From d3f8d858a992abf9d02513e79baf0b83056f94b1 Mon Sep 17 00:00:00 2001 From: aREversez <53844261+aREversez@users.noreply.github.com> Date: Sat, 26 Nov 2022 23:33:51 +0800 Subject: [PATCH] translated --- ...iend- The Facebook That Could Have Been.md | 242 ------------------ ...iend- The Facebook That Could Have Been.md | 238 +++++++++++++++++ 2 files changed, 238 insertions(+), 242 deletions(-) delete mode 100644 sources/talk/20200105 Friend of a Friend- The Facebook That Could Have Been.md create mode 100644 translated/talk/20200105 Friend of a Friend- The Facebook That Could Have Been.md diff --git a/sources/talk/20200105 Friend of a Friend- The Facebook That Could Have Been.md b/sources/talk/20200105 Friend of a Friend- The Facebook That Could Have Been.md deleted file mode 100644 index 024278de13..0000000000 --- a/sources/talk/20200105 Friend of a Friend- The Facebook That Could Have Been.md +++ /dev/null @@ -1,242 +0,0 @@ -[#]: subject: "Friend of a Friend: The Facebook That Could Have Been" -[#]: via: "https://twobithistory.org/2020/01/05/foaf.html" -[#]: author: "Two-Bit History https://twobithistory.org" -[#]: collector: "lujun9972" -[#]: translator: "aREversez" -[#]: reviewer: " " -[#]: publisher: " " -[#]: url: " " - -Friend of a Friend: The Facebook That Could Have Been -====== - -> _I express my network in a FOAF file, and that is the start of the revolution._ —Tim Berners-Lee (2007) - -The FOAF standard, or Friend of a Friend standard, is a now largely defunct/ignored/superseded[1][1] web standard dating from the early 2000s that hints at what social networking might have looked like had Facebook not conquered the world. Before we talk about FOAF though, I want to talk about the New York City Subway. - -The New York City Subway is controlled by a single entity, the Metropolitan Transportation Agency, better known as the MTA. The MTA has a monopoly on subway travel in New York City. There is no legal way to travel in New York City by subway without purchasing a ticket from the MTA. The MTA has no competitors, at least not in the “subway space.” - -This wasn’t always true. Surprisingly, the subway system was once run by two corporations that competed with each other. The Inter-borough Rapid Transit Company (IRT) operated lines that ran mostly through Manhattan, while the Brooklyn-Manhattan Transit Corporation (BMT) operated lines in Brooklyn, some of which extended into Manhattan also. In 1932, the City opened its own service called the Independent Subway System to compete with the IRT and BMT, and so for a while there were _three_ different organizations running subway lines in New York City. - -One imagines that this was not an effective way to run a subway. It was not. Constructing interchanges between the various systems was challenging because the IRT and BMT used trains of different widths. Interchange stations also had to have at least two different fare-collection areas since passengers switching trains would have to pay multiple operators. The City eventually took over the IRT and BMT in 1940, bringing the whole system together under one operator, but some of the inefficiencies that the original division entailed are still problems today: Trains designed to run along lines inherited from the BMT (e.g. the A, C, or E) cannot run along lines inherited from the IRT (e.g. the 1, 2, or 3) because the IRT tunnels are too narrow. As a result, the MTA has to maintain two different fleets of mutually incompatible subway cars, presumably at significant additional expense relative to other subway systems in the world that only have to deal with a single tunnel width. - -This legacy of the competition between the IRT and BMT suggests that subway systems naturally tend toward monopoly. It just makes more sense for there to be a single operator than for there to be competing operators. Average passengers are amply compensated for the loss of choice by never having to worry about whether they brought their IRT MetroCard today but forgot their BMT MetroCard at home. - -Okay, so what does the Subway have to do with social networking? Well, I have wondered for a while now whether Facebook has, like the MTA, a natural monopoly. Facebook does seem to have _a_ monopoly, whether natural or unnatural—not over social media per se (I spend much more time on Twitter), but over my internet social connections with real people I know. It has a monopoly over, as they call it, my digitized “social graph”; I would quit Facebook tomorrow if I didn’t worry that by doing so I might lose many of those connections. I get angry about this power that Facebook has over me. I get angry in a way that I do not get angry about the MTA, even though the Subway is, metaphorically and literally, a sprawling trash fire. And I suppose I get angry because at root I believe that Facebook’s monopoly, unlike the MTA’s, is not a natural one. - -What this must mean is that I think Facebook owns all of our social data now because they happened to get there first and then dig a big moat around themselves, not because a world with competing Facebook-like platforms is inefficient or impossible. Is that true, though? There are some good reasons to think it isn’t: Did Facebook simply get there first, or did they instead just do social networking better than everyone else? Isn’t the fact that there is only one Facebook actually convenient if you are trying to figure out how to contact an old friend? In a world of competing Facebooks, what would it mean if you and your boyfriend are now Facebook official, but he still hasn’t gotten around to updating his relationship status on VisageBook, which still says he is in a relationship with his college ex? Which site will people trust? Also, if there were multiple sites, wouldn’t everyone spend a lot more time filling out web forms? - -In the last few years, as the disadvantages of centralized social networks have dramatically made themselves apparent, many people have attempted to create decentralized alternatives. These alternatives are based on open standards that could potentially support an ecosystem of inter-operating social networks (see e.g. [the Fediverse][2]). But none of these alternatives has yet supplanted a dominant social network. One obvious explanation for why this hasn’t happened is the power of network effects: With everyone already on Facebook, any one person thinking of leaving faces a high cost for doing so. Some might say this proves that social networks are natural monopolies and stop there; I would say that Facebook, Twitter, et al. chose to be walled gardens, and given that people have envisioned and even built social networks that inter-operate, the network effects that closed platforms enjoy tell us little about the inherent nature of social networks. - -So the real question, in my mind, is: Do platforms like Facebook continue to dominate merely because of their network effects, or is having a single dominant social network more efficient in the same way that having a single operator for a subway system is more efficient? - -Which finally brings me back to FOAF. Much of the world seems to have forgotten about the FOAF standard, but FOAF was an attempt to build a decentralized and open social network before anyone had even heard of Facebook. If any decentralized social network ever had a chance of occupying the redoubt that Facebook now occupies before Facebook got there, it was FOAF. Given that a large fraction of humanity now has a Facebook account, and given that relatively few people know about FOAF, should we conclude that social networking, like subway travel, really does lend itself to centralization and natural monopoly? Or does the FOAF project demonstrate that decentralized social networking was a feasible alternative that never became popular for other reasons? - -### The Future from the Early Aughts - -The FOAF project, begun in 2000, set out to create a universal standard for describing people and the relationships between them. That might strike you as a wildly ambitious goal today, but aspirations like that were par for the course in the late 1990s and early 2000s. The web (as people still called it then) had just trounced closed systems like America Online and [Prodigy][3]. It could only have been natural to assume that further innovation in computing would involve the open, standards-based approach embodied by the web. - -Many people believed that the next big thing was for the web to evolve into something called the Semantic Web. [I have written about][4] what exactly the Semantic Web was supposed to be and how it was supposed to work before, so I won’t go into detail here. But I will sketch the basic vision motivating the people who worked on Semantic Web technologies, because the FOAF standard was an application of that vision to social networking. - -There is an essay called [“How Google beat Amazon and Ebay to the Semantic Web”][5] that captures the lofty dream of the Semantic Web well. It was written by Paul Ford in 2002. The essay imagines a future (as imminent as 2009) in which Google, by embracing the Semantic Web, has replaced Amazon and eBay as the dominant e-commerce platform. In this future, you can search for something you want to purchase—perhaps a second-hand Martin guitar—by entering `buy:martin guitar` into Google. Google then shows you all the people near your zipcode selling Martin guitars. Google knows about these people and their guitars because Google can read RDF, a markup language and core Semantic Web technology focused on expressing relationships. Regular people can embed RDF on their web pages to advertise (among many other things) the items they have to sell. Ford predicts that as the number of people searching for and advertising products this way grows, Amazon and eBay will lose their near-monopolies over, respectively, first-hand and second-hand e-commerce. Nobody will want to search a single centralized database for something to buy when they could instead search the whole web. Even Google, Ford writes, will eventually lose its advantage, because in theory anyone could crawl the web reading RDF and offer a search feature similar to Google’s. At the very least, if Google wanted to make money from its Semantic Web marketplace by charging a percentage of each transaction, that percentage would probably by forced down over time by competitors offering a more attractive deal. - -Ford’s imagined future was an application of RDF, or the Resource Description Framework, to e-commerce, but the exciting thing about RDF was that hypothetically it could be used for anything. The RDF standard, along with a constellation of related standards, once widely adopted, was supposed to blow open database-backed software services on the internet the same way HTML had blown open document publishing on the internet. - -One arena that RDF and other Semantic Web technologies seemed poised to takeover immediately was social networking. The FOAF project, known originally as “RDF Web Ring” before being renamed, was the Semantic Web effort offshoot that sought to accomplish this. FOAF was so promising in its infancy that some people thought it would inevitably make all other social networking sites obsolete. A 2004 Guardian article about the project introduced FOAF this way: - -> In the beginning, way back in 1996, it was SixDegrees. Last year, it was Friendster. Last week, it was Orkut. Next week, it could be Flickr. All these websites, and dozens more, are designed to build networks of friends, and they are currently at the forefront of the trendiest internet development: social networking. But unless they can start to offer more substantial benefits, it is hard to see them all surviving, once the Friend Of A Friend (FOAF) standard becomes a normal part of life on the net.[2][6] - -The article goes on to complain that the biggest problem with social networking is that there are too many social networking sites. Something is needed that can connect all of the different networks together. FOAF is the solution, and it will revolutionize social networking as a result. - -FOAF, according to the article, would tie the different networks together by doing three key things: - - * It would establish a machine-readable format for social data that could be read by any social networking site, saving users from having to enter this information over and over again - * It would allow “personal information management programs,” i.e. your “Contacts” application, to generate a file in this machine-readable format that you could feed to social networking sites - * It would further allow this machine-readable format to be hosted on personal homepages and read remotely by social networking sites, meaning that you would be able to keep your various profiles up-to-date by just pushing changes to your own homepage - - - -It is hard to believe today, but the problem in 2004, at least for savvy webizens and technology columnists aware of all the latest sites, was not the lack of alternative social networks but instead the proliferation of them. Given _that_ problem—so alien to us now—one can see why it made sense to pursue a single standard that promised to make the proliferation of networks less of a burden. - -### The FOAF Spec - -According to the description currently given on the FOAF project’s website, FOAF is “a computer language defining a dictionary of people-related terms that can be used in structured data.” Back in 2000, in a document they wrote to explain the project’s goals, Dan Brickley and Libby Miller, FOAF’s creators, offered a different description that suggests more about the technology’s ultimate purpose—they introduced FOAF as a tool that would allow computers to read the personal information you put on your homepage the same way that other humans do.[3][7] FOAF would “help the web do the sorts of things that are currently the proprietary offering of centralised services.”[4][8] By defining a standard vocabulary for people and the relationships between them, FOAF would allow you to ask the web questions such as, “Find me today’s web recommendations made by people who work for Medical organizations,” or “Find me recent publications by people I’ve co-authored documents with.” - -Since FOAF is a standardized vocabulary, the most important output of the FOAF project was the FOAF specification. The FOAF specification defines a small collection of RDF _classes_ and RDF _properties_. (I’m not going to explain RDF here, but again see [my post about the Semantic Web][4] if you want to know more.) The RDF _classes_ defined by the FOAF specification represent subjects you might want to describe, such as people (the `Person` class) and organizations (the `Organization` class). The RDF _properties_ defined by the FOAF specification represent logical statements you might make about the different subjects. A person could have, for example, a first name (the `givenName` property), a last name (the `familyName` property), perhaps even a personality type (the `myersBriggs` property), and be near another person or location (the `based_near` property). The idea was that these classes and properties would be sufficient to represent the kind of the things people say about themselves and their friends on their personal homepage. - -The FOAF specification gives the following as an example of a well-formed FOAF document. This example uses XML, though an equivalent document could be written using JSON or a number of other formats: - -``` - - - Dan Brickley - - - - - -``` - -This FOAF document describes a person named “Dan Brickley” (one of the specification’s authors) that has a homepage at `http://danbri.org`, something called an “open ID,” and a picture available at `/images/me.jpg`, presumably relative to the base address of Brickley’s homepage. The FOAF-specific terms are prefixed by `foaf:`, indicating that they are part of the FOAF namespace, while the more general RDF terms are prefixed by `rdf:`. - -Just to persuade you that FOAF isn’t tied to XML, here is a similar FOAF example from Wikipedia, expressed using a format called JSON-LD[5][9]: - -``` - - { - "@context": { - "name": "http://xmlns.com/foaf/0.1/name", - "homepage": { - "@id": "http://xmlns.com/foaf/0.1/workplaceHomepage", - "@type": "@id" - }, - "Person": "http://xmlns.com/foaf/0.1/Person" - }, - "@id": "https://me.example.com", - "@type": "Person", - "name": "John Smith", - "homepage": "https://www.example.com/" - } - -``` - -This FOAF document describes a person named John Smith with a homepage at `www.example.com`. - -Perhaps the best way to get a feel for how FOAF works is to play around with [FOAF-a-matic][10], a web tool for generating FOAF documents. It allows you to enter information about yourself using a web form, then uses that information to create the FOAF document (in XML) that represents you. FOAF-a-matic demonstrates how FOAF could have been used to save everyone from having to enter their social information into a web form ever again—if every social networking site could read FOAF, all you’d need to do to sign up for a new site is point the site to the FOAF document that FOAF-a-matic generated for you. - -Here is a slightly more complicated FOAF example, representing me, that I created using FOAF-a-matic: - -``` - - - - - - - - - - Sinclair Target - Sinclair - Target - - - - - John Smith - - - - - - - -``` - -This example has quite a lot of preamble setting up the various XML namespaces used by the document. There is also a section containing data about the tool that was used to generate the document, largely so that, it seems, people know whom to email with complaints. The `foaf:Person` element describing me tells you my name, email address, and homepage. There is also a nested `foaf:knows` element telling you that I am friends with John Smith. - -This example illustrates another important feature of FOAF documents: They can link to each other. If you remember from the previous example, my friend John Smith has a homepage at `www.example.com`. In this example, where I list John Smith as a `foaf:person` with whom I have a `foaf:knows` relationship, I also provide a `rdfs:seeAlso` element that points to John Smith’s FOAF document hosted on his homepage. Because I have provided this link, any program reading my FOAF document could find out more about John Smith by following the link and reading his FOAF document. In the FOAF document we have for John Smith above, John did not provide any information about his friends (including me, meaning, tragically, that our friendship is unidirectional). But if he had, then the program reading my document could find out not only about me but also about John, his friends, their friends, and so on, until the program has crawled the whole social graph that John and I inhabit. - -This functionality will seem familiar to anyone that has used Facebook, which is to say that this functionality will seem familiar to you. There is no `foaf:wall` property or `foaf:poke` property to replicate Facebook’s feature set exactly. Obviously, there is also no slick blue user interface that everyone can use to visualize their FOAF social network; FOAF is just a vocabulary. But Facebook’s core feature—the feature that I have argued is key to Facebook’s monopoly power over, at the very least, myself—is here provided in a distributed way. FOAF allows a group of friends to represent their real-life social graph digitally by hosting FOAF documents on their own homepages. It allows them to do this without surrendering control of their data to a centralized database in the sky run by a billionaire android-man who spends much of his time apologizing before congressional committees. - -### FOAF on Ice - -If you visit the current FOAF project homepage, you will notice that, in the top right corner, there is an image of the character Fry from the TV series Futurama, stuck inside some sort of stasis chamber. This is a still from the pilot episode of Futurama, in which Fry gets frozen in a cryogenic tank in 1999 only to awake a millennium later in 2999. Brickley, whom I messaged briefly on Twitter, told me that he put that image there as a way communicating that the FOAF project is currently “in stasis,” though he hopes that there will be a future opportunity to resuscitate the project along with its early 2000s optimism about how the web should work. - -FOAF never revolutionized social networking the way that the 2004 Guardian article about it expected it would. Some social networking sites decided to support the standard: LiveJournal and MyOpera are examples.[6][11] FOAF even played a role in Howard Dean’s presidential campaign in 2004—a group of bloggers and programmers got together to create a network of websites they called “DeanSpace” to promote the campaign, and these sites used FOAF to keep track of supporters and volunteers.[7][12] But today FOAF is known primarily for being one of the more widely used vocabularies of RDF, itself a niche standard on the modern web. If FOAF is part of your experience of the web today at all, then it is as an ancestor to the technology that powers Google’s “knowledge panels” (the little sidebars that tell you the basics about a person or a thing if you searched for something simple). Google uses vocabularies published by the schema.org project—the modern heir to the Semantic Web effort—to populate its knowledge panels.[8][13] The schema.org vocabulary for describing people seems to be somewhat inspired by FOAF and serves many of the same purposes. - -So why didn’t FOAF succeed? Why do we all use Facebook now instead? Let’s ignore that FOAF is a simple standard with nowhere near as many features as Facebook—that’s true today, clearly, but if FOAF had enjoyed more momentum it’s possible that applications could have been built on top of it to deliver a Facebook-like experience. The interesting question is: Why didn’t this nascent form of distributed social networking catch fire when Facebook was not yet around to compete with it? - -There probably is no single answer to that question, but if I had to pick one, I think the biggest issue is that FOAF only makes sense on a web where everyone has a personal website. In the late 1990s and early 2000s, it might have been easy to assume the web would eventually look like this, especially since so many of the web’s early adopters were, as far as I can tell, prolific bloggers or politically engaged technologists excited to have a platform. But the reality is that regular people don’t want to have to learn how to host a website. FOAF allows you to control your own social information and broadcast it to social networks instead of filling out endless web forms, which sounds pretty great if you already have somewhere to host that information. But most people in practice found it easier to just fill out the web form and sign up for Facebook than to figure out how to buy a domain and host some XML. - -What does this mean for my original question about whether or not Facebook’s monopoly is a natural one? I think I have to concede that the FOAF example is evidence that social networking _does_ naturally lend itself to monopoly. - -That people did not want to host their own data isn’t especially meaningful itself—modern distributed social networks like [Mastodon][14] have solved that problem by letting regular users host their profiles on nodes set up by more savvy users. It is a sign, however, of just how much people hate complexity. This is bad news for decentralized social networks, because they are inherently more complex under the hood than centralized networks in a way that is often impossible to hide from users. - -Consider FOAF: If I were to write an application that read FOAF data from personal websites, what would I do if Sally’s FOAF document mentions a John Smith with a homepage at `example.com`, and Sue’s FOAF document mentions a John Smith with a homepage at `example.net`? Are we talking about a single John Smith with two websites or two entirely different John Smiths? What if the both FOAF documents list John Smith’s email as `johnsmith@gmail.com`? This issue of identity was an acute one for FOAF. In a 2003 email, Brickley wrote that because there does not exist and probably should not exist a “planet-wide system for identifying people,” the approach taken by FOAF is “pluralistic.”[9][15] Some properties of FOAF people, such as email addresses and homepage addresses, are special in that their values are globally unique. So these different properties can be used to merge (or, as Libby Miller called it, “smoosh”) FOAF documents about people together. But none of these special properties are privileged above the others, so it’s not obvious how to handle our John Smith case. Do we trust the homepages and conclude we have two different people? Or do we trust the email addresses and conclude we have a single person? Could I really write an application capable of resolving this conflict without involving (and inconveniencing) the user? - -Facebook, with its single database and lack of political qualms, could create a “planet-wide system for identifying people” and so just gave every person a unique Facebook ID. Problem solved. - -Complexity alone might not doom distributed social networks if people cared about being able to own and control their data. But FOAF’s failure to take off demonstrates that people have never valued control very highly. As one blogger has put it, “‘Users want to own their own data’ is an ideology, not a use case.”[10][16] If users do not value control enough to stomach additional complexity, and if centralized systems are more simple than distributed ones—and if, further, centralized systems tend to be closed and thus the successful ones enjoy powerful network effects—then social networks are indeed natural monopolies. - -That said, I think there is still a distinction to be drawn between the subway system case and the social networking case. I am comfortable with the MTA’s monopoly on subway travel because I expect subway systems to be natural monopolies for a long time to come. If there is going to be only one operator of the New York City Subway, then it ought to be the government, which is at least nominally more accountable than a private company with no competitors. But I do not expect social networks to stay natural monopolies. The Subway is carved in granite; the digital world is writ in water. Distributed social networks may now be more complicated than centralized networks in the same way that carrying two MetroCards is more complicated than carrying one. In the future, though, the web, or even the internet, could change in fundamental ways that make distributed technology much easier to use. - -If that happens, perhaps FOAF will be remembered as the first attempt to build the kind of social network that humanity, after a brief experiment with corporate mega-databases, does and always will prefer. - -_If you enjoyed this post, more like it come out every four weeks! Follow [@TwoBitHistory][17] on Twitter or subscribe to the [RSS feed][18] to make sure you know when a new post is out._ - -_Previously on TwoBitHistory…_ - -> I know it's been too long since my last post, but my new one is here! I wrote almost 5000 words on John Carmack, Doom, and the history of the binary space partitioning tree. -> -> — TwoBitHistory (@TwoBitHistory) [November 6, 2019][19] - - 1. Please note that I did not dare say “dead.” [↩︎][20] - - 2. Jack Schofield, “Let’s be Friendsters,” The Guardian, February 19, 2004, accessed January 5, 2020, . [↩︎][21] - - 3. Dan Brickley and Libby Miller, “Introducing FOAF,” FOAF Project, 2008, accessed January 5, 2020, . [↩︎][22] - - 4. ibid. [↩︎][23] - - 5. Wikipedia contributors, “JSON-LD,” Wikipedia: The Free Encyclopedia, December 13, 2019, accessed January 5, 2020, . [↩︎][24] - - 6. “Data Sources,” FOAF Project Wiki, December 11 2009, accessed January 5, 2020, . [↩︎][25] - - 7. Aldon Hynes, “What is Dean Space?”, Extreme Democracy, accessed January 5, 2020, . [↩︎][26] - - 8. “Understand how structured data works,” Google Developer Portal, accessed January 5, 2020, . [↩︎][27] - - 9. tef, “Why your distributed network will not work,” Progamming is Terrible, January 2, 2013, . [↩︎][28] - - 10. Dan Brickley, “Identifying things in FOAF,” rdfweb-dev Mailing List, July 10, 2003, accessed on January 5, 2020, . [↩︎][29] - - - - --------------------------------------------------------------------------------- - -via: https://twobithistory.org/2020/01/05/foaf.html - -作者：[Two-Bit History][a] -选题：[lujun9972][b] -译者：[aREversez](https://github.com/aREversez) -校对：[校对者ID](https://github.com/校对者ID) - -本文由 [LCTT](https://github.com/LCTT/TranslateProject) 原创编译，[Linux中国](https://linux.cn/) 荣誉推出 - -[a]: https://twobithistory.org -[b]: https://github.com/lujun9972 -[1]: tmp.mJHAgyVHGr#fn:1 -[2]: https://en.wikipedia.org/wiki/Fediverse -[3]: https://en.wikipedia.org/wiki/Prodigy_(online_service) -[4]: https://twobithistory.org/2018/05/27/semantic-web.html -[5]: https://www.ftrain.com/google_takes_all -[6]: tmp.mJHAgyVHGr#fn:2 -[7]: tmp.mJHAgyVHGr#fn:3 -[8]: tmp.mJHAgyVHGr#fn:4 -[9]: tmp.mJHAgyVHGr#fn:5 -[10]: http://www.ldodds.com/foaf/foaf-a-matic -[11]: tmp.mJHAgyVHGr#fn:6 -[12]: tmp.mJHAgyVHGr#fn:7 -[13]: tmp.mJHAgyVHGr#fn:8 -[14]: https://en.wikipedia.org/wiki/Mastodon_(software) -[15]: tmp.mJHAgyVHGr#fn:9 -[16]: tmp.mJHAgyVHGr#fn:10 -[17]: https://twitter.com/TwoBitHistory -[18]: https://twobithistory.org/feed.xml -[19]: https://twitter.com/TwoBitHistory/status/1192196764239093760?ref_src=twsrc%5Etfw -[20]: tmp.mJHAgyVHGr#fnref:1 -[21]: tmp.mJHAgyVHGr#fnref:2 -[22]: tmp.mJHAgyVHGr#fnref:3 -[23]: tmp.mJHAgyVHGr#fnref:4 -[24]: tmp.mJHAgyVHGr#fnref:5 -[25]: tmp.mJHAgyVHGr#fnref:6 -[26]: tmp.mJHAgyVHGr#fnref:7 -[27]: tmp.mJHAgyVHGr#fnref:8 -[28]: tmp.mJHAgyVHGr#fnref:9 -[29]: tmp.mJHAgyVHGr#fnref:10 diff --git a/translated/talk/20200105 Friend of a Friend- The Facebook That Could Have Been.md b/translated/talk/20200105 Friend of a Friend- The Facebook That Could Have Been.md new file mode 100644 index 0000000000..3d46287b2b --- /dev/null +++ b/translated/talk/20200105 Friend of a Friend- The Facebook That Could Have Been.md @@ -0,0 +1,238 @@ +[#]: subject: "Friend of a Friend: The Facebook That Could Have Been" +[#]: via: "https://twobithistory.org/2020/01/05/foaf.html" +[#]: author: "Two-Bit History https://twobithistory.org" +[#]: collector: "lujun9972" +[#]: translator: "aREversez" +[#]: reviewer: " " +[#]: publisher: " " +[#]: url: " " + +Friend of a Friend: The Facebook That Could Have Been +====== + +> 我把自己的网络写进 FOAF 文件，由此开启了变革。——互联网之父蒂姆·伯纳斯·李（2007） + +目前，FOAF 标准（Friend of a Friend）基本上已经不再使用了，或者说被人们忘记了，亦或者说已经被取代了[1][1]。回顾本世纪第一个十年，假如 Facebook 没有征服世界的话，FOAF 将会定义我们今天所用的社交网络。不过在开始探讨这一 web 标准之前，我想先来谈一谈纽约地铁。 + +目前，纽约地铁的唯一管理机构是 大都会运输署Metropolitan Transportation Agency，简称 MTA。MTA 对纽约市的地铁交通拥有垄断权。在纽约乘地铁必须先从 MTA 购买车票，否则属于非法行为。也就是说，MTA 在地铁行业没有任何竞争对手。 + +不过，以前可不是这样。说起来可能有些让人吃惊，纽约市的地铁交通曾由两家企业相互竞争，共同运营。跨区捷运公司Inter-borough Rapid Transit Company（IRT）的势力范围是穿过曼哈顿的线路，布鲁克林-曼哈顿运输股份有限公司Brooklyn-Manhattan Transit Corporation（BMT）管理的则是布鲁克林区内的线路，其中也有部分线路延伸到了曼哈顿。1932 年，纽约市投入使用了独立地铁系统，与 IRT 和 BMT 展开竞争，所以当时纽约共有三家不同公司控制着市内地铁交通。 + +可能会有人觉得这样的地铁运营效率不高。事实上也确是如此。由于 IRT 和 BMT 投入使用的列车宽度不同，所以在不同的运营系统之间建造换乘站十分困难。此外，乘客换乘时还需向不同的运营商支付费用，这就意味着在换乘站至少要设置两个不同的检票区域。后来，纽约市于 1940 年取缔了 IRT 和 BMT，将整个纽约市的地铁交通置于一家运营商的管理之下，不过由于此前分而治之而造成的效率低下问题时至今日依然存在：能在 BMT 的线路上运行的列车无法在 IRT 的线路上运行，因为 IRT 的线路隧道比较窄。因此，MTA 不得不同时管理这两种互不兼容的列车系统，由此带来的支出可能远比世界上其他单一隧道宽度的地铁系统要多得多。 + +IRT 和 BMT 之间的竞争所造成的历史遗留问题告诉我们，地铁系统本身就趋向于垄断经营。相比较于两家运营商相互竞争，只有一家运营商更能解决问题。乘客们虽然失去了选择的余地，但再也不用担心带了一张地铁卡却忘记了另一张的问题。 + +那么，地铁和社交网络又有什么关系呢？我在想，Facebook 是否和 MTA 一样都有自然垄断的属性呢？事实上，无论是自然垄断还是非自然垄断，Facebook 貌似确有垄断能力。当然它垄断的不是社交媒体本身（我在 Twitter 上面花的时间更多），而是垄断了与现实中认识的人之间的线上联系。Facebook 能够垄断所谓的“社交图谱”；如果我不用担心会失去与别人的联系方式，那我明天就会卸载 Facebook，因为我对 Facebook 的垄断权力感到非常气愤。不过，我却不会生 MTA 的气，即便从字面上和比喻义上来讲，纽约市地铁都是一堆焚烧着的、火舌乱窜的垃圾。说到底，我愤怒是因为我觉得不同于 MTA 的自然垄断，Facebook 的垄断属于非自然垄断。 + +我的意思是，如今 Facebook 之所以能拥有所有人的社交数据，因为它碰巧是第一个做大做强，确立巨头地位的社交平台，而不是因为其他社交平台难以或者无法与之竞争。不过，难道真是因为这样吗？许多事实告诉我们，原因并非如此。Facebook 真的是第一个社交平台吗？它提供的服务真的比其他社交平台要好吗？如果你想联系老朋友，只有 Facebook 这一个平台的话，不会方便许多吗？在 Facebook 有很多竞争对手的情况下，如果你和你男朋友 Facebook 上面的感情状态都显示“交往中”，但是他始终没来得及更新他在 VisageBook 上的感情状态，说不定现在上面还是他和他大学前任那时候的关系，那么这种情况意味着什么呢？人们信任的又是哪个社交网站呢？如果社交网站有很多，难道在填写信息上面不会很耗时间吗？ + +过去几年，由于中心化社交网络的缺陷暴露出来，许多人尝试构建去中心化的平台。基于开放标准，去中心化平台有望建立互通的社交网络生态（比如 [Fediverse][2]）。可惜的是，其中没有一个平台能够取代主流社交网络。一个比较明显的原因是 网络效应network effects的力量：既然每个人都在用 Facebook，那么任何想要放弃 Facebook 的人都将会付出巨大的代价。有人会说，这一点恰恰证明了社交网络属于自然垄断行业。但我想说，Facebook、Twitter等平台是自己选择封闭起来的。此外，鉴于人们已经设想出社交网络的互通性，并且付诸实践，那么封闭的社交平台引发的网络效应就无法证明社交网络具有自然垄断属性。 + +因此，在我看来，真正的问题是：之所以 Facebook 等平台到现在仍是主流社交网络，仅仅是因为网络效应，还是说与只有一家运营商的地铁系统一样，单一的主流社交网络效率更高？ + +最后，这些问题让我想起了 FOAF。尽管人们似乎已经忘记了 FOAF 标准，但是早在 Facebook 出现之前，人们就尝试使用 FOAF 建立开放的、去中心化的社交网络。如果过去有哪个去中心化社交网络有机会早于 Facebook 占领如今 Facebook 驻守的阵地，那只可能是 FOAF。考虑到世界上大部分人都有 Facebook 账号，而且了解 FOAF 的人相对较少，我们是否可以得到如下结论：同地铁一样，社交网络也有中心化和自然垄断的性质；亦或者，FOAF 项目说明，尽管去中心化社交网络可行，但由于其他原因，无法获得人们的广泛支持。 + +### 早期社交媒体的未来 + +FOAF 项目诞生于 2000 年，旨在建立一套表示个人身份以及人与人之间关系的通用标准。在今天看来，这一雄心勃勃的项目可能会让人感到惊讶，但是在上世纪末本世纪初，这样的想法再寻常不过了。当时在网络上，美国在线America Online 与 [Prodigy][3] 等封闭系统遭遇惨败。这让人很自然地想到，计算机领域的创新发展必须要保持开放、基于标准，而且这也正是网络的特点。 + +许多人认为，网络下一场重头戏会是 语义网Semantic Web。我有篇文章介绍了关于语义网概念与运行原理的设想，所以这里不再赘述。但是我会简单谈谈推动人们研究语义网技术的愿景，因为 FOAF 标准正是这一愿景在社交网络方面的应用。 + +一篇题为 [《谷歌如何击败亚马逊和易贝，朝着语义网进军How Google beat Amazon and Ebay to the Semantic Web》][5] 的文章很好地描绘了语义网这一崇高理想。文章写于 2002 年，作者是 Paul Ford。这篇文章设想了 2002 年至 2009 年的情景：通过使用语义网，谷歌取代了亚马逊和易贝，成为电商平台主导者。文章指出，在未来，如果你想买东西，比如说一把二手的马丁吉他，可以在谷歌中输入 `buy:martin guitar`。根据你的邮编，谷歌会告诉你附近哪些人在卖马丁吉他。谷歌之所以可以获取卖家及其吉他的信息，是因为它可以读取资源描述框架标记语言 RDF，该语言是语义网的核心技术，用于描述资源之间的关系。人们可以将 RDF 内容嵌入网页，能实现的用途比较广泛，比如给要卖的东西打广告。Ford 预测，随着使用这种方式搜索和售卖商品的人数增加，亚马逊和易贝将失去它们在电商领域近乎垄断的地位。如果可以搜索全网，又有谁会执着于某个封闭的数据库呢？Ford 写道，即便是谷歌，最终也会失势。因为理论上，任何一个人都可以检索网络，查阅 RDF，提供类似于谷歌的搜索功能。起码，如果谷歌打算对语义网上的每笔交易按一定比例收取费用，以此盈利，那么以后随着相关竞争越来越激烈，谷歌的抽成比例很有可能会被迫降低。 + +Ford 所设想的未来是关于 RDF 在电商领域的应用，不过 RDF 更振奋人心的地方在于，它或许可以应用于各个领域。RDF 标准以及一系列相关标准曾一度得到广泛应用，被认为可以突破基于数据库的开放软件服务面临的发展瓶颈，如同 HTML 为开放文档排版带来新的发展契机一般。 + +RDF 以及其他语义网技术唾手可得的另一个领域是社交网络。FOAF 项目最初的名字是“RDF Web Ring”，是语义网发展的产物，旨在实现语义网的设想。FOAF 自诞生之初就被人们看好，有人甚至认为，FOAF 必定会淘汰掉其他社交网站。2004 年《卫报》的一篇文章这样介绍该项目： + +> 最初是 1996 年，SixDegrees 开始运营；接着是去年，出现了 Friendster；上周是 Orkut；下周 Flickr 也会登上舞台。这些网站不胜枚举，都是为了建立社交网络。如今，它们处在互联网发展的最前沿。但是，如果它们无法提供更优质的服务，在 FOAF 标准得到广泛应用之后，它们就会很难存活下去。[2][6] + +文章继续指出，社交网络面临的最大问题就是网站数量过多。这就需要一种能够将所有这些网站连接起来的手段。可行方案就是 FOAF ，它终将变革整个社交网络。 + +根据该文章，FOAF 可将不同的社交网站紧密连接起来，实现途径有三个要点： + + * FOAF 将创建机器可读的社交数据格式，可为各个社交网站识别读取，避免让用户在不同的网站上重复输入信息。 + * FOAF 标准下，联系人Contacts（个人信息管理程序）可生成上述格式的文件，供用户在各社交网站使用。 + * FOAF 标准下，该文件可寄放在个人主页上，可为各社交网站读取。这样一来，用户只需将修改过的信息推到主页，其他平台就会同步更新。 + + + +在今天可能难以想象，但在 2004 年，至少在熟悉技术的网民和技术专栏记者看来，当时社交网络并不算少，但是每个网络的用户群体都很小。虽然与我们十分遥远，但若能考虑到这一问题，我们就会明白为什么需要建立单一标准来推动社交网络扩大用户基础。 + +### FOAF 规则 + +根据 FOAF 项目官网现有的介绍，FOAF 指的是“一种计算机语言，用于生成与人相关的各种条目的字典，条目以结构化数据的形式储存”。2000 年，FOAF 的创始人 Dan Brickley 和 Libby Miller 发表了一份关于该项目目标的文件，给出了不同的解释，强调了 FOAF 的最终目标：作为工具，FOAF 可让计算机像人类一样读取用户主页的个人信息[3][7]。FOAF 将会“帮助网络提供当前只有中心化平台才能提供的服务”[4][8]。通过为个人以及人际关系定义标准词汇表，FOAF 可以理解用户输入的内容，比如“查找医院医疗人员今天给出的推荐”，或者“查找曾与我合作过的人最近发表的文章”。 + +由于 FOAF 是标准化的词汇表，所以该项目最重要的成果莫过于 FOAF 规则。FOAF 规则规定了 RDF 类 和 RDF 属性（这里我不再解释什么是 RDF，如果感兴趣可查阅[我关于语义网的文章][4]）。RDF 的类由 FOAF 规则规定，表示要描述的对象，比如人（`Person` 类）和组织（`Organization` 类）。RDF 属性由 FOAF 规则规定，表示针对不同对象所做的逻辑声明。例如，一个人可以有一个名字（`givenName` 属性）、一个姓氏（`familyName` 属性），可能还有人格类型（`myersBriggs` 属性）以及与他人的距离或者位置信息（`based_near` 属性）。FOAF 规则的思想是，这些类和属性要足以表示人们在个人主页上显示的身份信息和朋友信息。【注：Myers–Briggs 即迈尔斯布里格斯类型指标，是一种人格类型理论模型】 + +FOAF 规则给出了 FOAF 文件的一份范例文档。该实例的格式是 XML，不过也可以使用 JSON 等格式进行编写： + +``` + + + Dan Brickley + + + + + +``` + +这份 FOAF 文件对一个人进行了描述，他的名字叫做 Dan Brickley，在 `http://danbri.org` 网站上设有个人主页，他还有个叫做“open ID”的东西，`/images/me.jpg` 表示图片，可能和 Brickley 的主页地址相关。FOAF 的元素名称都会有 `foaf:` 前缀，表示它们是 FOAF 命名空间的一部分。相应地，RDF 的元素名称前面也都会有 `rdf:`。 + +为了说明 FOAF 不限于 XML 格式，这里从维基百科摘取了一个相似的例子，格式为 JSON-LD [5][9]： + +``` + + { + "@context": { + "name": "http://xmlns.com/foaf/0.1/name", + "homepage": { + "@id": "http://xmlns.com/foaf/0.1/workplaceHomepage", + "@type": "@id" + }, + "Person": "http://xmlns.com/foaf/0.1/Person" + }, + "@id": "https://me.example.com", + "@type": "Person", + "name": "John Smith", + "homepage": "https://www.example.com/" + } + +``` + +上面这份 FOAF 文件也描述了一个人，他的名字叫 John Smith，在 `www.example.com` 网站上有自己的主页。 + +理解 FOAF 原理的最好方法可能就是使用 [FOAF-a-matic][10]，一个在线生成 FOAF 文档的工具。你可以在工具页面的表单里输入自己的相关信息，创建表示自己的 FOAF 文档（XML 格式）。FOAF-a-matic 说明了 FOAF 是如何避免在注册不同社交网站账号时重复输入社交信息的麻烦：如果每个社交网站都可以读取 FOAF，你只需要在没有注册过帐号的网站上引用你在 FOAF-a-matic 生成的 FOAF 文档，就可以注册一个新帐号了。 + +下面这个实例是我用 FOAF-a-matic 生成的，稍微复杂一些，表示我自己： + +``` + + + + + + + + + + Sinclair Target + Sinclair + Target + + + + + John Smith + + + + + + + +``` + +本例中，主要信息之前有很多其他内容，用于设置文档使用的各种 XML 命名空间。其中就有文档生成工具的信息，这样用户就能明白出了问题要向谁进行反馈。`foaf:Person` 元素给出了我的名字，电子邮箱和主页。其中嵌套了 `foaf:knows` 元素，说明我有个叫 John Smith 的朋友。 + +该例还体现了 FOAF 文档的另外一个重要功能：相互关联。还记得之前 John Smith 的例子吗？他在 `www.example.com` 上面有自己的主页。在我的这个例子中，我将 John Smith 列在了 `foaf:person` 元素里，上一级元素是 `foaf:knows`，表示我认识的人。此外，我还加入了 `rdfs:seeAlso` 元素，放了 John Smith 主页的 FOAF 文档链接。由于加入了这一链接，程序在读取我的 FOAF 文档时，就能根据该链接读取他的 FOAF 文档，查找到更多关于 John Smith 的信息。在之前 John Smith 的 FOAF 文档里，John 并没有提供任何有关朋友的信息（包括我在内），这意味着程序无法确定我们两人之间的朋友关系。但如果他加入了朋友信息，程序在读取我的文档之后，不仅会发现我，也会发现 John、他的朋友、他的朋友的朋友，以此类推，直到程序穷尽我和 John 各自的社交图谱。 + +该功能在 Facebook 用户的大家看来应该很熟悉了。FOAF 没有 `foaf:wall` 属性和 `foaf:poke` 属性，无法完美复制 Facebook 的功能。很明显，FOAF 也没有漂亮的蓝色界面，无法为用户提供可视化的 FOAF 社交网络，它只是一个词汇表。不过，Facebook 的核心功能是以分布式的方式提供的，我认为这正是 Facebook 垄断能力的关键。在 FOAF 标准下，好友可以将 FOAF 文档上传至个人主页，数字化展示他们真实的社交图谱，用户无需将个人数据的控制权交给 Facebook 这样一个中心化的数据库。要知道，由于对用户个人数据管理不当，扎克伯格在国会委员会召开之前的大多数时间都在向公众道歉。 + +### 暂时搁置的 FOAF + +浏览 FOAF 项目主页，你会发现在页面的右上角，有一张喜剧动画《飞出个未来》主角弗莱躺在休眠舱内的图片。这张图片是《飞出个未来》某一集的截图，讲的是弗莱在 1999 年不小心跌进了低温休眠舱，到了 2999 年再次苏醒过来的故事。我曾和 Brickley 在 Twitter 上简短地聊了一下，他告诉我，挂这张图片是为了告诉人们，未来 FOAF 项目将有机会重获新生，继续探索 20 世纪初关于网络运作方式的设想。 + +FOAF 绝不可能像《卫报》期望的那般变革社交网络。一些社交网站选择支持 FOAF 标准，比如 LiveJournal 和 MyOpera [6][11]。FOAF 甚至还在 2004 年霍华德·迪恩竞选总统时发挥了一定作用：一群博主和程序员合力搭建起了一个将网站连接起来的网络，称其为“迪恩空间DeanSpace”，帮助迪恩竞选，并在网站上使用 FOAF 记录迪恩的支持者和帮助迪恩竞选的志愿者[7][12]。不过，今天人们了解到 FOAF 主要还是因为它是 RDF 应用最为广泛的词汇表之一，而 RDF 正是现代网络的一个重要标准。如果在今天还能用到 FOAF 的话，可能就是谷歌“知识面板knowledge panels”所用技术的原型。知识面板是在用谷歌搜索时，出现在搜索结果右侧的一小块内容，会提供搜索关键词的基本信息。谷歌为推行其知识面板，使用了语义网项目的“后继者” schema.org 项目发布的词汇表[8][13]。schema.org 用来描述人物的词汇表似乎有着 FOAF 的影子，两者的目的大多也是相同的。 + +那么，为什么 FOAF 还是失败了呢？为什么人们都在用 Facebook 呢？且不提 FOAF 只是一个简单的标准，没有 Facebook 那么丰富的功能，如果 FOAF 发展势头保持下去，很有可能就会出现相关软件和应用，带来像 Facebook 那样的体验。问题是，在 Facebook 还未发展到能与之分庭抗礼之时，FOAF 这股分布式社交网络的新生力量为什么没能得到广泛应用呢？ + +恐怕这个问题的答案并不唯一，不过非要我说的话，我觉得最关键的一点是，只有在每个人都有个人网站的情况下，FOAF 才有意义。19 世纪 90 年代末到 20 世纪初，人们理所当然地觉得网络最终会出现这种情况，因为就我所知，互联网的早期用户多是高产的博客写手，参政的技术专家，他们都希望能有个自己的平台。但是，现实情况却是，普通用户并不愿意学习怎么搭建和运营网站。FOAF 的用户可以掌控自己的社交信息并将其推送到各类社交网络上，省去了到处注册账号的麻烦。如果你已经有了储存社交信息的个人网站，那么这个想法应该很诱人。但实际上，相比较于买域名、折腾 XML 文档，大多数人觉得填写信息、注册 Facebook 账号来得更容易些。 + +那么，这与我最初的问题（Facebook 是否属于自然垄断）有什么相关呢？我不得不承认，FOAF 的案例说明，社交网络 _的确_ 拥有自然垄断属性。 + +其实，关于用户不愿管理自己的数据这一问题，本身并没有那么重要，因为通过让普通用户在熟悉技术的用户所设置的节点上储存个人信息，[Mastodon][14] 等现代分布式社交网络已经解决了这个问题。这也表明，人们多么不愿意折腾复杂的东西。对去中心化社交网络来说，这无疑是个坏消息，因为相较于中心化网络，去中心化网络更为复杂，这是用户再清楚不过的。 + +FOAF：如果我要写一个能读取个人网站上 FOAF 数据的程序，假设 Sally 的 FOAF 文档提到了 John Smith，说他在 `example.com` 有自己的主页；Sue 的 FOAF 文档也提到了 John Smith，说他在 `example.net` 有自己的主页。在这种情况下，我应该怎么办呢？到底是只有一个 John Smith 而他正好有两个主页呢，还是这两个 John Smith 是不同的人呢？如果两个 FOAF 文档中 John Smith 的邮箱都是 `johnsmith@gmail.com`，又该怎么办呢？这种身份问题是 FOAF 的软肋。在一封 2003 年的邮件里，Brickley 写道，由于不存在而且可能也不应该存在一个“全球性的身份识别系统”，FOAF 采取的方法只能是“多元的”[9][15]。FOAF 用户的邮件地址和主页地址等部分属性具有特殊性，因为邮件地址和主页地址都是独一无二的。因此，这些内容不可能相同的属性可以将人们的多个 FOAF 文档合并起来（用 Libby Miller 的话来说，“挤”在一起）。不过这些特殊属性不存在所谓优先级的说法，所以前面 John Smith 的问题还是不好解决。换句话说，是该相信主页，判定他们不是同一个人呢？还是要相信邮件地址，判定他们是同一个人呢？我真的能够在不干扰到用户的前提下，写出一个程序，解决这类问题吗？ + +Facebook 拥有单一的数据库，不用顾虑政治性问题，有条件创建“全球性的身份识别系统”，给每个人发行独一无二的身份 ID，于是问题就迎刃而解了。 + +如果人们真的在乎对自己数据的持有权和掌控权，单是因为复杂难解应该不足以导致分布式社交网络的失败。但是 FOAF 的失败表明，人们从未重视过对自己数据的掌控权。正如一位博主所说，“所谓‘用户想要拥有自己的数据’只不过是一个想法，和实际应用没有关系”[10][16]。如果用户不够重视个人数据，无法忍受过于复杂的平台，如果中心化系统比去中心化系统更为简单易用，如果中心化系统有发展为封闭系统的趋向，借此取得成功，享受网络效应带来的巨大效益，那么社交网络确实属于自然垄断。 + +即便如此，我认为地铁系统的案例和社交网络的案例仍存在不同之处。我可以欣然接受 MTA 对地铁交通的垄断，因为我希望地铁系统本身就应该是长期垄断行业。如果纽约地铁只有一家运营商，那么它只能是政府，至少在名义上，政府比没有竞争对手的私企更加负责。但是我却不希望社交网络属于自然垄断。地铁建好了基本上就是一成不变的，但数字世界却在不断演变发展。在今天，分布式社交网络也许比中心化网络更加复杂，就好比带两张地铁卡总是比只带一张要麻烦的多。不过，在未来，互联网会发生根本性变革，那时分布式技术将会更易于使用。 + +如果未来果真如此，FOAF 可能会作为建立分布式社交网络的第一次尝试为人们记住。在企业大型数据库所驱动的中心化网络时代结束之后，分布式网络将会得到人们的长期青睐。 + +_如果你喜欢这篇文章，欢迎关注推特 [@TwoBitHistory][28]，也可通过 [RSS feed][29] 订阅，获取更多最新文章。_ + + + + 1. 请注意，这里我没有用“消亡”一词。[↩︎][20] + + 2. Jack Schofield, “Let’s be Friendsters,” The Guardian, February 19, 2004, accessed January 5, 2020, . [↩︎][21] + + 3. Dan Brickley and Libby Miller, “Introducing FOAF,” FOAF Project, 2008, accessed January 5, 2020, . [↩︎][22] + + 4. 同上。[↩︎][23] + + 5. Wikipedia contributors, “JSON-LD,” Wikipedia: The Free Encyclopedia, December 13, 2019, accessed January 5, 2020, . [↩︎][24] + + 6. “Data Sources,” FOAF Project Wiki, December 11 2009, accessed January 5, 2020, . [↩︎][25] + + 7. Aldon Hynes, “What is Dean Space?”, Extreme Democracy, accessed January 5, 2020, . [↩︎][26] + + 8. “Understand how structured data works,” Google Developer Portal, accessed January 5, 2020, . [↩︎][27] + + 9. tef, “Why your distributed network will not work,” Progamming is Terrible, January 2, 2013, . [↩︎][28] + + 10. Dan Brickley, “Identifying things in FOAF,” rdfweb-dev Mailing List, July 10, 2003, accessed on January 5, 2020, . [↩︎][29] + + + + +-------------------------------------------------------------------------------- + +via: https://twobithistory.org/2020/01/05/foaf.html + +作者：[Two-Bit History][a] +选题：[lujun9972][b] +译者：[aREversez](https://github.com/aREversez) +校对：[校对者ID](https://github.com/校对者ID) + +本文由 [LCTT](https://github.com/LCTT/TranslateProject) 原创编译，[Linux中国](https://linux.cn/) 荣誉推出 + +[a]: https://twobithistory.org +[b]: https://github.com/lujun9972 +[1]: tmp.mJHAgyVHGr#fn:1 +[2]: https://en.wikipedia.org/wiki/Fediverse +[3]: https://en.wikipedia.org/wiki/Prodigy_(online_service) +[4]: https://twobithistory.org/2018/05/27/semantic-web.html +[5]: https://www.ftrain.com/google_takes_all +[6]: tmp.mJHAgyVHGr#fn:2 +[7]: tmp.mJHAgyVHGr#fn:3 +[8]: tmp.mJHAgyVHGr#fn:4 +[9]: tmp.mJHAgyVHGr#fn:5 +[10]: http://www.ldodds.com/foaf/foaf-a-matic +[11]: tmp.mJHAgyVHGr#fn:6 +[12]: tmp.mJHAgyVHGr#fn:7 +[13]: tmp.mJHAgyVHGr#fn:8 +[14]: https://en.wikipedia.org/wiki/Mastodon_(software) +[15]: tmp.mJHAgyVHGr#fn:9 +[16]: tmp.mJHAgyVHGr#fn:10 +[17]: https://twitter.com/TwoBitHistory +[18]: https://twobithistory.org/feed.xml +[19]: https://twitter.com/TwoBitHistory/status/1192196764239093760?ref_src=twsrc%5Etfw +[20]: tmp.mJHAgyVHGr#fnref:1 +[21]: tmp.mJHAgyVHGr#fnref:2 +[22]: tmp.mJHAgyVHGr#fnref:3 +[23]: tmp.mJHAgyVHGr#fnref:4 +[24]: tmp.mJHAgyVHGr#fnref:5 +[25]: tmp.mJHAgyVHGr#fnref:6 +[26]: tmp.mJHAgyVHGr#fnref:7 +[27]: tmp.mJHAgyVHGr#fnref:8 +[28]: tmp.mJHAgyVHGr#fnref:9 +[29]: tmp.mJHAgyVHGr#fnref:10