Reidar Conradi, IDI, NTNU
Reidar.Conradi@idi.ntnu.no, reidar.conradi@yahoo.no

Reidar Conradi, IDI, NTNU, 2 Jan. 2008:
Episodes in and Comments on the Development of Internet

- NB: p.t. not quite complete??!!

Preface

This document started as some personal Comments on Terje Rasmussen's book on Internet decision making processes (2007), see Appendix X and [Rasmussen07]. This endevor, however, soon demanded further and delightful "studies" in odd hours to better acquaint myself with the Internet TCP/IP protocols and the story behind. My notes from the last half year therefore constitute the bulk of this document, which has been uploaded on the Web via my home web-page. Any comments on contents or form are solicited and appreciated!

0. Table of Contents

1. Introduction

2. ARPA startup and the ARPANET

3. Prelude on Packet Switching: how, why, and by whom?

4. First Wave, 1967-1972: first ARPANET development

5. Years 1965-1972, Comments on openness and independence

6. Second Wave, 1973-1990: New TCP/IP protocol

7. Technical remarks on the new TCP/IP protocol

8. Third Wave, 1991-2000: the World Wide Web (WWW, or just web)

9. Fourth Wave, 2000-now: "New times" for interactive, distributed multimedia work

10. Some current and future issues and challenges for Internet

11. Literature and References

12. Appendix A: Internet institutions

13. Appendix B: Mini-glossary of IT terms

14. Appendix C: More on Turing Machines

15. Appendix D: The Technical Computer Revolution

16. Appendix E: Gallery of Persons related to Computer- and Internet Technology

17. Appendix X: Comments on Terje Rasmussen's book on Internet decision making processes (2007)

18. Appendix Y: Comments on Thomas L. Friedman's book on the World is Flat" (2006)

1. Introduction

The Internet and the services it supports, especially email and the Web, has made IT essential for most human and societal activities. And it has happened in only 30-40 years. What can be learned from this development, seen from a philosophy of science perspective?

For instance, great innovations can never be planned. Just think of agriculture, the wheel, the alphabet, iron production, typesetting of books, electricity, telephone, computers, mobile phones, and similar important innovations. Both the Internet and Web were developed for other purposes than their actual use today, respectively, mutual use of computer resources (telnet login service) and navigation in hypertext-connected, scientific documents at CERN (the Web).

Indeed, the Internet and Web together rank as one of mankind's most applicable and penetrating technologies. New work modes and usage patterns pop up every day, like net shops and e-businesses, web-search applications, and "social computing" phenomena like Open Source, Wikipedia, Youtube, and Facebook. The difference between information consumers (watching common TV series) and producers (writing own blogs) is diminishing. Furthermore, the fusion of Internet and mobile technologies proceeds at full speed, with GBs of local storage and 10 GIPS computing capacity in mobile handsets.

Berkeley professor Manuel Castells claims in an invited keynote to ICSE'2001 (International Conference on Software Engineering in 2001 in Limerick, Ireland), that the Web has reduced the cost of becoming an international organization by two orders of magnitude. We are now witnessing a massive turn-around in how we interact and organize production and consumption of both hard and soft artifacts.

The work mode of the Internet creators was a combination of extreme openness and rough consensus making. Yet most established stakeholders, like the large IT and Telecom companies (IBM, ATT), standardization bodies (ISO, CCITT), and international organizations (UN, ITU), were passive. This was due to a mismatch between their perceived technical and economic interests and the potential benefits of the new net protocols. Now, many private and public actors want to have a hand on the steering wheel. So, who should control and evolve common and international technology that represents the core infrastructure of our societies? That is, how to combine peer-driven technological skills from above, with democratic representation from below?

As a tribute to the creators of the Internet and web, let us paraphrase Churchill's speech to the RAF pilots during the battle for Britain in 1940-1941: Never have so few meant so much to so many so fast.

2. ARPA startup and the ARPANET

A comment on terminology: we will refer to ARPA, not DARPA as it now is called. The name has changed as follows: ARPA (starting in 1958), DARPA (1972), ARPA (1993), and finally DARPA (1996). Sources: [ArpaNetxx] and [Leiner99] for more information and inside stories.

3. Prelude on Packet Switching: how, why, and by whom?

4. First Wave, 1967-1972: first ARPANET development

5. Years 1965-1972, Comments on openness and independence

6. Second Wave, 1973-1990: New TCP/IP protocol

See Policy issues in 2nd Wave.

7. Technical remarks on the new TCP/IP protocol

8. Third Wave, 1991-2000: the World Wide Web (WWW, or just web)

9. Fourth Wave, 2000-now: "New times" for interactive, distributed multimedia work

Basic ICT infrastructure: we have got cheap and powerful laptop PCs, PDAs/ mobile-phones, Internet, LAN, wireless WiFi and fast UMTS/GPRS nets to stay connected with high bandwidth at nearly at all times and places (nomadic computing), not to mention web++ on these devices.

Construction tools: In addition comes a large set of common and partly freeware construction tools for e.g. text processing, spreadsheets, graphics/images/video etc. - and with use of standardized data formats and transfer protocols. For instance, a TV director can now compose and edit a whole TV program on a normal laptop, sitting on a street cafe, and sending the final program via the local WiFi net to the "home TV station" for further broadcast or network dissemination. Twenty years ago, such an editing process would have required a "studio" of the size of a 50m2 apartment. Another example: due to the penetration of laptop PCs and local WiFi nets, our university's specialized computer rooms and reading rooms stand almost empty. The students sit "everywhere" in small groups, and work and interact through their computers.

All are equal in the name of ICT: The pervasive access and versatile use of ICT has caused several authors to claim that the earth is flat [Friedman06]. This means that all countries and companies now stand on an equal foot wrt. competitive selling of brainpower and in production of ICT-enabled and ICT-reliant services and products. This "flatness" will accelerate outsourcing not only of labor-intensive physical products, like clothes, but also of many services that by their nature is labor-intensive, like legal advice. Many "low-end" services have been outsourced to low-wage places like Bangalore, e.g. "call centers" to help you fill in IRS forms, "hot-line" help for PC users (e.g. how can grandmother write a postcard before Christmas?), or standard booking of plane tickets. More "high-end" services include specialized medical diagnoses and outsourced software development.

This illustrates the new interaction patterns and media made possible through the web, and enthusiasticly adopted by young people (less than 30!). The ICT support for symmetric and diverse participation seems groundbraking. Any "producer" can upload on the web their own texts, drawings, photos, speak, music, videos, and combinations of these - being written and sent from mobile devices - and later being downloaded from the web by some "consumer". Anybody can thus be a writer or graphical artist - no more "quality control" by book publishers or art galleries. The relevant web resources will be own blogs, the new FaceBook site (already subscribed by 100,000s of Norwegian users), home videos by Utube etc.

The possibilities and implications of creative commons ("dugnadsbasert felleskunnskap"), as an extension of OSS, also deserves some thought. The common artifacts can be almost anything: (young) people's blogs and email messages, arbitrary personal notes and essays, online discussions, home photos and videos, software components, construction engineering and architectural designs, books, scientific papers, ... This raises issues of privacy, copyrights and ownership, business models with attached services, quality (not offending others - use moderators?), ...

Of web-enabled tools we should not forget the newest and most spectacular innovation of them all, the textual search engines a la Google. These operate internally with a flat, textual search space - regardless any pre-existing structure (document format, contents, topic, origin, internal data organization). This search space is then (magically) "indexed" ca. once per day. The query responses are sorted after asssumed relevance, and we usually get a response time of a textual query of less than a second. The three most frequent search words are "sex", "god" and "work" - in that order [Friedman06]. For instance, it takes me shorter time to use google to get the phone number of my colleague down the hall (searching among 20 billion webpages) than to navigate and search in the university's web portal. Or I can easily find the home address of my colleague through her phone number.

A more sophisticated use is to analyze the text of new articles in the Medline database, where 2100 are added every day (!). It is infeasible even for a top medical researcher to keep updated in her field of specialization and interest. Dr. Rune Sætre, a NTNU Ph.D. graduate from 2006, made a text-comparison tool called GENETUC [Sætre06] to analyze in which articles and by which authors the same human gene were described in the Medline articles. He thereby coupled hundreds of researchers that had been unaware of each other's existence and overlapping research. This "meta-research" lead to almost ground-breaking results. So let the earth be one.

Most written knowledge from the last 15 "post-web" years are already on digital form, and thus searchable by common search engines. A similar situation is for the (mostly unstructured) internal archives of large organizations (StatoilHydro, NTNU). So they ought to apply modern search technology on their own data. For the "pre-web" books, there are ambitious scanning initiatives by Google, EU (Gutenberg project), Norway (Runeberg effort) and other actors to make available and later searchable, most older books, magazines, newspapers etc. - mainly from paper-based libraries and publishers. This raises, regrettably (?) the issues of copyrights to books less than, say, a hundred years old. So, is this medium-old literature "owned" by their living writers, their heirs, their publishers, or by the greater society (at least eventually)?

A united world: Regardless of copyrights, mankind will soon be connected to each other and to most existing knowledge - in completely new ways by just a few decades. What will this combined "access" mean? Let us live to try it out!

10. Some current and future issues and challenges for Internet

Some more technical aspects: (mainly from Cerf's talk at NTNU 21 Nov. 2007): Some policy aspects: Number of users per summer 2007 - and counting: (source: Cerf's slides of 21 Nov. 2007)

11. Literature and References

12. Appendix A: Internet institutions

13. Appendix B: Mini-glossary of IT terms

14. Appendix C: More on Turing Machines

Alan Turing proposed in 1936 a general computing machine, the "Universal Turing Machine" or UTM [Turing36]. It is essentially a finite-state machine extended with an infinite external memory. The hypothesis is, that such an UTM can simulate any computer, or generally any other UTM. Such a UTM has: Remark on UTM variability: When a UTM shall tackle a new problem or task, the finite-state machine table must be filled up or "programmed" with relevant problem-solving instructions. When merely another data set or example shall be analysed, the machine table must be initialized once with an existing program for the task at hand, while the input data on the tape varies between executions. A classical challenge will be to first put a description of some formal language on the tape and then a statement in that language. The UTM task is now to analyze the possible truthfulness of the latter statement, by expecting a Yes or No after some years!

Lastly and most importantly, Turing was in 1936 able to prove, that such a UTM may not always stop to execute its program, the so-called halting problem. However, that a UTM may simulate any other UTM or even any other computer, is p.t. only a very well justified hypothesis, but nobody has so far managed to outline a computer that cannot be simulated. So "proof" by generalization from overwhelming pluraltiy, since no falsification so far.

15. Appendix D: The Technical Computer Revolution

The above examples are all on the hardware side. The ACM now operates with 14 knowledge focus groups in their Curriculum for CS studies:
  1. Discrete Structures (DS)
  2. Programming Fundamentals (PF)
  3. Algorithms and Complexity (AL)
  4. Architecture and Organization (AR)
  5. Operating Systems (OS)
  6. Net-Centric Computing (NC)
  7. Programming Languages (PL): e.g. syntax and parsing, semantics and compiling. Languages like Lisp, Fortran, Algol, Cobol, Pascal, Modula, Ada, Smalltalk, C, C++, Java, and Php.
  8. Human-Computer Interaction (HC)
  9. Graphics and Visual Computing (GV)
  10. Intelligent Systems (IS)
  11. Information Management (IM): including databases.
  12. Social and Professional Issues (SP)
  13. Software Engineering (SE)
  14. Computational Science (CN)

16. Appendix E: Gallery of Persons related to Computer- and Internet Technology

Computer (pre-)pioneers 1930-1960

Famous ARPANET Managers

Some "ARPANET managers" in rough chronological sequence are:
  1. Joseph Carl Robnett Licklider ("Slick", 1915-1990): both engineering and psychology degree. He came to ARPA from project MAC, via BBN. He served as visionary IPTO chair in 1962-1964, with grand plans for the Intergalactic Computer Network (now Internet?). In 1968-1985 he returned as director of project MAC and professor at MIT.

  2. Ivan Edward Sutherland (b. 1938-??): Ph.D. in EECS from MIT in 1963 (25 years old), and recipient of the Turing award in 1988 for his work on Sketchpad, an interactive graphical editor. He was head of ARPA in 1964-1966, after Licklider returned to MIT and before Taylor took over. He later was at Harvard (1966-1968), University of Utah (1968-1974) which was an early ARPANET node in 1969, Caltech (1974-1978) where he founded its CS department, and recently adjunct professor at UC Berkeley (2005-2008). He co-founded Evans & Sutherland in 1968 to develop specialized hardware and software for advanced computer graphics. He also co-founded Sutherland, Sproull and Associates in 1980 for consulting services on computer graphics. The latter company was in 1990 purchased by Sun Microsystems, where he now is a Vice President and Fellow.

  3. Robert Taylor (b. 1932): BA and MA from University of Texas. IPTO leader in 1966-1970 and ARPANET reponsible 1968-1970. Later he established and managed the Computer Science Laboratory at Xerox PARC in 1970-1983 (spin-offs: 3Com, Apple, Adobe, ...). Then founded and run the Systems Research Center (SRC) of Digital Equipment Corporation in 1984-1996, until his retirement.

  4. Lawrence ("Larry") G. Roberts (b. 1937): Ph.D. from MIT in 1963, and first worked at Lincoln Labs on computer networks. Inspired by a meeting with Licklider in 1964, he finally joined ARPA in 1967, and became IPTO chair (after Taylor) and ARPANET chair in 1970-1972. The ARPANET project was approved by ARPA on June 21 1968, whereupon Roberts hired the future designer of TCP/IP, Robert E. Kahn, who had worked on the Interface Message Processor?? at BBN. Roberts left IPTO in 1973 for Telenet, the first packet switching network carrier, which was sold to GTE in 1979 and became part of Sprint. Since 1982, Roberts has worked in DHL, NetExpress, Packetcom, and Caspian Networks.

  5. Robert E. Khan (b. 1938): Ph.D. in 1964 from Princeton. He then worked at Bell Labs and MIT. He then had a leave of absence to work at BBN with the ARPANET system design. He joined ARPA to be the director of IPTO in 1972-1985 and main responsible for the ARPANET project in 1972-1980??. In 1986 he established the Corporation for National Research Initiatives, onto which Cerf joined in 1986-l994.
    In December 1997, President Clinton presented the U.S. National Medal of Technology to Khan and his partner, Vinton G. Cerf, for founding and developing the Internet. Similarly, he and his partner got the ACM Turing Award in 2004, regarded as the "Nobel Prize in Computer Science". Finally, Khan and Cerf got the President's Medal of Freedom in 2005 by George W. Bush.

  6. Vinton Gray Cerf (b. 1943): B.Sc. in Math/CS from Stanford in 1965. At IBM in 1965-67. MSc in 1970 and Ph.D. in 1972, both in CS, from UCLA and with prof. Kleinrock as his advisor. Worked on network protocols at UCLA in 1967-1972 and at Stanford in 1972-76. He designed TCP together with Robert E. Khan in 1973-1974, and the same two invented IP in March 1978. In 1976-1982 he worked at ARPA on Internet architecture and security. In 1982-2005 he worked for MCI, but on leave in 1986-1994 at the Corporation for National Research Initiatives (run by Khan). From 2005 a Google Vice President and Chief Internet Evangelist. He has recently been working on the Interplanetary Internet, together with NASA's Jet Propulsion Laboratory since 1998.
    Co-founder in 1992 of the umbrella organization ISOC (Internet SOCiety), to serve as a civil and formal "anchor" on top of IAB, ITEF and ICANN and to create public awareness on Internet issues. In 2000 he volonteered as chairman of the board in ICANN, and in 1999 and 2001-2007 as ICANN board member.
    In December 1997, President Clinton presented the U.S. National Medal of Technology to Cerf and his partner, Robert E. Kahn, for founding and developing the Internet. Similarly, he and his partner got the ACM Turing Award in 2004, regarded as the "Nobel Prize in Computer Science". Finally, Cerf and Khan got the President's Medal of Freedom in 2005 by George W. Bush.

  7. Barry M. Leiner (1945-2003): Ph.D. in 1973 from Stanford. At ARPA he managed the Internet Project in 1980-85.

  8. Stephen Wolff (b. ??), at NSF: Ph.D. in 19xx from XX. When ARPA gradually reduced its ARPANET effort in 1983-1990, NSF gradually stepped in on the Internet side. Wolff from NSF managed the operative and public Internet in 1985-95, after which it was deemed sufficiently commercialized. NSF contributed 20 mill. USD for software development at US computer manufactors in 1985-1995 to support TCP/IP protocols on their computers. NSF also upgraded the net to serve as communication channels to and between US supercomputing centers.

More network pioneers

  • David Crocker (b. 194??): PhD from UCLA around 1975. Early ARPANET pioneer and best known as the author of RFC 822 that defines the Internet protocol suite. He is cynical about the size of the IETF vs. that of the ANSI EDI subcommittee. He is now at Silicon Grahics.

  • Stephen Crocker (b. 194??): PhD from UCLA around 1975. Early ARPANET pioneer and initiated RfC and the NWG. Now working for ICANN on security issues.

    Free and Open Software missionaires

    • Richard Stallman (b. 1953): BA in Physics from Harvard in 1974, abandoned Ph.D. studies in AI at MIT in Jan. 1984 after working there since 1974.
      Developed EMACS text editor, QCC compiler, GDB debugger; initiator of the GNU (Gnu is Not Unix) "Unix project" and of the Free Software movement where he is President.
    • Eric Steven Raymond (b. 1957): Revealed confidential Microsoft documents, boasting of the quality of Linux, to the public. Open Source propagantist with the book "The Cathedral and the Bazaar", but with quite extreme attitudes to weapons and blacks.

    • OSS examples: over 150,000 software items in the www.SourceForge.net portal alone, and 10,000s in similar portals, e.g. software as:
      Lamp platform (Linux, Apache, MySQL, PHP),
      Gimp and Trolltech graphical library,
      CVC version control system, Gentoo release management system,
      Bugzilla bug reporting system.
      Netscape => Mozilla in Nov. 2004.
      New Firefox browser from a New Zealand team.

    Unix pioneers

    • Kenneth ("Ken") Thomson (b. 1943) and Dennis MacAlistair Ritchie (b. 1941): created Unix operating system in 1969 at Bell Labs on a PDP-7, later ported it to PDP-11 and gave away the system with full source code to research institutions.
      Digression: ATT tried to patent Unix in 1985, but eventually sold the Unix rights and software to SCO, who again resold it to Novell. During the patenting process ATT found out that all source code had to be made publicly available as part of the patent! ATT also discovered trivial copyright violations just by starting (booting) a desktop computer over a net of LAN-connected computers!! The patent application was therefore withdrawn.
    • William ("Bill") Nelson Joy (b. 1954): Ph.D. from Berkeley in 1983, improved Unix implementation on PDP-11 in 1977, TCP/IP on Unix, later ported Unix to VAX as the v4.1BSD delivery in 1981. Co-founder of Sun Micros ystems in 1982.
    • Linus Torvalds (b. 1969): master in informatics from Helsinki University (1988-1996).
      Creator of Linux Unix kernel in 1993 (his share is now 2%!), and indirectly of Open Source Software. Recently at Redhat which sells Unix services.

    Web pioneers

    • Tim-Berners Lee (b. 1955): Invented and implemented web at CERN in 1990-1991, later director in W3C consortium from April 1995.
    • Robert Caillau (b. 1947): co-inventor of web at CERN.

    Norwegian Internet Contributors

    • Pål Spilling (b. 1934): then at FFI, participated in 1975 in the first intercontinental performance test of TCP, i.e. to UCL in London from Stanford University, USA.
    • Dag Belsnes, UiO (b. 1941): helped proving TCP/IP protocol at Stanford in 1974.
    • Yngvar Lund, FFI and UiO (b: 1932): ARPANET promoter in Norway around 1980.
    • Harald Tveit Alvestrand, then at UNINETT (b. 1959): leader of IETF (Internet Engineering Task Force) in 2001-2005. Before that, he was a member at the Internet Architecture Board.
    The first three were among the 30 persons invited to celebrate the 30th anniversary of the Internet in autumn 2007.

    17. Appendix X: Comments on Terje Rasmussen's book on Internet decision making processes (2007)

    Re: [Rasmussen07]

    Really a delightful, scholarly and captive book! I learned a lot, and read even more from other sources afterwards, mostly taken from the Internet/Web - where else! :-)

    Below comes a page-sorted list of minor comments, mostly typos. After that comes some editorial and then some more general comments, partly pertaining to earlier policy issues: P1 (rights to decision making e.g. in ICANN), P2 (unhindered net access), P3 (bundling of net and contents) and P4 (media ownership).

    List of minor comments:

    1. p.28: "crechendo" => "crescendo".
    2. p.30: Mention also Jon Postel, Steve Crocker and David Crocker among the UCLA graduate students around 1970.
    3. p.30: Mention how RAND lost most of their public contracts, reborn as ISI.
    4. p.33: The stuff on "entscheidung" (norsk: "beslutning", ikke "skjelning") and "Turing machine" is not understandable even for experts, so please rewrite.
    5. p.37: "skal ut og fly" => "skal ut å fly", twice at the top.
    6. p.39: Name explicitly the first four IMP sites: UC Barbara, UCLA, SRI and Univ. Utah - and their projects and contact persons.
    7. p.42: "webside" => "nettside".
    8. p.4x: "webserver" => "nett-tjener (and elsewhere).
    9. p.45: Please explain "LAN".
    10. p.45: In case of a busy Ethernet-bus, say that the delay time before another bus-access request, is doubled before each retrial - in order to ease congestion.
    11. p.48: MILNET and ARPANET: split in 1983, not 1982?
    12. p.50,middle of page; "var buffer" => "fungerte som en buffer".
    13. p.54: Say that Cerf got a PhD from UCLA in 1972.
    14. p.56: "Sammentenke" => "Sammenlenke".
    15. p.61: "Tannum" => "Tanum", satellite link station in Bohuslaen, Sweden.
    16. p.63: "CEC" => "DEC"?
    17. p.72,nine lines from below: "annen funksjonalitet" => "annen ekstra funksjonalitet" =>
    18. p.91: "QoS" is a "neutral" measure of some aspects (e.g. reliability, robustness, performance, ...) of an ITC service. It is up to the service user to decide whether the reported measure is acceptable, e.g. by setting a max/min value for some QoS aspects.
    19. p.91: "header" => "hode"!
    20. p.94: "selverhvervede" => "...erverv...".
    21. p.103 Mention that UC Berkeley Ph.D. graduate, Bill Joy, co-founded Sun Microsystems in 1984.
    22. p.108: "GNUs not Unix" => "GNU's not Unix".
    23. p.120: "Open source" => "Open Source".
    24. p.120: "reliable/reliability" => "pålitelig/pålitelighet" = "The probability for incorrect behaviour/crashes (vs. documented requirements) in an IT-system in a given usage/operative context and time period". E.g. a probability of 10**(-4) means one hour of "down" time per year. Very seldom to get under 10**(-6), and often dominated by non-IT errors ("force majeur") or human operator errors.
    25. p.120ff: "gaveøkonomi" => "dugnads.." - word not in English!
    26. p.127: "Len Kleinrock" => "Leo K..."?
    27. p.130: Network Working Group (NWG) established in 1969, not 1972?
    28. p.137: "maskin og programvare" => "maskin- og programvare".
    29. p.137: The TCP/IP split-up was in March 1978, not 1980.
    30. p.138: "en enhelhetlig" => "en helhetlig".
    31. p.138: Almost incredible bargaining with certain domain names: .am, .fm., .cd etc. !!
    32. p.153: "Trusted Computing" - forget it!
    33. p.157: "Creative Commons" i.e. "Dugnads ..." (no English term for this!)
    34. p.165: "server" => "tjenermaskin", her "adressetjener".
    35. p.169,middle: "websted" => "nettsted", "webadresse" => "nettadresse". - And elsewhere.
    36. p.176: Say explicitly that WIPO was established in 1967.
    37. p.178: "hverken" => "verken".
    38. p.180,line nine and eleven from top: "WIPI" => "WIPO"?
    39. p.180,middle: "dagens ording" => "dagens ordning".
    40. p.182,middle: "nettets kode" is ambiguous: "programvare" or "codeks"?
    41. p.184,bottom: "kode er politikk" - similar ambiguity.
    42. p.185,middle: "mindr" => "mindre".
    43. p.186,top: "åpen arkitektur best effort" => "åpen arkitektur, best effort" =>
    44. p.186: Confusing use of the word "kode". "kode" in the meaning of formats/structures exists on all three levels.
    45. p.186,bottom: "Ogsp" => "Også".
    46. p.187: Say that in the telecom network, most advanced logic and features lie in the inner, kernel layers - but opposite in the Internet with its enduser-to-enduser architecture, see discussion in [Saltzer84]. So are all the recent QoS features posed to the Internet (price-dependent quality, high streaming efficiency, flexible security/authentification policies, ...) really implementable by mere add-ons in the outermost Internet layers?
    47. p.191: "digitale skiller" => "digitale klasseskiller" (e.g.)?
    48. p.191: -------- Noter: (mainly not commented)
    49. p.200,note 47 on object-oriented (OO) programming by Dahl and Nygaard:
      "Deres ide var at programmeringen kunne gjøres enklere og mer fleksibelt ved at små subrutiner med hukommelse kunne brukes i programeringen. Tanken var helt uvant og brøt helt med den sekvensielle skritt-for-skritt-prosessen som programmeringen innebar ...".
      - Well, most non-OO subroutines (like those in FORTRAN) are organized in a rather "flat" name space. I will use the word "linear", not "sequential", about such a list of subroutine definitions. In OO languages, the subroutines are defined inside classes that stand in a hierarchy, so the flat name space is gone. But most non-OO and OO languages are still supporting mostly sequential (not parallel) execution of subroutine calls. So your text is imprecise.
    50. p.200: "Programmene Simula68 og ..." => "Programmeringsspråkene Simula67 og ...".
    51. p.200,note 71: "Java(CLI)" => "Java(C#)" - pronunced C-sharp.
    52. p.210: -------- Litteraturliste:
    53. p.211: "CarpenterB" => "Carpenter B".
    54. p.212: Daling etc.: a PhD thesis?
    55. p.212: Gilder ... Metcalfes Law" => Gilder ... Metcalfe's Law".
    56. p.213: "Hamm, Ingrid, Macel" => "Hamm, Ingrid og Macel".
    57. p.214: "Jackson ... Technologogical" => "Jackson ... Technological".
    58. p.215: "Moody ... WIRED" => "Moody ... Wired" (as earlier)?
    59. p.215: "Moscovitis .. et.al." => "Moscovitis .. et al.".
    60. p.215: "Postel Jon, og" => "Postel, Jon og".
    61. p.215: "Raymond ... revolutionary" => "Raymond ... Revolutionary".
    62. p.215: "Reed, David P. Jerome" => "Reed, David P., Jerome".
    63. p.216: "Reid ... 1000 Day" => "Reid ... 1000 Days".
    64. p.216: "Reingold ... Cambiridge" => "Reingold ... Cambridge".
    65. p.216: "Saltzer Jerome H et al." => "Saltzer, Jerome H et al.".
    66. p.216: "Segaller ... LLC" (Limited Liability Company) => "Segaller ... L.L.C." (as earlier?
    67. p.216: "Streeter ... incoporating" => "Streeter ... incorporating".a
    68. p.218: ------------- Ordliste:
    69. p.218: "BBN Bolt Beranek and Newman" => "BBN Bolt, Beranek and Newman".
    70. p.219: "GNU ..." - NB: It is Linux that is the kernel!
    71. p.220: "NCP ... ARPANETs program for" => "NCP ... ARPANETs protokoll for".
    72. p.221: "Nettsted En samling nettsteder ..." (circular def.) => "Nettsted En samling nettsider ...".
    73. p.221: "Node En side ..." => "Node En nettside ...".
    74. p.221: "Operativsystem ... ustyr" => "Operativsystem ... utstyr".
    75. p.223: ------------- Register (over personer og organisasjoner):
    76. p.224: Heart, Frank (BBN employee): please include.
    77. p.225,228: IANA: placed wrongly under letter "T".
    78. p.225: ISP (... Providers) => ISP (... Provider).
    79. p.226: Joy, Bill (Sun co-founder): please include.
    Three editorial comments:
    1. Protocol example: make a static one to show OSI-style layering, and a dynamic one to demonstrate how a single message is split up in standardized packets, which are sent through the layers, cf. above.
    2. The glossary ("ordliste") should couple persons better to institutions, and inversely. Ex. Vint Cerf worked at ... (whole list?).
    3. Make a good timeline / chronology with major events indicated.
    Four more general comments on policies:
    1. Discuss innovation vs. technology - to what degree can such be "planned"?
    2. Point P1 on technocratic vs. democratic control: Discuss different stakeholder roles, e.g. technology developer (academic researcher, hacker, IT industry), international standardization organ (ITU, ISO), international "government" (UN, EU), national government (political actor in office, ministry), national control body (ICANN, FNC) Internet provider (Telenor), content provider (Disney), end-user (any citizen or organization), ...
    3. Point P2 and P3 on "bbundling": Discuss overlapping roles between infrastructure/Internet providers (CanalDigital) and cultural agents (bookstores) vs. content providers: split these to avoid conflict of interest (NRK, TV2, Digitalt bakkenett, CanalDigital).
    4. Point P4 on media power structure: Discuss media owning or controlling other media (Schibsted, TV2, A-pressen).
    So on points P2-P4, stricter national legislation seems in place.

    18. Appendix Y: Comments on Thomas L. Friedman's book on the World is Flat" (2006)

    Re: [Friedman06], but also see [Hira05]. ++??

    - reidar

    www.idi.ntnu.no/~conradi/internet-history.html
    RC: 28 Nov. 2007, 10 Dec. 2007, and 1 and 4 Jan. 2008.