Talk:InterPlanetary File System

Tags

Why are these tags there without talk page discussion? Endercase (talk) 01:33, 14 August 2017 (UTC)

Too technical? Internet for Dummies is what they're after?

The article went a long way since May 2016 (even if comparing to 2017) and, IMO, is simple and readable enough to remove the technical template. K4rolB (talk) 12:20, 28 August 2019 (UTC)

Some news in 2017?

Examples

• http://www.bitcoinisle.com/2017/06/26/what-is-the-interplanetary-file-system/
• "(...) IPFS connects all these different blockchains ...", https://www.coindesk.com/ethereum-meets-zcash-why-ipfs-plans-a-multi-blockchain-browser/

Latest addition for the Catalonian referendum

How about redacting this addition in a more neutral point of view? Something along the lines "IPFS was used before the October 1st Catalonian independence referendum, after the webpage for consulting census data was closed by the Spanish Government, in order to replicate the database, which lead to a security breach, etc."? I think it's better to focus on the technical subject rather than on the ideological subject, which has already its good deal of problems, hatred, etc. Let's not add more fuel to the already perfectly burning fire, because otherwise this addition is going to be used as an editing war by both sides, which will lead to vandalism, etc. Let's focus on the technical issue at hand, I think.

I won't do any editing myself, let's comment this and discuss this before, if the editors involved agree. DervishD (talk) 15:36, 5 October 2017 (UTC)

I agree; in the spirit of WP:BOLD I did some editing of this section. Rather than "censorship", I opted for the word "blocking". Also instead of the original external links (EFF.org, Hacker News and a personal blog), I added two news stories which should satisfy WP:Reliable sources.

I think the part about "security breach" doesn't belong in this article: it's not an inherent issue of IPFS as far as I can see and it's not covered by any news sources that I could find. -- intgr [talk] 13:13, 6 October 2017 (UTC)

You're right, the "security breach" doesn't belong here because as you mentioned it's not an inherent issue of IPFS, but of the particularly weak crypto used on the database. The redaction looks much more appropriate for Wikipedia now. Thanks for editing. DervishD (talk) 13:36, 6 October 2017 (UTC)

browser support from 2018?

Removed this text, as I can't find evidence this actually happened:

Web browsers like Firefox, Brave, and Opera are also taking steps to add built-in support for IPFS.

The cite was to Leonard, John. "Firefox 59 will support decentralised internet protocols". The Inquirer. Retrieved 27 August 2019.

This is future-looking, and no future-looking claim for anything within a mile of crypto is worth citing. So it's a year later, Firefox 59 is long past, and I can't find evidence that this ever eventuated. Is there any? - David Gerard (talk) 11:49, 28 August 2019 (UTC)

You're right, no native support in Firefox: "Support for the IPFS protocol". {{cite web}}: Cite has empty unknown parameter: |dead-url= (help), or Brave: "IPFS Integration - Roadmap and discussion". {{cite web}}: Cite has empty unknown parameter: |dead-url= (help) Both require ipfs companion plugin and desktop app. "IPFS Integration in Web Browsers Working Group". {{cite web}}: Cite has empty unknown parameter: |dead-url= (help) — K4rolB (talk) 13:03, 28 August 2019 (UTC)

There is nothing future-looking about it. As of Firefox 59, FireFox recognizes IPFS, DAT, and other p2p protocols as valid protocols. That is, it can handle ipfs://, dat:// just like it can handle ftp://, file://, :mailto, etc. Read the blog post here: https://blog.mozilla.org/addons/2018/01/26/extensions-firefox-59/. The information should be reinstated. 23.241.127.109 (talk) 06:07, 31 August 2019 (UTC)

says "approved for use by extensions", not "supports" - claiming FF "supports" IPFS when it doesn't work without a third-party extension is puffery at best - David Gerard (talk) 10:12, 31 August 2019 (UTC)

I'm not saying FireFox supports IPFS - Mozilla is. In the official documentation, they say they have "Full Support" for IPFS as a protocol handler which is exactly what the article claims was added in FireFox 59. 23.241.127.109 (talk) 15:57, 31 August 2019 (UTC)

that looks very like they added the text string "ipfs" to a whitelist, as a thing you could write a handler for. This is really not noteworthy - David Gerard (talk) 17:24, 31 August 2019 (UTC)

"Notability guidelines do not apply to content within an article" as per WP:NNC. 23.241.127.109 (talk) 22:51, 31 August 2019 (UTC)

It does when there's nothing significant about it, and you're casting about for any puffery on IPFS you can find. "Additionally, Firefox whitelisted a string" really doesn't rate mention - David Gerard (talk) 23:39, 31 August 2019 (UTC)

That's not for you to decide. A reliable source thought it was notable enough to write an entire article about. 23.241.127.109 (talk) 00:01, 1 September 2019 (UTC)

• How do we get from Mike Conca saying "Firefox itself does not implement these protocols" to "Mozilla wrote an entire article about it"? Uncle G (talk) 10:57, 1 September 2019 (UTC)
  • The inquirer wrote the article, not Mozilla. 23.241.127.109 (talk) 16:46, 1 September 2019 (UTC)

Semiprotected

I have semiprotected this article due to what looks on the face of it very much like spamming. Guy (Help!) 11:17, 1 September 2019 (UTC)

The whitewashing comment was related to the merge of the malware section with the notable users section as per WP:UNDUE. David Gerard (correction: roxy the dog) has taken this opportunity to revert all edits I made to the article including basic things like adding additional sources. Please review the changes. 23.241.127.109 (talk) 16:44, 1 September 2019 (UTC)

Disputable sentence in article

There's a small edit war going on around this sentence: In 2019 the project has been criticized for being "still not usable for websites" in spite of attracting huge investments and creating an "overextended, under-documented, and unfinished constellation of projects". The source is MacWright, Tom. "IPFS, Again". macwright.org. Retrieved 2019-06-13., which to me looks like an ordinary blog post and should not be used as a reference. Although, there was at least one voice (@Roxy the dog:) that removing it might be whitewashing. I would happily remove this sentence, what are other opinions? — K4rolB (talk) 12:04, 1 September 2019 (UTC)

My whitewashing comment related to the whole series of edits made by what appears to be an interested party. It was a response to the whole series. If consensus is that that source is unreliable then so be it, but my comment, whitewash, stands. Roxy, the dog. wooF 12:56, 1 September 2019 (UTC)

• The editor who restored the edit has an undisclosed connection to the subject. He promoted the article on his personal Twitter account. 23.241.127.109 (talk) 16:48, 1 September 2019 (UTC)

If you continue making baseless accusations of conflict of interest, you will be blocked. Guy (Help!) 21:18, 1 September 2019 (UTC)

12 June 2019: Gerard promotes blog post on Twitter. 16 July 2019: Gerard restores same blog post to article without a disclosure. That's not baseless - those are facts. 23.241.127.109 (talk) 21:26, 1 September 2019 (UTC)

Take it to WP:COIN. Oh wait, you did and it was rejected, because that doesn't in any way constitute a conflict of interest - David Gerard (talk) 21:28, 1 September 2019 (UTC)

Any external relationship can trigger a conflict of interest, which you would know if you read WP:COIN. Regardless, this edit breaks plenty of other rules. 23.241.127.109 (talk) 21:33, 1 September 2019 (UTC)

Based on my understanding of Wikipedia's guidelines - the source, content, and tone of the quote in question are all in violation of guidelines. "In 2019 the project has been criticized for being "still not usable for websites" in spite of attracting huge investments and creating an "overextended, under-documented, and unfinished constellation of projects"" is from an unreliable source (the author's personal blog), and gives undue weight to a single perspective that is not representative of a majority or large minority opinion (as demonstrated in a basic google search where many individuals describe using it successfully for websites). In addition, the content itself has a biased tone and implies that one individual's experience is a factual assessment of the project's current state. Do folks agree with that assessment? - Thanks! Nala28 (talk) 6:16, 2 Sept 2019 (UTC)

I agree. I would also add the fact that the style of citation is far from perfect, bc it is cited word by word while it's not necessary, unless you want to it sound extra harsh—"unfinished constellation of projects" is far from neutral language. — K4rolB (talk) 07:18, 2 September 2019 (UTC)

I looked through the recent edits in the article and I get the impression that David Gerard has an axe to grind with IPFS. His edits have only added negative material, but Wikipedia editors should make some attempt to keep their edits in balance. Now pretty much half the article is criticism. I'm not very familiar with the subject but that seems WP:UNDUE. Anyway, this one paragraph based on an unreliable blog post can be removed. -- intgr [talk] 07:48, 2 September 2019 (UTC)

Wikipedia has a long-standing problem with advocacy editing around cryptobollocks. David Gerard is one of the small number of reality-based editors maintaining our standards there. So it's very likely you'll see his name prominently featured in any article that's a target for cryptobollocks promoters. Guy (Help!) 08:10, 2 September 2019 (UTC)

'"reality-based" editors. Do you mean RationalWiki editors? It sounds like you are supporting Gerard because you share his biases. 23.241.127.109 (talk) 17:05, 2 September 2019 (UTC)

Five years later now. IPFS still doesn't work. I'd say it was crypto-bollocks all along. Sadly, I spent more than a month trying to build an app for it. In retrospect, IPFS is a low-quality DHT and kind of nothing more. Multi-minute timeouts trying to fetch data from a server that is in the same room is frankly absurd. Whatever. Off-topic. I'm re-visiting IPFS and this wiki page with the vague hope that maybe something, somehow got better. From what I can tell, it hasn't. Yes, this is a rant. Yes, I enjoy ranting. 67.198.37.16 (talk) 04:33, 19 November 2024 (UTC)

References

Cloudflare IPFS gateway removed as notable user?

I notice that in this diff, the Cloudflare IPFS gateway was removed as a notable user of IPFS with the rhetoric "if only the primary source says it, it's not noteworthy". However, the launch of the Cloudflare IPFS gateway has been reported on in a number of articles from other sources including ZDNet - "Cloudflare IPFS Gateway boosts decentralized website development" and Yahoo! Finance - "Web Server Security Firm Cloudflare Announces Launch of Ethereum Gateway". Given that, can we reinstate the bullet below (optionally referencing one of these additional sources)? Thanks for your help! - Nala28 (talk) 6:46, 2 September 2019 (UTC)

* In September 2018, Cloudflare launched an IPFS gateway, as well as an IPFS-backed version of their services. The gateway permits https read access to most types of files located within IPFS (but e.g. not to streaming video files).

The Yahoo! source is a reprint of CoinTelegraph, i.e. a crypto source - David Gerard (talk) 06:53, 2 September 2019 (UTC)

While I agree CoinTelegraph might be considered to be an opinionated source - the content seems factual/unbiased. Other examples of notable sources that have reported on this usage include MIT Technology Review, siliconANGLE and Forbes - so I think we can agree that the previous logic doesn't hold. Can you please undo the deletion? In general, notability doesn't apply inside articles - so while we should look for reliable references to any edit, notability should be relative to the subject at hand (aka common uses of this project) - not global notability guidelines. - Nala28 (talk) 18:11, 2 September 2019 (UTC)

Forbes contributor blogs are not acceptable sources for Wikipedia - David Gerard (talk) 18:30, 2 September 2019 (UTC)

I already demonstrated the contradiction and have since referenced 2 additional sources you don't have a problem with - in addition to the Cloudflare blog as an authoritative source on itself. This should be more than enough to undo the deletion - do you agree? - Nala28 (talk) 22:18, 2 September 2019 (UTC)

I had previously restored the edit in this diff [1] adding in a couple secondary sources. Unfortunately, this was washed away in the massive revert from roxy the dog. 23.241.127.109 (talk) 17:01, 2 September 2019 (UTC)

Cryptocurrency ???

Why is this article categorized as "Cryptocurrency?" It's a distributed filesystem, or alternatively, a file distribution system which uses cryptographic algorithms for content-addressing storage or filesystem encryption. In the same way that all modern internet protocols like HTTPS, SSH, etc are not cryptocurrencies, it is not a cryptocurrency. The categorization and edit lock is bizarre. - KMeyer (talk) 01:48, 9 September 2019 (UTC)

it's ... not? If you mean "wikiproject", that covers all blockchainy things - David Gerard (talk) 08:23, 9 September 2019 (UTC)

Yes, it is currently marked as "Cryptocurrency:"

• On this very talk page: "{{WikiProject Cryptocurrency|class=Start |importance=Low}}."
• On the article page: "{{Cryptocurrencies|state=collapsed}}."

If the category "Cryptocurrencies" is actually "All blockchainy things," perhaps it should have a name that reflects what it is?

On the other hand, how is IPFS a blockchainy thing? As far as I can tell, bitcoin and other altcoin spammers just went around trying to spoof their coins on as many articles as possible and some editor(s) made the mistake of converting that into a category annotation before the spam was removed. Here's a timeline of events:

• 3 April 2018, a registered spammer (User:JasonCarswell) adds a ton of random links to the page with absolutely no justification, including "[[Category:BitTorrent]]."
• 7 April 2018, an unregistered spammer adds their self-promoted altcoin to the page.
• 18 July 2018, you convert User:JasonCarswell's inexplicably not-yet-reverted "[[Category:BitTorrent]]" into "[[Category:Cryptocurrencies]]."
• 26 September 2018, User:Audacity adds the "{{Cryptocurrencies}}" tag, perhaps based on confusion about the the self-promoted altcoin thing or the link to the category page you added.
• Immediately subsequent to Audacity's edit, on 26 September 2018, you remove the altcoin spam, but not the bogus category tag.
• 11 January 2019, User:Balkywrest finally removes the "[[Category:Cryptocurrencies]]" derived from User:JasonCarswell's drive-by spam in April 2018, and only the bogus "{{Cryptocurrencies}}" tag is left.

I think the categorization is a mistake and the correct thing to do here is to remove the category tag. As far as I am aware, nothing about IPFS is implemented on or using a blockchain. It's just a P2P content-addressed data store. (If you want to call any distributed content-addressed storage system a blockchain, you might as well categorize Git as one. Good luck.)

- KMeyer (talk) 23:31, 14 September 2019 (UTC)

This is nonsensical. IPFS is not only the standard file system for a ton of crypto projects, its development is largely funded by an ICO. And - most tellingly - its editing has all the problems of the crypto articles. There's no way on earth it's not related, no way it shouldn't be categorised as such, and no way it's not within the ambit of the cryptocurrency project and - most importantly - the blockchain-related sanctions - David Gerard (talk) 00:22, 15 September 2019 (UTC)

Ext4 is likely the standard filesystem for a ton of (I think you meant) cryptocurrency/ies. ("Crypto" is ambiguous and IMO, chiefly means cryptography.) Are you going to slap a cryptocurrency category label on it? How about Git? It's more or less a distributed blockchain.

I can see the problems with editing on the article and support the idea of having a general blanket limitation policy in place, it just shouldn't use a "cryptocurrency" label. I don't know how to refute a statement like "there's no way on earth it's not related." It seems you've just given up on debate and are declaring your opinion fact. I totally agree the page should have edit restrictions given it is the target of spammers! I just don't think it should have an inaccurate categorization. Take care, KMeyer (talk) 20:09, 15 September 2019 (UTC)

FWIW I agree that it seems weird to categorize this page inside {{WikiProject_Cryptocurrency}} and add the {{Cryptocurrencies}} section to the bottom of the page since it's neither a blockchain nor a cryptocurrency. While used by some blockchains and cryptocurrencies, so are many other technologies (which aren't similarly tagged) - and it's clearly got a lot of non-blockchain-related users too, so seems weird to pigeonhole. - User:Nala28 (talk) 6:07, 16 September 2019 (UTC)

Small update: Now I understand why you see everything through blockchain-colored glasses: "I have written and self-published a book on cryptocurrencies and blockchains," linked to https://davidgerard.co.uk/blockchain/. Potential, if not clear, WP:COI. KMeyer (talk) 22:06, 15 September 2019 (UTC)

As you can see above, a previous drive-by here tried making the same personal attack. If you seriously think you can substantiate it, take it to WP:COIN or please desist in such - David Gerard (talk) 22:47, 15 September 2019 (UTC)

David, pointing out your potential COI isn't a personal attack. It is relevant for the same reason WP has policies about COI editing. Not sure what I'm being asked to desist doing. It only needed pointing out the once. If it becomes obvious that it is an issue, I will take your recommendation to WP:COIN. - KMeyer 05:45, 17 September 2019 (UTC)

You have literally no basis for this claim that fits a single element of WP:COI. If you do, and aren't just blowing smoke, then you need to make your case - David Gerard (talk) 06:39, 17 September 2019 (UTC)

You may call me a spammer, a Nazi, an SJW, or any other bad name you can think of - but it doesn't make it true. I attempted to improve this article and was denied on several fronts to allow the other VERY interconnected [[Protocol Labs]] projects on Wikipedia. If you really understood IPFS then you would know that Protocol Labs is developing 5 interrelated projects - as are reflected on their websites and in this uncensored article: https://infogalactic.com/info/InterPlanetary_File_System. If you really understood IPFS, one of the five, then you would realize that another of those 5, [[Filecoin]], is very connected - and THAT is where all this cryptocurrency hubbub originated. ~ JasonCarswell (talk) 02:25, 15 September 2019 (UTC)

Jason, I didn't call you any of those things except a spammer, and you know it. I call a spade a spade and am happy to defend that choice of categorization. It's not a baseless attack, just my opinion formed based on the facts available to me. Your edit was extremely characteristic of typical self-promotion spam. Your user page, etc, suggest more of the same. Even your response here attempts to insert links back to cryptocurrency topics. Re: Filecoin, how fortunate is it that we can have more than one article on this online encyclopedia? The separate Filecoin page can and should to have a cryptocurrency label and perhaps link to the separate IPFS page. It sometimes makes sense to discuss the underlying technologies something is built upon. But references in the other direction are less obviously relevant or interesting, and clearly unhelpful here. I think IPFS stands on its own quite well, which by definition means that it is NOT especially interconnected with whatever else "Protocol Labs" is doing. - KMeyer (talk) 20:22, 15 September 2019 (UTC)

And yet, it's clearly relevant to the project, so I've put the tag back. The tag is not some sort of ontological statement about IPFS - David Gerard (talk) 22:46, 15 September 2019 (UTC)

You've asserted that it's clear, but haven't shown that to be true. I don't follow the sentence about a category tag not being an ontological statement about a page. That's like, exactly the opposite of true. "Ontology, n. A systematic arrangement of all of the important categories of objects or concepts which exist in some field of discourse, showing the relations between them." Finally: I've removed the tag again, per the above. Cheers, KMeyer 05:45, 17 September 2019 (UTC)

It's a statement about the project, not a statement about IPFS. At this point, you're just being disruptive - David Gerard (talk) 06:39, 17 September 2019 (UTC)

I don’t think it’s reasonable to just claim that anyone who disagrees with you is disruptive. Seriously, look at your arguments in this thread. They’re unsourced and appeals to your own authority. There’s no compelling arguments for categorizing a file system as a currency presented; it’s bizarre. You don’t have some moral high ground here to claim disruption. KMeyer (talk) 10:42, 17 September 2019 (UTC)

Fixed it. - KMeyer (talk) 20:26, 15 September 2019 (UTC)

Unfixed again

References

References

Wikipedia-IPFS

I noticed the following article on "Hacker News" at https://news.ycombinator.com/item?id=23127829

Wikipedia-IPFS - An exploration to host Wikipedia in IPFS. • Sbmeirow • Talk • 04:30, 10 May 2020 (UTC)

Protocol Labs + article quality

I added a subsection titled "Protocol Labs" under "History". The sources I found seemed somewhat unreliable (the official organization website, some strange third party investment/promotional websites), so I added better source needed tags; if anyone could find better sources on Protocol Labs, I'd appreciate that a lot. IMHO, Protocol Labs is different enough from and well-known apart from IPFS that it should have its own article, rather than piggybacking off this one.

Also, I'll rate this thing as Start-class and Low-importance for every WikiProject listed, since it's already rated that way (and I don't think IPFS is super important to cryptography or Internet culture now anyways). Please review; thanks. Duckmather (talk) 16:38, 28 April 2021 (UTC)

Reliability

Is there any reliably sourceable information about why IPFS is so terribly unreliable? Anecdotally, I find that it works only about may 5% of the time (after trying dozens of IPFS gateways, most of which fail to resolve outright or fail to provide the requested file if they can be reached). I figure there must be some explanation for this, though it's outside my experience area.  — SMcCandlish ☏ ¢ 😼  01:13, 6 November 2024 (UTC)

I have the same question. What's up? What gives? What's the status?

A few years ago, I tried to develop an app on IPFS, and it was a failure. Multiple minute timeout trying to fetch data on a server across the room. IPNS lookups that would take 5 minutes. That was then. I'm revisiting, because with a catchy name like IPFS, it must be good, right? In the last week I've read through assorted reddit and y-combinator pages (and this talk page). None of these are "reliably sourceable"; they're quasi-organized conversations between semi-informed individuals.

At any rate, the story is grim. Cobbling together these sources, I'd summarize as follows: (1) IPFS is not actually a file system in the conventional POSIX (or MSWindows) sense. It's actually a DHT, and their own website now comes clean about this. I quote: "Our peer-to-peer content delivery network is built around the innovation of content addressing: store, retrieve, and locate data based on the fingerprint of its actual content rather than its name or location." This is from https://ipfs.tech earlier today. (2) It's not a very good DHT. Based on my personal experience (of developing an app) its a terrible DHT. Why would that be? Per discussion forums, this is because the parent company Protocol Labs halted development to focus on filecoin, instead. (3) I've yet to read a single positive report about it that isn't from the marketing dept. of Protocol Labs. There do seem to be many, and I mean lots, like hundreds? of attempted pilot projects started by hopeful onlookers who thought they might be at the start of something cool, keen, wonderful or at least useful and forward-looking. I haven't seen any reports of any of these pilot projects moving beyond the explore stage, and certainly not to implementation.

So that's it, then. Its not a file-system, as claimed. It is a DHT, as now acknowledged, but one that works poorly (as your own personal experience attests). Management has focused efforts on creating yet more cypto coin markets, rather than developing the purported foundation-stone. And, basically, no one is using it. If there were authoritative or at least "reliably sourceable" reports of its status that aren't corporate vaporware and "forward-looking statements", they should be added to this article. Or at least, dropped here on the talk page for discussion. Meanwhile, I'm vaguely irked that IPFS has managed to suck up yet another few days of my time, with zero in return. 67.198.37.16 (talk) 05:09, 19 November 2024 (UTC)

As to the technical question of why it works so poorly, I can speculate. Every blob of data is associated with a "content ID" (CID, its a hash or fingerprint). To find that data, the IPFS network has to be queried for that CID. The nearest neighbor (that you're connected to) probably doesn't have a copy of the data, but it does have a table of IP addrs that might have it. Those have to be queried, maybe recursivly, for a few more hops. Meanwhile, these network hosts are coming and going all the time. Why? I dunno: maybe they're other people experimenting with servers. Maybe they are browser plugins trying to cache some of the data. Maybe these are just people opening and closing their laptop lids, and their laptop happens to run an IPFS node. I dunno. But with each of these network join/leave events, the DHT needs to be rebalanced. Who has what needs to be updated, and maybe some of the big nodes that have a lot redistribute to those that have only a little. Or something. Read the DHT article for a general overview. So anyway, these tables of who-has-what are constantly changing, every node is updating it's map of the topology all the time. Even on an idle node, this generates a lot of traffic: you'd notice it on your wifi or your internet gateway. Given this constant background chatter of syncing and resyncing, forwarding, querying, timing out and reforwarding, I'm hardly surprised that connections are dropped. I'm guessing that the route to the host that supposedly has a copy of the data is just constantly in flux, and mostly you're getting routes to hosts that don't yet have the data, will have the data, have just joined or have already left the network. So of course they can't respond, and meanwhile, there's no fallback mechanism to re-route your request somewhere else. At least, not within a minute or two. This is my educated guess. 67.198.37.16 (talk) 17:58, 19 November 2024 (UTC)