Page MenuHomePhabricator

IABot should not edit only to change archive.is to archive.today
Open, LowPublicBUG REPORT

Description

https://it.wikipedia.org/w/index.php?diff=135572845

https://nl.wikipedia.org/w/index.php?title=Prog_(tijdschrift)&diff=next&oldid=66920158

While archive.today is the preferred domain, an edit consisting *only* of this change is an unnecessary cosmetic edit.

Event Timeline

Harej triaged this task as Low priority.Apr 3 2024, 8:20 PM
Harej added a subscriber: Green_Cardamom.

@Green_Cardamom this appears to be a result of your process

It's not cosmetic. Archive.today uses 5 or 6 domains such as .is .ph, etc.. these are where the content is located. The domain archive.today does not contain content, it's only function is to route requests to one of the other domains. Thus if archive.is becomes unavailable, the admins only need to reroute traffic to archive.ph .. however if the incoming link is archive.is then it won't function. Note that "becomes unavailable" is a domain-level outage, because this archive provider has had problems with domain registars and/or DNS resolvers that blackhole it for policy reasons.

To add, the owners of archive.today requested we at Wikipedia use archive.today as a safeguard against potential future outages of the other domains, such as archive.is, which have in the past been subject to problems.

I was the one who reported this over on InternetArchiveBot's talk page. This is unacceptable on the Dutch Wikipedia as part of our “if it ain't broke, don't fix it” policy. And requests on behalf of site owners are also looked upon as unwanted on the Dutch Wikipedia. And I agree with that: it's the exact same site.

And to add to that, the bot doesn't know what to do with archive.today either as it changed .today to .is in this example: https://nl.wikipedia.org/w/index.php?diff=prev&oldid=67337586

It is being done entirely for the benefit on Wikipedia, as explained above. Not because the site owner requested it. The site owner suggested this after I asked what they thought would be best, for the benefit of Wikipedia. The site owner could care less what we do, it does not matter to them. It's not a branding issue, if that is your concern.

Your statement "it's the exact same site" is both correct and incorrect. The content that is returned is the same. However the traffic is routed through different servers that have different behaviors. One route is prone to outages, the other route is robust ie. it is not prone to outages. It's like a 5 lane road, and a 1 lane road. They both work, but if one lane goes offline, then at least you still have 4 other lanes on the first road, but the other road is completely shut down.

So we are doing this for the benefit of Wikipedia, so it is more robust and less prone to outages.

Are you against making Wikipedia more robust? If so, why? Can you address these issues. Why do you believe the 1 lane road is better than the 5 lane road.

(As for IABot adding .is that will be fixed it's on my to do list).

Well, I chose to use archive.ph or archive.is. Even if archive.today would be better, I do not appreciate an automatic change if that's the only thing that's being changed (it would be different if it was part of a bigger changeset). Our policy also states that. And Harej agrees, given that he filed the report. But like I said: it also changes .today to .is, so apparently .today is not preferred either.