3 ms·

I considered this a few months ago, but it's a bad idea for a few reasons. One is that it increases the load on archive.org. (There are a significant number of

I considered this a few months ago, but it's a bad idea for a few reasons. One is that it increases the load on archive.org. (There are a significant number of bots, and adding 30 million new links for them to spider is no small thing.) The more important reason is that it's hard to do programmatically. For every link, you'd want to link to the most recent, active archive entry. As far as I know, this isn't easy to determine. Have you ever had the experience of going to archive.org, plugging in a url, and seeing that the most recent one is actually dead, but then you can go back in time and find what you're after? It's easy to say "Well, that's ok. Take me to the dead link, and I'll go back myself." Maybe. But with Dan's exacting standards, I somehow don't think that'll pass muster. The ultimate reason this probably won't happen, though, is because people can do it themselves. It's not hard to copy the link and visit archive.org. Every other link (past, flag, hide, favorite) is there because it's hard to do. EDIT: On the other hand, https://news.ycombinator.com/item?id=30384630 https://news.ycombinator.com/item?id=30384630 shows a way it could be done. I didn't know about the date wildcard feature.

ffpip 5y ago
> most recent, active archive entry. As far as I know, this isn't easy to determine. https://web.archive.org/$link https://web.archive.org/$link will always take you to the most recently archived version of the link. Eg: https://web.archive.org/https://example.com https://web.archive.org/https://example.com
- sillysaurusx 5y ago
  Yes. The point is that if you try doing that for links far in the past, you'll often get a dead page. By "dead" I mean different content than the one that was submitted to HN at the time. So you need a way of specifying the date. There's a way to do that, though, which is interesting. My edit above points out a comment to get it.
- ss48 5y ago
  This could be challenging for existing links, but at least moving forward, when a submission is made, HN could autorequest an archive be made on archive.org and then use that when link is available in past or search.
Fnoord 5y ago
Should be doable via a user script in e.g. Tampermonkey or Greasemonkey. It would avoid 30 million links for bots but would allow the feature for power users.
joecool1029 5y ago
>One is that it increases the load on archive.org. (There are a significant number of bots, and adding 30 million new links for them to spider is no small thing.) I mean, Wikipedia did it and archive.org survived. That has to be a totally different ballpark with regards to backlinks and traffic than niche HN.
- sillysaurusx 5y ago
  Niche? :) Maybe a few years ago. But you’re probably right about the load.
jka 5y ago
> For every link, you'd want to link to the most recent, active archive entry. Hrm, are you sure about that? I think I'd generally prefer to see each article as it looked when it was posted to HN, at least by default. (that would translate to: minimum archive.org timestamp that is greater than the HN post timestamp) Edit: ah, I've read your reply where you mention this exact use case. All good.
- sillysaurusx 5y ago
  Yeah, the more we talk it over, the better of an idea it sounds. It could even be restricted to submissions more than a year old. It would be an interesting experiment. You could prototype it with a tampermonkey script and post it as a Show HN.

ffpip 5y ago

> most recent, active archive entry. As far as I know, this isn't easy to determine. https://web.archive.org/$link https://web.archive.org/$link will always take you to the most recently archived version of the link. Eg: https://web.archive.org/https://example.com https://web.archive.org/https://example.com

sillysaurusx 5y ago
Yes. The point is that if you try doing that for links far in the past, you'll often get a dead page. By "dead" I mean different content than the one that was submitted to HN at the time. So you need a way of specifying the date. There's a way to do that, though, which is interesting. My edit above points out a comment to get it.
ss48 5y ago
This could be challenging for existing links, but at least moving forward, when a submission is made, HN could autorequest an archive be made on archive.org and then use that when link is available in past or search.

Fnoord 5y ago

Should be doable via a user script in e.g. Tampermonkey or Greasemonkey. It would avoid 30 million links for bots but would allow the feature for power users.

joecool1029 5y ago

>One is that it increases the load on archive.org. (There are a significant number of bots, and adding 30 million new links for them to spider is no small thing.) I mean, Wikipedia did it and archive.org survived. That has to be a totally different ballpark with regards to backlinks and traffic than niche HN.

sillysaurusx 5y ago
Niche? :) Maybe a few years ago. But you’re probably right about the load.

jka 5y ago

> For every link, you'd want to link to the most recent, active archive entry. Hrm, are you sure about that? I think I'd generally prefer to see each article as it looked when it was posted to HN, at least by default. (that would translate to: minimum archive.org timestamp that is greater than the HN post timestamp) Edit: ah, I've read your reply where you mention this exact use case. All good.

sillysaurusx 5y ago
Yeah, the more we talk it over, the better of an idea it sounds. It could even be restricted to submissions more than a year old. It would be an interesting experiment. You could prototype it with a tampermonkey script and post it as a Show HN.