Talk:Internet

Resynchronizing

I have taken the article back to before the numerous large recent changes started. There's certainly no reason there cannot be changes. The changed text is in Internet/Alternate article.

It is my understanding, Dan, that the lede (i.e., before the first heading) is generally acceptable. You have questions about having the history, or as much of it, here. I can discuss why I believe it should be, but it may be shortened.

Some of the services you have in mind may be in the article convergence of communications, which could be merged here in whole or part. Services that Peter mentioned, such as newsgroups, are, I think, there, but certainly could be added.

It is my ruling that Comer is not as authoritative as RFCs, and I'm simply not familiar with the second book that was cited. Let's work to build from what is here, not replace it en masse. Let's make links to Internet Protocol Suite, which perhaps should become more of an architectural model, and be sure that when there is a wikilink to a protocol, and there is an article extant on that protocol, we don't get a redlink. Many of the redlinks came from writing such things as Simple Network Management Protocol (SNMP) rather than Simple Network Management Protocol, or not having the precise name of a protocol such as UDP.

Howard C. Berkowitz 04:57, 20 September 2009 (UTC)

---

Howard. Before you rolled back the article text, I stated my position why I think that is a bad idea in a new section on this talk page. Would you respond to that? Dan Nessett 05:01, 20 September 2009 (UTC)

I have done so. My rollback, however, is largely procedural -- CZ customs do not approve of major unilateral rewrites without discussion, especially when a workgroup Editor is available for active participation.

Now, within those customs and rules, if we agree on the need to change the earlier articles, we will then change the article. We do not do bold-rewrite-revert as does WP. Howard C. Berkowitz 05:09, 20 September 2009 (UTC)

Why did you delete most of the text in the Internet architecture section?

Howard,

Most of the material in the Internet architecture section is deleted. Why did you do that? Dan Nessett 17:34, 18 September 2009 (UTC)

Hmm. This is a mystery. I just looked at the history of the article and it doesn't look like Howard deleted anything. And yet, the text is now missing. I will contact CZ admins for some help with this. Dan Nessett 17:57, 18 September 2009 (UTC)

See my note on Matt's talk page. Hayford Peirce 18:19, 18 September 2009 (UTC)

Newsgroups

In the first paragraph, newsgroups/usenet should be mentioned along with email and ftp. (I don't add it myself, so that noone can claim I added content in case of an approval.) Peter Schmitt 01:21, 19 September 2009 (UTC)

Done. Dan Nessett 03:38, 19 September 2009 (UTC)

Need to expand section on history of the Internet

In some side conversations, it is noted that an article on the Internet would benefit from a discussion of its history. This would include a discussion of the technologies that led up to the internet as well as a discussion of the technologies used currently. This is a reasonable point and deserves discussion. So, let me open the discussion in this section. Those who have an opinion on this issue are hereby invited to provide it here. Dan Nessett 04:31, 20 September 2009 (UTC)

I will start this conversation by providing my view. I would expect that an article named Internet would provide a broad overview of the subject. This includes its history, its structure and its uses. I think the article would benefit from an expansion of the history section, but I think the level of detail in the article Development of the Internet is inappropriate for a high-level article. So, my position is those who are interested in expanding the history section should summarize the material in the Development of the Internet article and place that summary in the history section of the Internet article. Some have suggested rolling back the Internet article to the old material (which was renamed Development of the Internet). I am opposed to this because it misrepresents the meaning of the article title (i.e., Internet). If an article with the title Jet Planes contained only a detailed description of their history and no description of their design and use, I would say the article is misnamed. By the same argument, I think an article titled Internet that only discusses its development is misnamed. Dan Nessett 04:48, 20 September 2009 (UTC)

Well, while composing the above text, Howard decided to rollback the article to the old text. As I state above, I am opposed to this. I won't get into an edit war with Howard, because that wouldn't accomplish anything. I think the ball is now in Howard's court to explain his action. Dan Nessett 04:54, 20 September 2009 (UTC)

My action, as a Computers Workgroup Editor, was to put the text back to a version level before large rewrites were made without discussion. This is more basic than what your view or my view is of the definition of the Internet; it has to do with CZ procedure about large edits without consensus.

I am willing to discuss changes in the development section, perhaps better linking it to the material that is in the lede and perhaps moving some of it, but I am not willing to start with an essentially new article and new definition of the Internet and the scope of the Internet article. If you want to make arguments on changing what was there, in a conversational way on the talk page, I'll be happy to do so — although I don't agree to agree with rewrite of the initial paragraphs. If an architecture article is appropriate, fine -- I would note that I was an IETF reviewer of some of the draft architecture documents, as well as being directly involved in OSI architecture; this is not just a side area of my being a Computers editor as a compiler expert, but as a specialist in network architecture and network architecture standards. Howard C. Berkowitz 05:07, 20 September 2009 (UTC)

You still have not addressed the point I made. I am happy to follow CZ procedures and perhaps I violated some of these unknowingly. Let me restate my position. An article titled Internet is misnamed if all it does is cover the history of its development. That material is better named something like Development of the Internet, which is why I moved it there (I thought with your agreement, but it seems I was mistaken). So, at the very least a great deal of the material in the rolled back version of the article should be moved somewhere else. The current material is too detailed for an overview article. Dan Nessett 05:17, 20 September 2009 (UTC)

Let me ask you a serious question: what do you consider the starting date of the Internet in this article?

In my mind, it does not only cover the history, certainly not in the introduction. I have a suspicion, however, that convergence of communications may be what you have in mind. Just as a guess, you tend to think more Internet applications and I tend to think more from an ISP standpoint.

Let's start systematically. Do you have any problem with the 3-4 paragraphs before the first heading? Next, might we introduce a new next section that speaks of motivations and broad architectural assumptions (e.g., end-to-end principle) that might substitute for a more detailed development? Howard C. Berkowitz 05:27, 20 September 2009 (UTC)

[unindent]

I am about to go out of the door and don't have much time to craft a reply, so I will only add a few comments. There is no way to identify a starting date of the Internet. The concept was "hanging in the air" when I became involved back in the late 70s, when most of what eventually became the internet was still running on the ARPAnet. At that time RFCs really were "Request for Comments" and not standards, as they are today. Even at that time there was activity focusing on what eventually became each of the 4 layers in the Internet protocol architecture. Email was running and DNS was in development. So, early on applications were part of the design and implementation activity. Cerf and Kahn were developing TCP/IP, so transport and internet layer services were part of the mix. When I joined LLNL (then simply LLL) in 1977, Network Systems Corporation had just put its 50 Mbps product on the market. This preceded the commercialization of Ethernet and involved network adapters, which offloaded link layer services from the supercomputers of the time. So, link layer activity was also part of the mix. So, I would not characterize what I have in mind as limited to convergence of communications, which seems to focus specifically on Internet applications. Dan Nessett 15:06, 20 September 2009 (UTC)

Dan, I'm trying to focus on what you would change here, not in another article.

For the record, I might have come to this a little earlier and been a little closer to the development. In 1970, I was running a medical computing facility for Georgetown University Medical Center, in Bethesda, MD, with NIH and NIST, both on ARPANET, a few miles away; I then moved over to military networks for a while. My first internetworking involvement besides proprietary and military and ARPANET was in the X.25 world, and I would drop in on Barry Wessler at Telenet when it was still over a strip club on 14th Street in Washington in 1972; Larry Roberts was out doing fundraising. From 1976, I was on the Federal Telecommunications Standards Committee, and in the ANSI DISY project in 1979, which was the U.S. feeder to ISO. Simultaneously, I grew more involved in OSI as the first technical staff member at the Corporation for Open Systems, but also with TCP/IP as Vint Cerf was chair of our User Committee and, while I had met him earlier, saw him with some regularity from about 1986, when I still went to OSI meetings. Bob Metcalfe was our chairman, so we had Ethernet/link layer involvement, as well as his stage magic tricks. Vint, along with Scott Bradner and Lyman Chapin, were later book advisers of mine at Wiley. My first in-person IETF meeting was Danvers in 1995, but I had been contributing remotely long before, as well as in IEEE 802 and ISO.

I agree that it is somewhat hard to assign a date to "the Internet" as an administrative entity, but it is certainly not uncertain when ARPANET cut over to the TCP/IP protocol suite: 1 January 1983.

This article could do well to set aside some myths such as the Internet being created to be a survivable nuclear warfighting network, which is what we were doing in the Minimum Essential Emergency Communications Network (MEECN), although packet switching later came to military networks.

Again, what would you change in the starting paragraphs? Would you introduce new transitional sections following it? Howard C. Berkowitz 15:51, 20 September 2009 (UTC)

Changed name of the Internet subgroup to the Internetworking subgroup

For those who are intested, I have changed the name of the Internet subgroup to the Internetworking subgroup. I have copied the text from Internet/Alternate_article to the Internetworking subgroup main article page. This allows us to continue our discussions of the Internet article without those discussions becoming entangled with discussion of the subgroup formation. I did this because I created this subgroup and am currently its only member. I have not yet edited the main article to reflect its title, but will do that in the next few days. Dan Nessett 05:40, 20 September 2009 (UTC)

After going to bed to get some much needed rest, I think I should explain the motivation behind the change of subgroup name in more detail. We have two activities under way, which unfortunately seem to have been conflated. There is the writing and editing of the Internet article and there is the formation of an Internet (currently Internetworking) subgroup. These two activities have unfortunately collided simply because the name of the subgroup happens to be the same as that of an existing article. This was causing a great deal of confusion and unnecessary friction. Perhaps more importantly, some of us were working on the definition of the scope of the subgroup and had provided edits to what is now the Internet/Alternate_article article. I think it is important to understand that defining the scope of the subgroup is independent of working on the structure and content of the existing Internet article. In my original conception, the subgroup would focus on the design, implementation and use of the technologies/services defined by the Internet protocol architecture. As we worked on the article that defined this scope, it developed into a much more articulate description than what I originally wrote. So, I moved the subgroup name and the article content to a new name so we would not lose that work.

Since I have other things I have to handle today, I cannot provide a comprehensive discussion of this issue. However, I would like to observe that separating the deliberations associated with subgroup formation from those involved with editing the Internet article simply recognizes that two different sets of policies and procedures apply. The first activity is directed by the guidance provided in CZ:Subgroups, which roughly is 1) a Citizen forms a subgroup, 2) its main article is developed and moved toward approval, 3) once approval is obtained, the subgroup members ask for the association of the subgroup with existing workgroups. If these steps all succeed, the subgroup is established. The second activity involves editorial decisions and the crafting of text that eventually becomes part of the article main namespace. Dan Nessett 15:25, 20 September 2009 (UTC)

I will note, and for this reason am starting long-needed top-level articles on network architecture, that internetworking has a fairly well-defined meaning, a superset of the Internet, of the interconnection of dissimilar networks. The scope of the proposed subgroup is still very unclear to me.

Now, rather than confuse the main namespace further, I had taken your rewritten Internet article and put it as Internet/alternate article. "Internetworking", to me, has a quite different meaning that includes some aspects of the Internet, but applied just as well to X.25, SNA, and mu-law versus A-law digital telephony. Howard C. Berkowitz 15:56, 20 September 2009 (UTC)

I chose the term "Internetworking" because it allows the formation of a subgroup that focuses on the design, implementation and use of those technologies/services that comprise the Internet protocol architecture without the confusion that has arisen when the subgroup's name was "Internet". We have regrettably conflated two subjects: 1) the editing of an existing article in the main namespace with the title "Internet", and 2) the formation of a subgroup specified as described. Changing the subgroup's name allows these two activities to proceed in parallel and without the sort of unproductive churning that has recently occurred. The subgroup I have created, the continued existence of which depends on an approved main article titled with the subgroup's name and the association of at least one workgroup, is not focused primarily on the history or development of the Internet, although that topic would fall within its purview. Using different names for the two articles solves a significant problem and allows progress on both activities.

In regards to the definition of the term "Internetworking" here is a sample: 1) "The process of interconnecting two or more individual networks to facilitate communications among their respective nodes. Note: The interconnected networks may be different types. Each network is distinct, with its own addresses, internal protocols, access methods, and administration." This is taken from Federal Standard 1037C ("Glossary of Telecommunication Terms"), a publication of the Institute for Telecommunication Sciences, the research and engineering branch of the National Telecommunications and Information Administration (NTIA), a part of the U.S. Department of Commerce (DOC). 2) From our competitor WP, "Internetworking involves connecting two or more computer networks via gateways using a common routing technology. The result is called an internetwork (often shortened to internet)." 3) Cisco offers the following definition (http://www.cisco.com/en/US/docs/internetworking/technology/handbook/Intro-to-Internet.html): "An internetwork is a collection of individual networks, connected by intermediate networking devices, that functions as a single large network. Internetworking refers to the industry, products, and procedures that meet the challenge of creating and administering internetworks."

So, I think it is reasonable to conclude that internetworking is the activity that results in an internet. By making this change, the focus of the subgroup remains virtually the same, achieving the objective given above, viz., working on the two activities in parallel. Dan Nessett 01:06, 21 September 2009 (UTC)

I note you quote FED-STD-1037C; the emphasis is that it is a Federal Standard, to which ITS/NTIA/DOC submitted it for approval. The original 1037 came under the direction of Neal Seitz at ITS, and I was both a reviewer for ITS, and the Library of Congress representative to the Federal Telecommunications Standards Committee that approved it in 1978 or so. I think I drafted a few of the definitions. So, I think I have some fairly good insight into what and when went into that document. The update you cite is from 1996, so it's not the best of references for today's Internet. Neal is, however, absolutely authoritative on restoring Ford Mustangs.

We simply don't use WP as a reference.

Well, I am an (inactive) Certified Cisco Systems Instructor (#93005) and course developer, and, again, have some first-hand experience with how that was brought about. Those definitions largely come from tech writers. On the other hand, Fred Baker, Cisco Fellow and past chair of the IETF, was the lead editor of RFC 1812, which, while older, is still fairly definitive. The several IAB/IETF informational architectural documents, by Brian Carpenter, Randy Bush, Dave Meyer et al., might be a little more appropriate. After all, if one is trying to define the Internet, shouldn't the IETF be the primary reference? Howard C. Berkowitz 01:15, 21 September 2009 (UTC)

I think you have misunderstood why I supplied those definitions. My only objective was to support the argument that Internetworking is the activity that results in the Internet. The terms are similar enough that we can work on the main namespace article titled "Internet" in parallel with the main article of the Internetworking subgroup without those two activities inappropriately interfering with one another. Dan Nessett 01:44, 21 September 2009 (UTC)

We haven't had a situation before where there wasn't a substantial consensus, by several people, that a subgroup was needed. "Chemical engineering" is a good example, where the base article was in existence and stable. This particular situation, when we have one person pushing strongly for a subgroup for which he has a vision that may be a great one, but isn't clear to others and doesn't have a broad consensus, is something with which we have not dealt.

I don't know how best to proceed, without getting heavy-handed as an Editor. It would help, I suspect, if you'd work on this article, checking redlinks to be sure there isn't an existing article with a slightly different title. If you have problems with those existing articles, let us know.

If I were you, which I am not, I would have concentrated on trying to get consensus on changes to the Internet article first, so there were a consensus of what the Internet means. Really, I'm puzzled about the push for a subgroup; subgroups have been essentially indexing conveniences. Yes, there have been some suggestions about getting external involvement, for which there needs to be consensus whether there is a subgroup or not.

We are in the midst of charter group elections and then a charter process, and I believe we have cleared some obstacles to an Editorial Council election and, I hope, reviving. Those, and the Forums, are the place to present initiatives such as external relations -- and, in my opinion, generating sentiment for a subgroup if one is needed. Howard C. Berkowitz 02:14, 21 September 2009 (UTC)