7 ms·

Honestly, I’ve never donated bc “keeping Wikipedia up” could be a solved problem by now. Not just cheaper processing and bandwidth costs, but p2p hosting, etc.

Honestly, I’ve never donated bc “keeping Wikipedia up” could be a solved problem by now. Not just cheaper processing and bandwidth costs, but p2p hosting, etc. Every time I see that ad, I say to myself “it should not be that expensive”. Now, if they had said “We need 200k to send CC photographers to the Olympics”, count me in. That sounds bigger and way more interesting!

ReactiveJelly 2mo ago
I don't even use Wikipedia anymore, I use the free online P2P encyclopedia. It's called
Gander5739 2mo ago
Hosting isn't the expensive part, maintaining mediawiki is. For example, unifying the parser for the markup language: https://www.mediawiki.org/wiki/Parsoid/Parser_Unification/en https://www.mediawiki.org/wiki/Parsoid/Parser_Unification/en
- account42 2mo ago
  Maintaining a shitty PHP website isn't that expensive either.
- bawolff 2mo ago
  Depends on how you define "hosting". CapEx costs for buying the servers is a non trivial expense (note: in accounting statements this isn't considered part of internet hosting). Not as much as software, but still expensive.
bawolff 2mo ago
P2P hosting doesn't really work in reality, especially for websites that change regularly and require quick propagation of those changes. Its not just Wikipedia. Name a high traffic site that actually uses p2p hosting. I'll wait. Bandwidth is also the sort of thing that gets way cheaper at scale than most people expect. When you are big enough there comes a point where the telecom companies want to connect to you because all their users want to access your website. It wouldn't surprise me if a lot (not all) of Wikimedia's bandwidth is actually free.
- joshspankit 2mo ago
  Actually, I think your argument is almost entirely about P2P tech inspired by BitTorrent. We’ve had good be the enemy of great here. No, no high-traffic sites actually uses P2P hosting but that doesn’t mean we’re not at a point where they can. In fact, I’d say the biggest impediment to P2P sites in 2026 is the absolute unit that is advertiser data. We have P2P databases, and round trip on the earth is ~300ms (aka “quick propagation” can be seconds which is more that enough for 98%). We have VPS and serverless. We also have content addressing and symmetric cert verification, but when a site’s “traffic” is P2P you lose visibility on most of it. And can you imagine the nonsense of advertisers adding client side tracking data to the public db so that they can query it? Woof.
  - bawolff 2mo ago
    > No, no high-traffic sites actually uses P2P hosting but that doesn’t mean we’re not at a point where they can. Even if true, being the first to do anything is usually an expensive proposition. > We have P2P databases, and round trip on the earth is ~300ms (aka “quick propagation” can be seconds which is more that enough for 98%) I dont know how you are defining "p2p database" or which product you mean, but in general distributed p2p solutions where you do not trust the peers cannot propagate information in a single round-trip, or anything even remotely close to that. > We have VPS and serverless Neither of these are p2p technologies. They are also a terrible fit for something like Wikipedia. > We also have content addressing Sure, but that is like 1% of a solution > symmetric cert verification Do you mean mutual TLS? That is indeed a technology that exists, but i dont know how its relavent. > And can you imagine the nonsense of advertisers adding client side tracking data to the public db so that they can query it? Which kind of brings up the other flaw with p2p - privacy. It obviously depends on what technology is chosen, but p2p often has poor privacy properties. Think media companies sending threatening notices to people who download movies. People often look up private things on Wikipedia. Think sexual health topics. Do we really want malicious people to be able to track that?

ReactiveJelly 2mo ago

I don't even use Wikipedia anymore, I use the free online P2P encyclopedia. It's called

Gander5739 2mo ago

Hosting isn't the expensive part, maintaining mediawiki is. For example, unifying the parser for the markup language: https://www.mediawiki.org/wiki/Parsoid/Parser_Unification/en https://www.mediawiki.org/wiki/Parsoid/Parser_Unification/en

account42 2mo ago
Maintaining a shitty PHP website isn't that expensive either.
bawolff 2mo ago
Depends on how you define "hosting". CapEx costs for buying the servers is a non trivial expense (note: in accounting statements this isn't considered part of internet hosting). Not as much as software, but still expensive.

bawolff 2mo ago

P2P hosting doesn't really work in reality, especially for websites that change regularly and require quick propagation of those changes. Its not just Wikipedia. Name a high traffic site that actually uses p2p hosting. I'll wait. Bandwidth is also the sort of thing that gets way cheaper at scale than most people expect. When you are big enough there comes a point where the telecom companies want to connect to you because all their users want to access your website. It wouldn't surprise me if a lot (not all) of Wikimedia's bandwidth is actually free.

joshspankit 2mo ago
Actually, I think your argument is almost entirely about P2P tech inspired by BitTorrent. We’ve had good be the enemy of great here. No, no high-traffic sites actually uses P2P hosting but that doesn’t mean we’re not at a point where they can. In fact, I’d say the biggest impediment to P2P sites in 2026 is the absolute unit that is advertiser data. We have P2P databases, and round trip on the earth is ~300ms (aka “quick propagation” can be seconds which is more that enough for 98%). We have VPS and serverless. We also have content addressing and symmetric cert verification, but when a site’s “traffic” is P2P you lose visibility on most of it. And can you imagine the nonsense of advertisers adding client side tracking data to the public db so that they can query it? Woof.
- bawolff 2mo ago
  > No, no high-traffic sites actually uses P2P hosting but that doesn’t mean we’re not at a point where they can. Even if true, being the first to do anything is usually an expensive proposition. > We have P2P databases, and round trip on the earth is ~300ms (aka “quick propagation” can be seconds which is more that enough for 98%) I dont know how you are defining "p2p database" or which product you mean, but in general distributed p2p solutions where you do not trust the peers cannot propagate information in a single round-trip, or anything even remotely close to that. > We have VPS and serverless Neither of these are p2p technologies. They are also a terrible fit for something like Wikipedia. > We also have content addressing Sure, but that is like 1% of a solution > symmetric cert verification Do you mean mutual TLS? That is indeed a technology that exists, but i dont know how its relavent. > And can you imagine the nonsense of advertisers adding client side tracking data to the public db so that they can query it? Which kind of brings up the other flaw with p2p - privacy. It obviously depends on what technology is chosen, but p2p often has poor privacy properties. Think media companies sending threatening notices to people who download movies. People often look up private things on Wikipedia. Think sexual health topics. Do we really want malicious people to be able to track that?