What part of the archived web



/ > the time Machine Internet Archive is the largest and well-known archive that preserves web pages since 1995. Besides him there are a dozen other services, which are also archived to the web: the indexes of search engines and industry-specific archives like Archive-It, UK Web Archive, Cite Web, ArchiefWeb, Diigo. Interesting to know how many web pages gets into these files, relative to the total number of documents in the Internet?

It is known that the base of the Internet Archive for 2011 contains more than 2.7 billion URIs, many of them in multiple copies taken at different points in time. For example, the home page of Habra "photographed" already 518 times since July 3, 2006.

It is also known that the base reference Google five years ago overstepped the mark in trillion unique URLS, although many of the documents were duplicated. Google is not able to perform all URLS, so the company decided to count the number of documents in the Internet are endless.

As an example, "infinity" results in Google web calendar. No need to download and index all of its pages on millions of years ahead, because each of the pages is generated on demand.

However, scientists are interested at least roughly know which part of the web is archived and preserved for future generations. Until now nobody could answer this question. Experts from Old Dominion University in Norfolk spent a study and got a rough estimate.

For data processing they used web-framework Memento, which operates on the following concepts:

    the
  1. URI-R for identifying address of the original resource.
    2. the
  2. URI-M to identify the archived status of this resource at time t.

Accordingly, each URI-R can be zero or more States URI-M.

From November 2010 to January 2011 continued the experiment for determination of the share of publicly accessible pages, which are in the archives. Since the number of URIs on the Internet endlessly (see above), it was necessary to find an acceptable sample, which is representative for the entire web. Here, the researchers used a combination of several approaches:

    the
  1. sample of the Open Directory Project (DMOZ).
    2. the
  2. Random sample of URIs from the search engines, as described in the work of Ziv Bar-Joseph and Maxim Gurevich, “Random sampling from a search engine''s index” (Journal of the ACM (JACM), 55(5), 2008).

    3. the Last of the URIs added to the social bookmarking website Delicious using a generator Delicious Recent Random URI Generator. the

  3. link shortening Service Bitly, links, selected with the help of the generator hashes.

For practical reasons, the size of each sample has restricted thousands of addresses. The results of the analysis are shown in the summary table for each of the four samples.



The study showed that from 35% to 90% of all URIs on the Internet have at least one copy in the archive. From 17% to 49% URI have 2 to 5 copies. From 1% to 8% URI "photographed" 6-10 times, and from 8% to 63% URI — 10 and more times.

With relative confidence we can say that not less than 31.3% of URIs are archived once a month or more often. At least 35% of all pages in the archives have at least one copy.

Of course, the above figures do not belong to the so-called Deep web (the Deep Web), which is used to refer to dynamically generated pages from a database, password-protected directories, social networks, paid archives of Newspapers and magazines, flash websites, digital books, and other resources hidden behind the firewall, in the closed access and/or not available for indexing by search engines. some estimates, a Deep network can be several orders of magnitude larger than the surface layer.
Article based on information from habrahabr.ru

Комментарии

Отправить комментарий

Популярные сообщения из этого блога

Automatically create Liquibase migrations for PostgreSQL

Изображение
Working with Liquibase always brings confidence in the fact that the database objects are in the same condition in which you want it. It is this confidence and makes system migrations so popular databases. Today I want to tell you how to make using liquibase a bit easier. If you write a migration by hand, and your target database is PostgreSQL, then read this article, I am sure, the time spent on reading the article will be repaid with interest. I hope you already know how to work with Liquibase, so only two words describe how the system works. In order to make changes in the database, you need to create a migration file (changeset) that you need to specify in the file change log (changelog), after which the migration can be successfully applied to the target database. Undeniable and obvious benefit of this approach is the ability to rollback the changes. In my opinion, liquibase, there is only one drawback — the automatic rollback of changes only works only if the mi
Далее...

Vkontakte sync with address book for iPhone. How it was done

Изображение
the Background application on androidphone when you install the application Vkontakte is possible to use it as a provider of contacts to synchronize the contacts from Facebook with contacts in your address book. In this case stretches the avatar, date of birth, address, phone numbers, etc. When I after prolonged use HTC Wildfire S rated and are so used to this that he wanted to find a similar to the iPhone. Not found, so we took it and made his own — Synchronizer VKontakte :) It was back in November of last year. The application was done fairly quickly — it uses quite a few functions of Vkontakte API and functionality with the address book of the iPhone. But when we started to deal with the automatic matching of contacts, it was fun. Compare head-on by name, last name and phone number was easy to do, easy, and worked through it quickly. But found only 20 percent of coincidences of the contacts from what is provider in Android. Of course, only direct hits do not sui
Далее...