出版社: Plume
副标题: How Everything is Connected to Everything Else and What It Means for Business, Science, and Everyday Life
出版年: 2003429
页数: 294
定价: GBP 11.24
装帧: Paperback
ISBN: 9780452284395
内容简介 · · · · · ·
A cocktail party. A terrorist cell. Ancient bacteria. An international conglomerate.
All are networks, and all are a part of a surprising scientific revolution. AlbertLászló Barabási, the nation's foremost expert in the new science of networks, takes us on an intellectual adventure to prove that social networks, corporations, and living organisms are more sim...
A cocktail party. A terrorist cell. Ancient bacteria. An international conglomerate.
All are networks, and all are a part of a surprising scientific revolution. AlbertLászló Barabási, the nation's foremost expert in the new science of networks, takes us on an intellectual adventure to prove that social networks, corporations, and living organisms are more similar than previously thought. Grasping a full understanding of network science will someday allow us to design bluechip businesses, stop the outbreak of deadly diseases, and influence the exchange of ideas and information. Just as James Gleick brought the discovery of chaos theory to the general public, Linked tells the story of the true science of the future.
作者简介 · · · · · ·
艾伯特拉斯洛·巴拉巴西是圣母院大学教授，主持对于复杂网络的研究。他在多个领域的开创性贡献也屡屡见诸媒体，广受赞誉。他出生于特兰西瓦尼亚，现居住在印第安纳州的南本德（south Bend）。
喜欢读"Linked"的人也喜欢的电子书 · · · · · ·
喜欢读"Linked"的人也喜欢 · · · · · ·
Linked的话题 · · · · · · ( 全部 条 )
Linked的书评 · · · · · · ( 全部 58 条 )
互联网：无组织的组织力量
《链接》一书的详细目录 & 评论
这篇书评可能有关键情节透露
全文链接：http://c2blog.blog.163.com/blog/static/5978008120080181541447 1.起源 最早知道“无尺度网络”这一概念，是在《科学美国人》中文版的2003.7期看到的，之后一直对各种网络的结构感兴趣。 上个月买来了《链接：网络新科学》这本书，才发现正是那篇文章的作者著的... (展开)总体还行，对翻译不太满意
Scalefree 奠基人Barabasi的经典之作
复杂网络如何影响商业社会
> 更多书评58篇
读书笔记 · · · · · ·
我来写笔记
junior (Nothing in heaven)
Milgram's six degrees and the Web's nineteen degrees suggest that small separations are common in just about every network scientists have had a chance to study. Indeed, species in food webs appear to be on average two links away from each other; modules in the cell are separated on average by three chemical reactions; scientists in different fields of science are separated by four to six coaut... (1回应)20181212 17:52
Milgram's six degrees and the Web's nineteen degrees suggest that small separations are common in just about every network scientists have had a chance to study.
Indeed, species in food webs appear to be on average two links away from each other; modules in the cell are separated on average by three chemical reactions; scientists in different fields of science are separated by four to six coauthorship links; and the neurons in the brain of the C. elegans worm are separated by fourteen synapses. Internet, a network of hundreds of thousands of routers, has a separation of ten.
P.35 Why? Consider a network in which the nodes have on average k links. This means that from a typical node we can reach k other nodes with one step. There are, however, k^2 nodes two links away and roughly k^d nodes d links away. Therefore, if k is large, for even small values of d the number of nodes you can reach can become very large. This explains why the average separation is so short in most networks.
The origin of small separation is a logarithmic term. The logarithm of even a very large number is rather small. The logarithm shrinks huge network and create small world around us.
p.37 The six/nineteen degrees phrase is deeply misleading because it suggest that things are easy to find in a small world. This could not be further from the truth! Even if it takes only one second to check a document, it would still take over 300 million years to get all documents that are nineteen clicks away!
The trick, of course, is that we do not follow all links. Rather, we used clues. By interpreting the links, we avoid having to check all the pages within nineteen degrees . While this method seems to be the most efficient, it almost always fails to find the shortest path.
1回应 20181212 17:52 
junior (Nothing in heaven)
P.21: The premise of random network model is deeply egalitarian: We place the links completely randomly; thus all nodes have the same chance of getting one. If the network is large, despite the links' completely random placement, almost all nodes will have approximately the same number of links. One way is to interview all guests as they leave the cocktail party, asking them how many acquaintan...20181210 23:49
P.21: The premise of random network model is deeply egalitarian: We place the links completely randomly; thus all nodes have the same chance of getting one.
If the network is large, despite the links' completely random placement, almost all nodes will have approximately the same number of links.
One way is to interview all guests as they leave the cocktail party, asking them how many acquaintances they made. When everybody leaves, we can draw a histogram by plotting how many of the guests have one, two or exactly k new acquaintances. For random network model of Erodos and Renyi the shape of the histogram was derived and proved exactly in 1982, one of Erodos's students, Bela Bolobas, professor of mathematics at the University of Memphis in the United States and Trinity College in the United Kingdomg. The result shows that the histogram follows a Poisson distribution. A poisson distribution has a prominent peak, indicating that the majority of nodes have the same number of linkes as the average node does.
It predicts that it is exponentially rare to find someone who deviateds from the average by having considerably more or fewer links than average. Therefore, we end up with an extremely democratic society.
ER's random universe is dominated by the average.
P.23 They never planned to provide a universal theory of network formation. They were far more intrigued by the mathematical beauty of random networks than by the model’s ability to faithfully capture the webs nature created around them.
回应 20181210 23:49 
hedgehog (靡不有初，鲜克有终)
As long as we thought of networks as random, we modeled them as static graphs. The scalefree model reflects our awakening to the reality that networks are dynamic systems that change constantly through the addition of new nodes and links. The fitness model allows us to describe networks as competitive systems in which nodes fight fiercely for links. Now BoseEinstein condensation explains how som...20130106 11:15
As long as we thought of networks as random, we modeled them as static graphs. The scalefree model reflects our awakening to the reality that networks are dynamic systems that change constantly through the addition of new nodes and links. The fitness model allows us to describe networks as competitive systems in which nodes fight fiercely for links. Now BoseEinstein condensation explains how some winner get the chance to take it all. Do the advances obtained by acknowledeing fitness toss out the scalefree model? By no means. In networks that display fitgetrichh behavior, competition leads to a scalefree topology. Most networks we have studies so far  the Web, the Internet, the cell, Holloywood, and many other real networks  belong to this category. The winner shares the spotlight with a continuous hierarchy of hubs. Yet BoseEinstein condensation offers the theoretical possibility that in some systems the winner can grab all the links. When that happens, the scalefree topology vanishes. So far among real systems, only the operations system market, with Microsoft as its dominating hub, appears to fit the bill. Are there other systems out there displaying a similar behavior? Very likely. It will take some time, however, to recognize them all.
回应 20130106 11:15 
Erdos and Renyi and its clusterfriendly extension by Watts and Strogatz both insisted that the number of nodes with k links should decrease exponentially a much faster decay than that predicted by a power law. Network of pornstars and Hollywood proved that the size does not always matter. The truly center position in networks is reserved for those nodes that are simultaneously part of many la...
20110316 09:34
Erdos and Renyi and its clusterfriendly extension by Watts and Strogatz both insisted that the number of nodes with k links should decrease exponentially a much faster decay than that predicted by a power law.Network of pornstars and Hollywood proved that the size does not always matter. The truly center position in networks is reserved for those nodes that are simultaneously part of many large networks Peaked distribution vs. power law Road map vs. airline mapScale (characteristic note) in random networks vs. hierarchy structure of power lower (real network)scale free80/20 rule and the fact that the networks behind the web, hollywood, scientists and the cell and many other complex systems all obey a power law allowed us toe paraphrase Pareto and claim for the first time that perhaps there were laws behind complex networks.In physics: how does order emerge from disorder?Hubs the consequences of power laws a hint of selforganization and order.Despite the diversity most real networks share an essential feature: growth. It ended up dethroning the first fundamental assumption of the random universe: its static character. Also we abandon another assumption inherent in random networks: democratic character. We find that real networks are governed by two laws: growth and preferential attachment (rich get richer phenomenon).Static versus growing; random versus scalefree, structure versus evolution In terms of topology all networks fall into one of only two possible categories:Scalefree topology: a fitgetrich behavior. We have a hierarchy of nodes whose degree distribution follows a power lawWinner takes all: not scalefree. Destroys the hierarchy of hubs characterizing the scalefree topology.Vulnerability due to interconnectivity A significant fraction of nodes can be randomly removed from any scalefree network without its breaking apart. A property not shared by random networksThe web of life determines whether a cell …”there are no good or bad genes, but only networks that exist at various levels” P181The full weblike molecular architecture of a cell is encoded in the cellular network, a sum of all cellular components, connected by sll physiologically relevant interactions. P183(SYSTEM AND SUB system)回应 20110316 09:34

Erdos and Renyi and its clusterfriendly extension by Watts and Strogatz both insisted that the number of nodes with k links should decrease exponentially a much faster decay than that predicted by a power law. Network of pornstars and Hollywood proved that the size does not always matter. The truly center position in networks is reserved for those nodes that are simultaneously part of many la...
20110316 09:34
Erdos and Renyi and its clusterfriendly extension by Watts and Strogatz both insisted that the number of nodes with k links should decrease exponentially a much faster decay than that predicted by a power law.Network of pornstars and Hollywood proved that the size does not always matter. The truly center position in networks is reserved for those nodes that are simultaneously part of many large networks Peaked distribution vs. power law Road map vs. airline mapScale (characteristic note) in random networks vs. hierarchy structure of power lower (real network)scale free80/20 rule and the fact that the networks behind the web, hollywood, scientists and the cell and many other complex systems all obey a power law allowed us toe paraphrase Pareto and claim for the first time that perhaps there were laws behind complex networks.In physics: how does order emerge from disorder?Hubs the consequences of power laws a hint of selforganization and order.Despite the diversity most real networks share an essential feature: growth. It ended up dethroning the first fundamental assumption of the random universe: its static character. Also we abandon another assumption inherent in random networks: democratic character. We find that real networks are governed by two laws: growth and preferential attachment (rich get richer phenomenon).Static versus growing; random versus scalefree, structure versus evolution In terms of topology all networks fall into one of only two possible categories:Scalefree topology: a fitgetrich behavior. We have a hierarchy of nodes whose degree distribution follows a power lawWinner takes all: not scalefree. Destroys the hierarchy of hubs characterizing the scalefree topology.Vulnerability due to interconnectivity A significant fraction of nodes can be randomly removed from any scalefree network without its breaking apart. A property not shared by random networksThe web of life determines whether a cell …”there are no good or bad genes, but only networks that exist at various levels” P181The full weblike molecular architecture of a cell is encoded in the cellular network, a sum of all cellular components, connected by sll physiologically relevant interactions. P183(SYSTEM AND SUB system)回应 20110316 09:34 
junior (Nothing in heaven)
P.21: The premise of random network model is deeply egalitarian: We place the links completely randomly; thus all nodes have the same chance of getting one. If the network is large, despite the links' completely random placement, almost all nodes will have approximately the same number of links. One way is to interview all guests as they leave the cocktail party, asking them how many acquaintan...20181210 23:49
P.21: The premise of random network model is deeply egalitarian: We place the links completely randomly; thus all nodes have the same chance of getting one.
If the network is large, despite the links' completely random placement, almost all nodes will have approximately the same number of links.
One way is to interview all guests as they leave the cocktail party, asking them how many acquaintances they made. When everybody leaves, we can draw a histogram by plotting how many of the guests have one, two or exactly k new acquaintances. For random network model of Erodos and Renyi the shape of the histogram was derived and proved exactly in 1982, one of Erodos's students, Bela Bolobas, professor of mathematics at the University of Memphis in the United States and Trinity College in the United Kingdomg. The result shows that the histogram follows a Poisson distribution. A poisson distribution has a prominent peak, indicating that the majority of nodes have the same number of linkes as the average node does.
It predicts that it is exponentially rare to find someone who deviateds from the average by having considerably more or fewer links than average. Therefore, we end up with an extremely democratic society.
ER's random universe is dominated by the average.
P.23 They never planned to provide a universal theory of network formation. They were far more intrigued by the mathematical beauty of random networks than by the model’s ability to faithfully capture the webs nature created around them.
回应 20181210 23:49 
junior (Nothing in heaven)
Milgram's six degrees and the Web's nineteen degrees suggest that small separations are common in just about every network scientists have had a chance to study. Indeed, species in food webs appear to be on average two links away from each other; modules in the cell are separated on average by three chemical reactions; scientists in different fields of science are separated by four to six coaut... (1回应)20181212 17:52
Milgram's six degrees and the Web's nineteen degrees suggest that small separations are common in just about every network scientists have had a chance to study.
Indeed, species in food webs appear to be on average two links away from each other; modules in the cell are separated on average by three chemical reactions; scientists in different fields of science are separated by four to six coauthorship links; and the neurons in the brain of the C. elegans worm are separated by fourteen synapses. Internet, a network of hundreds of thousands of routers, has a separation of ten.
P.35 Why? Consider a network in which the nodes have on average k links. This means that from a typical node we can reach k other nodes with one step. There are, however, k^2 nodes two links away and roughly k^d nodes d links away. Therefore, if k is large, for even small values of d the number of nodes you can reach can become very large. This explains why the average separation is so short in most networks.
The origin of small separation is a logarithmic term. The logarithm of even a very large number is rather small. The logarithm shrinks huge network and create small world around us.
p.37 The six/nineteen degrees phrase is deeply misleading because it suggest that things are easy to find in a small world. This could not be further from the truth! Even if it takes only one second to check a document, it would still take over 300 million years to get all documents that are nineteen clicks away!
The trick, of course, is that we do not follow all links. Rather, we used clues. By interpreting the links, we avoid having to check all the pages within nineteen degrees . While this method seems to be the most efficient, it almost always fails to find the shortest path.
1回应 20181212 17:52 
hedgehog (靡不有初，鲜克有终)
As long as we thought of networks as random, we modeled them as static graphs. The scalefree model reflects our awakening to the reality that networks are dynamic systems that change constantly through the addition of new nodes and links. The fitness model allows us to describe networks as competitive systems in which nodes fight fiercely for links. Now BoseEinstein condensation explains how som...20130106 11:15
As long as we thought of networks as random, we modeled them as static graphs. The scalefree model reflects our awakening to the reality that networks are dynamic systems that change constantly through the addition of new nodes and links. The fitness model allows us to describe networks as competitive systems in which nodes fight fiercely for links. Now BoseEinstein condensation explains how some winner get the chance to take it all. Do the advances obtained by acknowledeing fitness toss out the scalefree model? By no means. In networks that display fitgetrichh behavior, competition leads to a scalefree topology. Most networks we have studies so far  the Web, the Internet, the cell, Holloywood, and many other real networks  belong to this category. The winner shares the spotlight with a continuous hierarchy of hubs. Yet BoseEinstein condensation offers the theoretical possibility that in some systems the winner can grab all the links. When that happens, the scalefree topology vanishes. So far among real systems, only the operations system market, with Microsoft as its dominating hub, appears to fit the bill. Are there other systems out there displaying a similar behavior? Very likely. It will take some time, however, to recognize them all.
回应 20130106 11:15

junior (Nothing in heaven)
Milgram's six degrees and the Web's nineteen degrees suggest that small separations are common in just about every network scientists have had a chance to study. Indeed, species in food webs appear to be on average two links away from each other; modules in the cell are separated on average by three chemical reactions; scientists in different fields of science are separated by four to six coaut... (1回应)20181212 17:52
Milgram's six degrees and the Web's nineteen degrees suggest that small separations are common in just about every network scientists have had a chance to study.
Indeed, species in food webs appear to be on average two links away from each other; modules in the cell are separated on average by three chemical reactions; scientists in different fields of science are separated by four to six coauthorship links; and the neurons in the brain of the C. elegans worm are separated by fourteen synapses. Internet, a network of hundreds of thousands of routers, has a separation of ten.
P.35 Why? Consider a network in which the nodes have on average k links. This means that from a typical node we can reach k other nodes with one step. There are, however, k^2 nodes two links away and roughly k^d nodes d links away. Therefore, if k is large, for even small values of d the number of nodes you can reach can become very large. This explains why the average separation is so short in most networks.
The origin of small separation is a logarithmic term. The logarithm of even a very large number is rather small. The logarithm shrinks huge network and create small world around us.
p.37 The six/nineteen degrees phrase is deeply misleading because it suggest that things are easy to find in a small world. This could not be further from the truth! Even if it takes only one second to check a document, it would still take over 300 million years to get all documents that are nineteen clicks away!
The trick, of course, is that we do not follow all links. Rather, we used clues. By interpreting the links, we avoid having to check all the pages within nineteen degrees . While this method seems to be the most efficient, it almost always fails to find the shortest path.
1回应 20181212 17:52 
junior (Nothing in heaven)
P.21: The premise of random network model is deeply egalitarian: We place the links completely randomly; thus all nodes have the same chance of getting one. If the network is large, despite the links' completely random placement, almost all nodes will have approximately the same number of links. One way is to interview all guests as they leave the cocktail party, asking them how many acquaintan...20181210 23:49
P.21: The premise of random network model is deeply egalitarian: We place the links completely randomly; thus all nodes have the same chance of getting one.
If the network is large, despite the links' completely random placement, almost all nodes will have approximately the same number of links.
One way is to interview all guests as they leave the cocktail party, asking them how many acquaintances they made. When everybody leaves, we can draw a histogram by plotting how many of the guests have one, two or exactly k new acquaintances. For random network model of Erodos and Renyi the shape of the histogram was derived and proved exactly in 1982, one of Erodos's students, Bela Bolobas, professor of mathematics at the University of Memphis in the United States and Trinity College in the United Kingdomg. The result shows that the histogram follows a Poisson distribution. A poisson distribution has a prominent peak, indicating that the majority of nodes have the same number of linkes as the average node does.
It predicts that it is exponentially rare to find someone who deviateds from the average by having considerably more or fewer links than average. Therefore, we end up with an extremely democratic society.
ER's random universe is dominated by the average.
P.23 They never planned to provide a universal theory of network formation. They were far more intrigued by the mathematical beauty of random networks than by the model’s ability to faithfully capture the webs nature created around them.
回应 20181210 23:49 
hedgehog (靡不有初，鲜克有终)
As long as we thought of networks as random, we modeled them as static graphs. The scalefree model reflects our awakening to the reality that networks are dynamic systems that change constantly through the addition of new nodes and links. The fitness model allows us to describe networks as competitive systems in which nodes fight fiercely for links. Now BoseEinstein condensation explains how som...20130106 11:15
As long as we thought of networks as random, we modeled them as static graphs. The scalefree model reflects our awakening to the reality that networks are dynamic systems that change constantly through the addition of new nodes and links. The fitness model allows us to describe networks as competitive systems in which nodes fight fiercely for links. Now BoseEinstein condensation explains how some winner get the chance to take it all. Do the advances obtained by acknowledeing fitness toss out the scalefree model? By no means. In networks that display fitgetrichh behavior, competition leads to a scalefree topology. Most networks we have studies so far  the Web, the Internet, the cell, Holloywood, and many other real networks  belong to this category. The winner shares the spotlight with a continuous hierarchy of hubs. Yet BoseEinstein condensation offers the theoretical possibility that in some systems the winner can grab all the links. When that happens, the scalefree topology vanishes. So far among real systems, only the operations system market, with Microsoft as its dominating hub, appears to fit the bill. Are there other systems out there displaying a similar behavior? Very likely. It will take some time, however, to recognize them all.
回应 20130106 11:15 
Erdos and Renyi and its clusterfriendly extension by Watts and Strogatz both insisted that the number of nodes with k links should decrease exponentially a much faster decay than that predicted by a power law. Network of pornstars and Hollywood proved that the size does not always matter. The truly center position in networks is reserved for those nodes that are simultaneously part of many la...
20110316 09:34
Erdos and Renyi and its clusterfriendly extension by Watts and Strogatz both insisted that the number of nodes with k links should decrease exponentially a much faster decay than that predicted by a power law.Network of pornstars and Hollywood proved that the size does not always matter. The truly center position in networks is reserved for those nodes that are simultaneously part of many large networks Peaked distribution vs. power law Road map vs. airline mapScale (characteristic note) in random networks vs. hierarchy structure of power lower (real network)scale free80/20 rule and the fact that the networks behind the web, hollywood, scientists and the cell and many other complex systems all obey a power law allowed us toe paraphrase Pareto and claim for the first time that perhaps there were laws behind complex networks.In physics: how does order emerge from disorder?Hubs the consequences of power laws a hint of selforganization and order.Despite the diversity most real networks share an essential feature: growth. It ended up dethroning the first fundamental assumption of the random universe: its static character. Also we abandon another assumption inherent in random networks: democratic character. We find that real networks are governed by two laws: growth and preferential attachment (rich get richer phenomenon).Static versus growing; random versus scalefree, structure versus evolution In terms of topology all networks fall into one of only two possible categories:Scalefree topology: a fitgetrich behavior. We have a hierarchy of nodes whose degree distribution follows a power lawWinner takes all: not scalefree. Destroys the hierarchy of hubs characterizing the scalefree topology.Vulnerability due to interconnectivity A significant fraction of nodes can be randomly removed from any scalefree network without its breaking apart. A property not shared by random networksThe web of life determines whether a cell …”there are no good or bad genes, but only networks that exist at various levels” P181The full weblike molecular architecture of a cell is encoded in the cellular network, a sum of all cellular components, connected by sll physiologically relevant interactions. P183(SYSTEM AND SUB system)回应 20110316 09:34
这本书的其他版本 · · · · · · ( 全部4 )
 浙江人民出版社版 201381 / 456人读过 / 有售
 湖南科技出版社版 20070401 / 695人读过
 Basic Books版 200205 / 16人读过
以下豆列推荐 · · · · · · ( 全部 )
 复杂网络和社会网络分析 (marego)
 社区研究 (小白)
 互联网大潮 (浪子回头)
 interaction design (bark)
 ITP书单 (靠谱的玩伴)
谁读这本书?
二手市场
订阅关于Linked的评论:
feed: rss 2.0
0 有用 多多钱少少 20171217
感觉略有些过时。
0 有用 junior 20190423
learn how to write a popular book with academic concepts
0 有用 c h l 20190609
浓浓的古早味…
0 有用 Gandalf 20140428
One of those books that changes how you look at the world.
0 有用 哦薇安 20140816
laszlo这本书写的真是不错，很好的科普读物
0 有用 c h l 20190609
浓浓的古早味…
0 有用 junior 20190423
learn how to write a popular book with academic concepts
0 有用 多多钱少少 20171217
感觉略有些过时。
0 有用 JIDISI 20141119
成为我的导师吧，先生！
0 有用 [已注销] 20140502
inspiring