The Internet

The Internet began as a series of small-scale networking experiments that gradually expanded in scope. Persons developing these experiments had no clear idea that a nationwide, let alone worldwide, network of computers would result. The Internet is usually considered to have originated as a networking experiment funded by the Advanced Research Projects Agency (ARPA) of the Department of Defense, which interconnected computers at four academic institutions in California and Utah [1]. This network, ARPANET, became operational in 1969; soon, other academic institutions started connecting their networks to ARPANET. This interconnected series of networks began to be referred to as the Internet. In 1986, the National Science Foundation created NSFNET, an administrative and funding agency that supported the development of regional academic computer networks. This agency funded a private corporation, Advanced Network and Services, Inc. (ANS), to develop and operate a high-speed nationwide network, ANSNET, that interconnected the regional networks in all parts of the United States [2] (Figure 1). In early 1995, NSFNET terminated funding to ANS and awarded grants to the regional networks so that they could purchase their own Internet connections [3]. Currently, Sprint and MCI provide most of the long-distance Internet connectivity in the United States. Other countries, primarily Canada and countries in Western Europe, were developing nationwide computer networks at the same time as the United States. The interconnection of these national networks resulted in the globalization of the Internet.

View larger version (22K): [in this window] [in a new window]   Figure 1. The principal National Science Foundation network (NSFNET) in the United States. The structure for ANSNET (Advanced Network and Services network), the major high-speed Internet network in the United States through the end of 1994, is shown. Because most of the funding for this network came from the National Science Foundation, the network was usually referred to as NSFNET. The links in the network are fiberoptic cables leased from MCI that can transmit 45 millions bits of information, equivalent to 1000 pages of text, per second. In 1995, NSFNET stopped funding ANSNET. Currently, most of the long-distance Internet service in the United States is provided by telecommunications companies, such as Sprint and MCI.

Internet Access

The Internet can be accessed in many different ways. Most large academic, research, and medical institutions in the United States are connected to the Internet. Each institution determines the types of access and the services available to its members. Because these organizations typically have high-speed connections to the Internet and employ staff that can maintain computer hardware and software, persons at these organizations often have the best access to the Internet.

For many users, the Internet is most conveniently accessed through a personal computer and a modem connected to standard telephone lines. In general, three things are needed to use Internet resources: 1) an access provider that supplies the physical links to the Internet, 2) networking software (TCP/IP [transmission control protocol/Internet protocol] software) for connecting to the Internet, and 3) application programs for accessing resources on remote computers. Several Internet access providers now sell various services and access methods to individual users. It is beyond the scope of this article to describe these vendors, but lists of vendors and their products and services can be found in current books and articles about the Internet [4-8] and in advertisements in many computer magazines. A list may also be obtained by sending an electronic mail (e-mail) message to info-deli-server@netcom.com. No subject is required, and the text should be "send pdial" (without the quotes). An interesting new development is the ability to connect personal computers to the Internet through television cables. Television cables can transmit data at 3 to 10 million bits of data per second, more than 100 times faster than the fastest modems [9]. At these speeds, many new applications are possible, such as real-time video. This technology is being tested in restricted areas throughout the United States and should be available in 1996. This service is described on the World Wide Web (WWW) at http://www.home.net (see below).

From a user's point of view, the simplest way to connect to the Internet is through a dial-up connection to an Internet access provider using a modem and standard communication software. For this type of access, called a shell account, all the network (TCP/IP) software and Internet application programs are on the provider's computer, and the user accesses these programs from a remote location. The main drawback of this type of access is that users cannot run programs that display graphics or video and cannot directly download files from the Internet to their computers.

Users may decide to run the networking software and Internet applications programs on their own computers and purchase just an Internet access program (SLIP [serial line Internet protocol] or PPP [point-to-point protocol] access) from a commercial vendor. This method of access allows users to run graphic-based programs, such as Mosaic and Netscape (described below) and to download files from the Internet directly to their computers. Much Internet software can be purchased or obtained free from the Internet. Commercial Internet software can be purchased as stand-alone products or as part of larger packages. Microsoft, IBM, and Apple now offer Internet software as part of their current operating systems. All the major on-line computer services—CompuServe, America Online, and Prodigy—and some Internet access providers, such as MCI, PSI (Performance Systems International), and Netcom Online Communications, offer complete Internet packages consisting of an Internet connection, network software, and application programs.

Internet Services

The resources of the Internet are accessed by using application programs (often called Internet services) that connect to remote computers by various methods and perform different functions [4, 5, 10, 11]. The most commonly used services are FTP (file transfer protocol), telnet, gopher, and the WWW. All Internet services function in a type of computer interaction called the client-server model. An Internet resource on a remote computer is provided by a program called a server. A user runs a program on the local computer (the client) and supplies it with the name or address of the remote computer running the server. After the local computer is connected to the remote computer, the client passes a user's request for information to the server. The server returns the information to the client program, which then displays the results on the local computer. Thus, for example, a user runs a telnet client program to access a remote computer that runs a telnet server. In practice, the terms client and server are not often used, the meaning being derived from the context.

Computers on the Internet are identified by names and numbers. The computer naming system is called the domain name system. Computer names are strings of characters separated by periods that identify a computer and the networks it is connected to in a hierarchical fashion. Computer networks are grouped into categories called top-level domains. In the United States, seven top-level domains are used Table 1 and Figure 2. Outside of the United States, top-level domains are countries and are specified by two-letter codes. Computers actually communicate in terms of numbers called IP addresses. Computer names are automatically converted to their corresponding IP addresses when a name is specified in an application program (Figure 2).

View larger version (19K): [in this window] [in a new window]   Figure 2. Naming and numbering scheme for Internet computers. A computer at M.D. Anderson Cancer Center in Texas that can be reached by gopher is used as an example. Computer names are strings of characters separated by periods. The leftmost string is called the host name, and succeeding strings identify subnetworks, the network, and the top-level domain to which a computer belongs. A computer name may have several or no subnetwork names. Each computer name has an associated Internet Protocol address consisting of four numbers, each in the range 0 to 255, separated by periods. Depending on the type of network, the leftmost one to three numbers identify the network, and the remaining numbers identify the computer (the host).

Many different client software programs are available for Internet services, but they can be grouped into two general categories: graphic-based user interfaces and text-based interfaces. Graphic-based user interfaces present a full-screen display from which commands or options can be chosen with a mouse. Complex graphic images can be displayed with these programs. Apple computers and the Microsoft Windows operating system for personal computers can be used to run these types of programs. Text-based interfaces use operating systems such as DOS and Unix and require that a user type commands or choose items from a menu. Software programs for all Internet services are available in both graphic-based and text-based formats, but much of the functionality of the WWW is lost when the text-based programs are used.

In the following paragraphs on the main Internet services, examples of communication between the user and the remote computer are given in the text-based Unix format. Unix is the most widely used operating system on the Internet, and commands in other text-based systems are similar. In the examples below, messages or prompts from the remote computer appear as ordinary text, and user input appears in bold text. The last character of the user input must be followed by a hard-carriage return.

Telnet

The telnet service allows a user to log onto a remote computer and use its resources. At public sites, a login name is usually required and a password is occasionally needed. Most publicly accessible sites present the user with a menu system, and many allow the user to search for information using keywords. An especially useful telnet site that offers a service called Netlink is located at Washington and Lee University in Lexington, Virginia. To access this service, the user types on the local computer:

telnet liberty.uc.wlu.edu

When the connection is made, the remote computer requests a login and password.

login: lawlib

password: lawlib

Gopher

The gopher program, although superficially similar to telnet, allows access to a much wider range of resources. Unlike with telnet, a user only enters the name of a site that runs a gopher server program. No password or login is required. A menu is displayed after the site name is entered, but the menu items may be located on other computers on the Internet. Choosing an item causes the user's gopher client program to automatically connect to the computer on which the item is located and to download the information associated with the item to the local computer. Because gopher sites can be linked to each other, a user can access many computers on the Internet from a single site. The gopher server at the University of Minnesota, where the original gopher program was written, connects to thousands of gopher sites around the Internet. To access it, the user types:

gopher gopher.tc.umn.edu

File Transfer Protocol

The FTP service is designed to transfer files between computers. Vast quantities of documents and software for all types of computer systems are freely available at the several thousand publicly accessible FTP sites on the Internet. All FTP sites require a login and password, but at public FTP sites, the login name is "anonymous" and the password is the user's E-mail address. One example is the enormously popular Mosaic program for accessing the WWW, which can be obtained from the National Center for Supercomputing Applications at the University of Illinois. On the local system, the user types:

ftp ftp.ncsa.uiuc.edu

The remote system requests a login and a password.

Name: anonymous

password: glowniak@portland.va.gov

The password is the author's E-mail address. Files and directories at FTP sites are arranged in a manner similar to that of DOS-based systems. With text-based FTP programs, a user changes directories by typing "cd" followed by the directory name and downloads files by typing "get" followed by the file name. The command "dir" lists the contents of a directory at the FTP site.

World Wide Web

The Internet service that has generated the most excitement is the WWW. The resources at WWW sites consist of documents written in a hypertext format. These documents can contain graphics, video, and sound clips. Hypertext is used in several programs for Apple computers and in the on-line help system for Microsoft Windows. Keywords in hypertext documents are highlighted and linked to other documents that give further information on the keywords. Choosing a keyword retrieves the linked document, which itself can contain further links. As with the gopher program, the linked documents can reside anywhere on the Internet. World Wide Web documents are often referred to as Web pages.

The marked popularity of the WWW derives from a free program called Mosaic that can display the text, graphic, video, and sound capabilities of WWW documents [12-15]. Many proprietary versions of Mosaic are now available, the most popular of which is Netscape from Netscape Corporation. Mosaic and similar programs are called Web browsers. In addition to WWW sites, these programs can also access gopher, FTP, and telnet sites. Some free browser programs are listed in the Appendix.

All files on the Internet can be classified according to the method by which they are accessed and their location on a specific computer. A description of a file in this format is called the Uniform Resource Locator (URL) of a file (Figure 3). Web browsers use URLs to locate files.

View larger version (12K): [in this window] [in a new window]   Figure 3. Components of a Uniform Resource Locator (URL). World Wide Web browsers use this format to locate documents. The leftmost set of characters identifies the mode of access. For the World Wide Web, the access method is the hypertext transport protocol (http). Other common access methods are telnet, FTP [file transfer protocol], and gopher. A colon and two forward slashes follow the access method. The next entries are 1) a computer name; 2) the directories that lead to the document, with entries separated by forward slashes; and 3) the name of the document or file. If the directory path and file name are omitted, as in the second example, the main menu or Web page at the site is displayed.

Quality and Availability of Information

Security Issues

The Internet was originally designed to be an open system accessible to only a few users at academic and research institutions. The unanticipated growth of the Internet and the ability of any person with a personal computer and a modem to gain access to many computers has created security problems. Any computer on the Internet is at risk for unauthorized access by persons, often called hackers, who steal information or destroy computer files. Security issues are primarily the responsibility of network administrators, but the individual user should keep a few principles in mind. The most common form of computer break-in occurs when hackers guess passwords [16]. It is estimated that 70% of all computer break-ins occur because persons choose passwords that can be easily guessed. Passwords should not be English words and should contain some nonalphanumeric characters. Sensitive information, such as patient data, should be placed only on systems that have documented levels of security. Confidential data should be transmitted only over secure data channels or should be encoded. For average users, the main security problem is downloading computer viruses. Viruses can only infect a computer if users download a computer program and run the program on their personal computers. Files that are not programs, such as text or graphic files, present little, if any, risk for introducing viruses. Many programs can check downloaded files for computer viruses.

Information Retrieval on the Internet

Retrieving information from the Internet has often been daunting. Resources are scattered among thousands of computers worldwide, which have no logical connection. Because the Internet has no central controlling agency, there are no official registries of Internet resources. No organization is required to announce what, if any, services it offers (except for some government agencies that may be required by law or regulations to list their services). Fortunately, several organizations have voluntarily compiled indices of Internet resources and have developed free search programs that markedly simplify the task of finding resources on the Internet. In addition to the on-line search programs described below, many books about the Internet contain lists of resources. Large resource directories are also available [6, 7].

The Netlink service, mentioned previously, can be reached by telnet, gopher (netlink.wlu.edu 1020), or WWW (http://netlink.wlu.edu:1020) and contains one of the most exhaustive databases on the Internet of publicly available telnet, gopher, and WWW sites. Users can search the entire database by keyword, subject, access method, geographic location, or a combination of these variables.

Archie is a search program that locates files at public FTP sites. The most common method of accessing archie is by telnetting to a site that offers the program. The archie program is available on about 30 Internet sites. The login and password (if required) are both "archie." Archie searches for file names or strings of characters in file names. It does not search for file contents. For example, Netscape can be located by using the archie service at the Internet Network Information Center:

telnet archie.internic.net

login: archie

archie> prog netscape

The "prog" command is used to initiate a search for a keyword. The result of the search is a list of sites at which this program is available and the directory path to the program.

Veronica is a program that searches gopher menus for keywords. Veronica searches more than 5000 gopher servers, a large proportion of the publicly available gopher sites on the Internet. Approximately 15 to 20 sites offer the Veronica service, and many gopher servers also have links to Veronica sites. The gopher server at the University of Manitoba, gopher.umanitoba.ca, lists Veronica servers under the "Search Tools" item on the main menu. Users accesses this site with the gopher command, choose a Veronica site, and enter a keyword. A list of menu items in which the keyword appears is then displayed. Choosing an item connects users to its gopher site.

In January 1995, approximately 13 000 WWW sites allowed public access. Finding resources at these sites is facilitated by several search programs that do keyword searches on document titles and, more importantly, on document contents. Some of these search programs are listed in the Appendix. A catalog of all the major search programs can be found at the WWW site http://cuiwww.unige.ch/meta-index.html. This site also offers archie and Veronica searches and several other search tools. An interesting program for listing all the WWW sites at a particular institution is located at http://www.netgen.com/cgi/comprehensive.

Medical Resources on the Internet

USENET

The User's Network (USENET) is dedicated entirely to discussion groups [4-7]. It is accessed by a software program called a newsreader. Thousands of discussion groups are available in USENET; many of these are devoted to medical and health-related subjects. A user subscribes to a newsgroup through the newsreader program and can read, post, or answer messages. The user's Internet access provider must subscribe to the newsgroups so that the newsreader can have access to them. About 20 to 30 USENET newsgroups discuss medical subjects. Examples are sci.med.radiology, sci.med.immunology, and sci.med.nursing. Because anyone can read or post messages to these groups, the quality of information varies considerably from group to group.

Mailing Lists

Mailing lists provide a way to communicate with a well-defined group of persons. Messages to the group are sent to an electronic mailer that forwards the messages to all members on the mailing list. This service is often generically referred to as "listserv" (after list server), which was the name of one of the first programs to perform this function. Unlike USENET, in which messages are sent to a newsgroup, messages in mailing lists are delivered to a user's mailbox. No special program, aside from a mail program, is required to participate in a mailing list. Subscribing to mailing lists is more involved than subscribing to USENET. There are two addresses for the mailing list: one that accepts commands for subscribing, unsubscribing, and performing various help functions and one that sends messages to the list. Some mailing lists are moderated: The person who maintains the list screens messages before posting. Moderated lists tend to have more focused discussions.

Mailing lists provide some of the most varied and specific medical information on the Internet. There are hundreds of medicine-related mailing lists covering every specialty and subspecialty in medicine, as well as more general areas such as ethics, politics, and medical organizations. A catalog of nearly 400 mailing lists, with descriptions of their areas of interest and access methods, is contained in a file of general biomedical resources assembled by Lee Hancock at the University of Kansas. The file, medlst03.txt in the pub/hmatrix directory, can be obtained by anonymous FTP from ftp2.cc.ukans.edu. The number in the file name refers to the month of the last update.

Medical Subject Guides

The most comprehensive catalog of medical resources on the Internet is located at the University of Michigan. The University Library and the School of Information and Library Sciences have developed several subject-oriented guides through an ongoing project called the Clearinghouse for Subject-Oriented Internet Resource Guides. Information in these guides is arranged by subject rather than by location or type of access. The database currently contains about 160 guides, 10 of which are on medical subjects. Some guides are on specific topics, such as cancer and the neurosciences, whereas others are on general medical and health sciences resources. For example, the most current guide on clinical medical resources is by Lee Hancock and Gary Malet from the University of Kansas Medical Center. Resources in this guide are categorized primarily by disease and medical specialty. The guide also features sections on medical education, medical libraries, and other medical topics. The University of Michigan can be accessed by gopher to una.hh.lib.umich.edu. This database is more thoroughly described in the Appendix.

Medical Gopher Servers

More than 100 medical centers and organizations have publicly accessible gopher servers. Several gopher servers are particularly helpful in locating medical resources because they categorize information by subject and contain links to gopher servers at several medical centers and organizations. One such server, cwis.usc.edu, is located at the University of Southern California. The main menu contains the Health Sciences directory, which has the following subdirectories: Funding Resources, Resources by Subject, Resources by Organization, and Medical and Health-Related Electronic Periodicals. The Centers for Disease Control and Prevention's Morbidity and Mortality Weekly Report is found in this last subdirectory. Some of the listings under the Resources by Subject subdirectory are AIDS, Cancer, Drug Information, and Genetics and Molecular Biology. The Resources by Organization subdirectory lists menu items for the Food and Drug Administration, the National Institutes of Health (NIH), the American Heart Association, and the World Health Organization.

Another gopher that indexes medical resources is the server itsa.ucsf.edu at the University of California at San Francisco. The Bio and Medical Gophers directory on the main menu contains the subdirectory Medical and Health Gophers, which lists about 80 gophers for large medical centers and for specialty areas such as anesthesiology, cancer, family practice, and nursing.

Medical World Wide Web Servers

Several WWW servers have documents with links to many medical WWW and gopher servers. The WWW site at EINet Galaxy contains a document that arranges medical resources by medical subspecialty. The URL of this document is http://galaxy.einet.net/galaxy/Medicine/Medical-Specialties.html. The WWW site at Harvard, http://golgi.harvard.edu, has one of the most extensive lists of biomedical resources at WWW sites on the Internet. This list of resources (path: biopages/medicine.html) has links to approximately 400 sites and documents that cover all aspects of biology and medicine, although most of these resources are not directly related to clinical medicine.

The National Institutes of Health

The various institutes of the NIH house a large body of information and services. Several of the institutes can be reached individually with gopher or Web browsers. A complete index of these institutes and other organizations at the NIH is available by gopher at gopher.nih.gov or by WWW at http://www.nih.gov. A detailed explanation on how to reach all the resources at the NIH cannot be given here, but menus and documents are arranged in large categories that facilitate the location of information. Search programs are provided at several points throughout the gopher menus and WWW documents that can be used to find desired information. Examples of how to find resources by gopher are given below. Similar items can be seen by WWW access on various Web pages.

CancerNet is a comprehensive cancer information service. This service, which can be accessed from the Health and Clinical Information item on the main menu, offers CANCERLIT, a listing of citations and abstracts from the current cancer literature, and the Physician Data Query system, which gives up-to-date information on the cause, diagnosis, and clinical characteristics of all forms of cancer. A user with only E-mail access to the Internet can use the Physician Data Query system by sending an E-mail message to cancernet@icicb.nci.nih.gov with "help" as the body of the message. The NIH Guide to Grants and Contracts, found under the Grants and Research Information heading of the main menu, describes grants and program announcements at the NIH. Some other features include access to the Morbidity and Mortality Weekly Report, other publications from the Centers for Disease Control and Prevention, and CRISP (Computer Retrieval of Information on Scientific Projects), which is a large biomedical database containing information on research supported by the Public Health Services.

The National Library of Medicine at the NIH can be accessed through telnet, gopher, FTP, and WWW (Appendix). More than 20 databases are available through the National Library of Medicine, including MEDLINE, TOXLINE, and GenBank. One important resource at this site is Health Services/Technology Assessment Texts (HSTAT), a service of the Agency for Health Care Policy and Research. This service, located on the main gopher menu, provides a database of clinical practice guidelines on topics such as acute pain management, pressure ulcers, and evaluation and management of early human immunodeficiency virus infection. This database can also be accessed by telnet at text.nlm.nih.gov, login: hstat. The full text of these documents is contained under the Databases menu item and can be searched for keywords. These documents can be downloaded by FTP from nlmpubs.nlm.nih.gov from the hstat/ahcpr directory.

Medical Centers

Several medical centers allow public access, primarily through gopher and WWW servers. One of the most widely accessed medical services on the Internet is OncoLink at the University of Pennsylvania, which was created and is maintained be E. Loren Buhle and J. Goldwein. OncoLink can be reached by gopher (cancer.med.upenn.edu) or WWW (http://cancer.med.upenn.edu). OncoLink offers information on all aspects of cancer, including diagnosis, treatment, cause, and referral centers. Information is available for both patients and physicians. The site has a statistics area that a user can examine to determine the most frequently requested items.

The Virtual Hospital at the University of Iowa, accessed by WWW (http://vh.radiology.uiowa.edu), is oriented toward patient care and physician education. Physicians can use this site to find answers to specific clinical problems. The Virtual Hospital also offers the multimedia textbook program, which physicians can use as a learning resource and as a way to receive continuing medical education credits, and it features a bibliography of multimedia computer-based education.

One of the best gophers on the Internet for finding medical resources is located at Thomas Jefferson University in Philadelphia at tjgopher.tju.edu. The Medicine(TJU) item within the main menu is a directory that lists resources such as medical bulletin board systems, conference and seminar announcements, medical information arranged by subspecialty and disease, and one of the most extensive lists of medical gopher servers on the Internet. This last, in the subdirectory Medical gophers(TJU), arranges gopher servers into different categories that allow a more directed search for specific information.

More and more hospitals are setting up WWW servers that provide information for the general public. The Veterans Administration (VA) has several such servers. The Veterans hospitals at San Francisco (http://vairmsmac.ucsf.edu) and Long Beach, California (http://www.long-beach.va.gov), feature documents that show the medical center, describe the facility and services, and provide telephone numbers and hours of operation for various departments. A WWW server at http://www.va.gov describes the VA system and gives information about general VA services and benefits.

One of the best examples of the use of the Internet as an education resource is the availability of sites on the WWW that offer teaching files in nuclear medicine and radiology. Three sites have especially good material: Washington University in St. Louis, Missouri (http://gamma.wustl.edu), Harvard University in Cambridge, Massachusetts (http://count51.med.harvard.edu), and the University of Washington in Seattle (http://www.rad.washington.edu). Teaching cases are presented by organ system, disease category, or imaging method. Computed tomographic, magnetic resonance, radiographic, and nuclear medicine images are accompanied by explanatory text and references. A similar example in pathology is the Internet Pathology Laboratory at the University of Utah in Salt Lake City (http://www-medlib.med.utah.edu/WebPath/webpath.html).

Conclusion

Top Conclusion Author & Article Info References

In the last decade, the medical field has witnessed the gradual emergence of electronic networks and bulletin boards as important resources in research, education, and clinical practice. Within the last 2 years, the Internet has rapidly and unexpectedly became a widely used tool for electronic communication and resource sharing. The developments I describe are merely the beginning of new methods of computer-mediated long-range interactions in medicine. Real-time video, audio capabilities, and interactive services are relatively rudimentary at present, but the technology for these services is maturing rapidly and will be increasingly implemented on the Internet. The Internet is creating information and communication spaces that are removing the traditional boundaries of time and location; it is truly creating a "global village" [17].

Appendix

Entries in this appendix are categorized by the method by which they are accessed and by the site at which resources are located. For WWW sites, the URL format is used. To access these entries, enter the complete URL on a Web browser. (See Figure 3 for a description of URLs). For gopher sites, the site name followed by the directory path, when appropriate, is given. To access these entries, enter the site name for a gopher program or "gopher:/site name" on a Web browser. After connecting to the site, successively choose the directories listed. The directory names are printed as they appear on displayed menus. These directory paths are not the URLs for the gopher documents.

For readers who would like to download and install a Web browser, several browsers that can be obtained by anonymous FTP are listed in that section. The selections were current at the time this article was typeset, but the reader should examine the directory in which the program is listed before downloading to obtain the most up-to-date version of the program. Documentation files ("readme" files) should be downloaded for explanation of installation. To run Mosaic for Microsoft Windows, the w32s125.exe program must be installed first. A complete list of Web browsers is located at the WWW site http://www.yahoo.com/Computers/World Wide Web/Browsers (Table 2).