least they're having to pay something to squat it though. It isn't enough enough, but at least they have some overhead for all the time they're sitting on it.
An example that isn't already being sat on may have been better, although I guess it still fits my point that desirable domains are scarce. Sure there are tons of others, but those aren't the one I want and I'm not the only one to have wanted it.
(As someone with a domain under my belt currently) I like it!
My first thought was exactly "but what about IPv6" as well, but it makes sense to first introduce people to the topic with the more friendly-looking IPv4 addresses as long as those are still available, and once the system is in place, you can show an old address as an example first and then introduce the longer + more varied new ones in the same breath.
My only issue is that I'd sometimes like to connect to my own website from another device – but this isn’t really an issue given people also memorize their own plus maybe a few other important phone numbers all the time, too!
(Plus the fact you can just go to your contacts/address book and… look them up.)
Hell, now I’m wondering if I can get an IP address that matches my mobile number…
Edit: the other comments have convinced me – why not just not risk any technical issues and choose longer domain names, if it’s not meant to be rememberable addresses anyways?
Cue the land rush for ipv6 addresses that spell something. I'd settle for
A1:1B:ee:f7:ac:06:0d.Save b00b:1e5:d00b:1e5::1 for me.
I like the idea personally. Just for the purposes of getting nontechnical people to host websites I think it makes sense and is reminiscent of the "old days." Using a link shortening service someone could maintain a "readable" index of these as well.