Evolution and the World Wide Web

The rapid growth in size of the World Wide Web has been noted. It is also apparent that the Web is gradually increasing in complexity, with new higher-level features emerging: examples include a variety of search engines such as the WebCrawler, the Best of the Web competition based on mass voting, and the WHO's On-line hyperbiographical database.

One possible reason for the success of WWW on both these fronts is that it seems to contain components which provide the prerequisites of evolution, that is, variation, reproduction, and selection. The details no doubt differ from the structure of biological evolution, but one could draw a rough, somewhat artificial parallel as follows.

Biology
WWW
Individual
URL
Genotype
Document or CGI program pointed to by URL
Environment
Human WWW community
Reproduction (asexual: multiple individuals with same genotype)
Copying of URLs into multiple documents and hotlists
Selection by environmental forces
Humans copy URLs pointing to documents they value
Genes producing well-adapted individuals spread
Multiple copies of popular URLs bring more visitors to their documents
Population explosion exhausts resources
Too many URLs followed to the same document swamp the server
Variation by mutation creating new genotypes
Humans improve their documents and write more
Mixing of genotypes in sexual reproduction
Humans get ideas from each others' documents and make new (better?) ones
Note that in WWW there is not much concept of phenotype: all URLs for a given document are similar (except for moved documents, misspelling, and omitted trailing slash, which are often fatal!)

Note also that unlike (presumably) in biological evolution, variations should generally be beneficial, because of the intelligence and the conscious goal-directed behaviour of the humans who develop WWW documents.

At least one case of server swamping caused by a very popular page (WWWW) has already been documented. Such server exhaustion may act as a counterweight to another human motivation missing from the parallel: it's nice to see your site visited. But this should not discourage people from paying their favourite pages the compliment of quoting their URLs so that others can also enjoy them.

The details of the above analogy are arguable and not very serious. What is important is that characteristics inherent in WWW seem likely to make it a remarkably effective medium for development (variation), communication (reproduction) and selection -- that is, for the evolution of memes, methods, and ideas -- and also, we may hope, for the emergence of interesting and useful complexity.


JSR, Sydney Mathematics and Statistics, 28 Jun 1994 (amended 29 Jun 1994) . . . SMSsearch