This is an archive of past discussions about Ethernet. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.
None of the above. These designations are so often misquoted, but they are correctly 10BASE2, 10BASE5 and 10BASE-T
Similarly 100BASE-TX, 1000BASE-SX, etc.
I would correct the article now, but I'm not feeling up to learning how to rename all the relevent articles and create redirects. I would confirm, however, that I have as I type this a copy on my lap of the official IEEE Ethernet standard (2000 edition) -- all 1515 pages of it -- and I have verified the above designations against it. The general rule would appear to be that the designations are entirely uppercase, with no punctuation, except that if BASE is followed by an alphabetic (rather than numeric) designation, it is separated by a hyphen. I believe there are other minor errors in the article, I will come back to this when I have time. -roy —Preceding unsigned comment added by 62.253.128.7 (talk)
Ok I think I've done this right, I've fixed up the ones I've found, though there may well be others. As to whether we really need a separate entry for every flavour of ethernet, I'll leave that up to someone else... -roy —Preceding unsigned comment added by Roybadami (talk • contribs)
Also note that there is no hyphen between 10G and BASE in the 10Gbps standards, eg 10GBASE-SR -roy —Preceding unsigned comment added by Roybadami (talk • contribs)
Frame format
The frame format stuff needs checking, as it is severely confusing even with references open in front of me! —Preceding unsigned comment added by 213.253.40.212 (talk)
So much more to write
There is so much more to write about here: AUIs, heartbeats, spanning tree, VLANs, G Ethernet and its relationship to Fiber Channel, 10G Ethernet and its relationship to SONET... —Preceding unsigned comment added by 217.158.106.250 (talk)
I now added a "See also" section and listed some of the terms. Gigabit Ethernet and 10 Gigabit Ethernet have their own pages, so Fiber Channel and SONE should be mentioned there (SONET already is, very shortly). Colin Marquardt
100 megabit thicknet
I have seen 100base5 (100mbit thicknet) beeing refered to in the manual of my IBM rs/6000 unix machine, and when searching around the web it seems to really exist. Does anyone know more which they can add in the article? —Preceding unsigned comment added by 213.238.230.200 (talk)
Pronunciation dispute
According to most.I have reviewed, the standard pronunciation is with accent on the initial long "E". There does appear to be a conventionally used alternate pronunciation with the accent on the "eth" with a short "e". However this only applies to the generic use of the word, since the trademarked name uses the former pronunciation.
Since this is a somewhat disputed point, the least-contoversial approach is just to leave this out, since pronunciation is not typically included in Wikipedia articles (this is better handled in Wiktionary). If the point of dispute is important enough to include, it should at least address the commonality of both forms of pronunciation, support any assertions, and use IPA pronunctiaton markup as described in article: http://en.wiki.x.io/wiki/Wikipedia:Manual_of_Style_%28pronunciation%29 —Preceding unsigned comment added by Limbo socrates (talk• contribs)
Realtime Ethernet Issues
Does it make sense to include a chapter about Realtime Ethernet here? —Preceding unsigned comment added by 62.180.123.245 (talk)
Re-arrangement would increase readability
I suggest that if one has some spare seconds, moving the general information like history on the top would be a good idea. That way, when someone who isnt technically oriented ends up on the article, he or she can read the history section before moving along. —Preceding unsigned comment added by Wk muriithi (talk • contribs)
I reverted the addition of a section on LXI by Bctwriter (talk·contribs); I think it is does not fit the flow of the article and not significant. The section removed is below for reference and discussion.
Speed nit
As one of the original users, I recall that the patented version actually ran at something like 2.997 Mbit/second, slightly less than three. This was an exact multiple of the US NTSC video refresh rate of 29.97 frames/sec (not 30 as people often round off to). The hardware folks used standard off-the-shelf parts for video technology that rate at this rate. But have yet to put my finger on the exact number. We should at least say "less than 3". W Nowicki (talk) 23:59, 6 April 2011 (UTC)
Found it. According to a 1982 paper, the speed was originally 2.94 Mbit/second. Not sure why, but it did jog the memory of Metcalfe pointing out once that the difference between this and 3.00 Mbit/second was more than the 56 Kbps link speed of the ARPANET at the time. Yogen Dalal has the paper on his web site. Which we should put on here I think, since he was the Xerox person involved in the commercialization of the 10 Mbit version for the Xerox Star products etc. W Nowicki (talk) 22:25, 7 April 2011 (UTC)
Nice work on your recent revisions. I did not know that was possible to put a reference inside a note. --Kvng (talk) 16:38, 10 April 2011 (UTC)
I ran into it in a previous article. Now that you mention it, Ethernet physical layer needs this detail even more than this page. And that includes a wikilink to Xerox Ethernet which redirects here. It is also a misnomer since we did have it at Stanford, MIT, CMU, and Berkeley at least. Proper term I think it "Experimental Ethernet". And there were other PHY differences, like the coax impedence, as given in that 1982 paper. Thanks for pointing this out. 22:33, 10 April 2011 (UTC)
Lead improvement
I have reverted user:W Nowicki's changes to the lead. There is definitely room for improvement but I don't see the new version as an improvement. I beleive we can make better progress by going back to the previous version and improving from there. Please have a look at WP:LEAD for guidance. Specifically: --Kvng (talk) 16:35, 10 April 2011 (UTC)
The lead for a substantial topic like this should be more than one paragraph.
It is generally not necessary to to include citations in the lead. The lead should summarize the body and the citations should appear in the body.
It looks more to me like you reverted user:Gnuish, not mine. I did not change the lead, but just added the detail that the experimental version ran at 2.94 according to the verifiable source. Also changed the "jugle" word that seemed like unciteded opinion. My changes seem still there.I agree that the lead seemed fine. My only comment is that saying it was used after 1980 is a bit misleading since it had hooked up Altos by 1973, and was published in 1975. The 48-bit address version dates from then, although this might be a nit. I would say something like "commercially available" instead of "used", since the experimental version was only in research projects. And probably drop "to the present" since Wikipedia generally avoids dated statements. W Nowicki (talk) 22:20, 10 April 2011 (UTC)
Re the process, you're so right, Kvng, I should stop wasting whole weekend afternoons improving Wikipedia articles when somebody whose taste differs from mine will just revert them. I have to relearn this lesson every few years. Thank you in particular for agreeing that the lead needs improvement -- then failing to even try to improve it. I included references in the text I added because I discovered that when I didn't put references into my contributions, too many people came along and removed my edits, claiming that they needed more references. It's valuable to me to know that you reverted them because you thought they needed fewer references. That's a good lesson for those expecting consistency from the Wikipedia editing community. BTW, the lead you reverted to contains a reference. Gnuish (talk) 22:07, 11 April 2011 (UTC)
Re the lead itself, I think it's far more important to mention why Ethernet is important (billions of devices, most Internet accesses occur over Ethernet, etc) than to use jargon like "frame-based" or mention the hoary OSI model in the lead. I suppose I could have inserted those facts without references into the lead, then added another paragraph somewhere else that duplicated the information and included the references, but I didn't feel that duplication was good editing practice. I really don't think the second sentence in the article should be about "wiring and signalling standards and media access control layers"; it should be about something that a reader who has noticed Ethernet jacks in their own home and wondered, "What the hell is Ethernet?" can understand. I made it a single paragraph because the contents fit into a single paragraph; I don't think there's a need to arbitrarily make multiple paragraphs just to make "the lead for a substantial topic" longer. I'm going to shut up now and stop wasting even more time on Wikipedia. Gnuish (talk) 22:07, 11 April 2011 (UTC)
I apologize for confusing user:W Nowicki and user:Gnuish. It is unfair to assume that I did not try and improve on talk:Gnuish's edits. I spent at least 30 minutes reviewing and attempting to improve. I compared old and new text side-by-side and found the new deficient and tried to explain why. I did not wholesale revert all of the work. I appreciate the WP:BOLD contribution and I'm sorry you're put off by my intervention. I've put quite a bit more than an afternoon into the article over the past few moths but I recognize that I do not WP:OWN it. All edits are subject to review by the community. There are many editors watching this article. Your work is still in the edit history and can be restored by anyone who believes that the edits improved the article or who can use them as a better launching point than the current text for further improvement. --Kvng (talk) 23:59, 11 April 2011 (UTC)
Changes to Evolution section
I have respectfully reverted user:W Nowicki's changes to the Evolution section. I'm not sure where you're going with this. I don't immediately see an improvement. Please discuss. --Kvng (talk) 16:56, 10 April 2011 (UTC)
Again, not mine. I would agree a few copyedits might be called for in this section, especially when just summarizing other articles. But removal of content might be better after consensus. W Nowicki (talk) 22:25, 10 April 2011 (UTC)
If you're Bill Nowicki, nice to see you here. Where I was going was: to more succinctly show how as Ethernet evolved, not only did it increase in speed, but also removed key barriers to acceptance, and expanded its area of application (e.g. from LANs into WANs). I tried to remove jargon and detail from the lead graf of the section, moved from the general to the specific in later grafs, and in the specifics (much of which was moved from History) showing WHY things changed rather than just what changed. I also got editorially tired of the passive voice and "With the advent of..." at every generation. Gnuish (talk) 22:07, 11 April 2011 (UTC)
I'm all for removing jargon and passive voice but that's not what I saw when I reviewed the edits. I've had another look and see that your new contributions are written better than the adjacent existing text. It looked to me like you added echoes what appears later in the Evolution section. I don't clearly see how this helps. You've deleted whole paragraphs, not just jargon. --Kvng (talk) 00:19, 12 April 2011 (UTC)
I tried to do some of this, mostly surgical copyediting so far. There indeed is too much jargon. I moved out the mention of FDDI and ARCNET from the lead, for example, and moved the citation to "external links" since it is a wiki anyway. That might be a compromise. There still is duplication between history and standardization, for example, and again in the evolution section. W Nowicki (talk) 03:29, 12 April 2011 (UTC)
OK, I replaced the first discussion of standardization in the history section with just a link to the standardization section, and then tried to spell out the alphabet soup in that section a bit. It still needs sources, and mention ECMA if they had a role? still not sure about the history vs. evolutiotion & other sections.
Review
I was asked to do a review for the computing project - this article is C class but with more sources and maybe some trimming and some "dumbing down" it'd be B class or even a GA. -- Eraserhead1 <talk> 21:51, 13 September 2010 (UTC)
This article is very difficult to follow. I was lost in the first two sentences. Could someone please provide a simple translation, at the front of the article, suitable for non-computer geeks. ProfKevinT (talk) 17:13, 26 July 2011 (UTC)
I've had a go at rewriting the intro to explain what I think are the most important thigns about ethernet without getting too technical. There is probablly still room for improvement though. Plugwash (talk) 16:08, 29 July 2011 (UTC)
To be honest that revision strikes me as more of a step backwards than an improvement. With these kinds of copyedits it is always very easy for errors to creep in even when making changes that do not appear to be significant.
In general stylistic terms, it is littered with typos and grammatical errors, to the extent that on occasion I'm not even clear as to what the original intended meaning was. There are also a few Easter egg links that give no clue as to where the link target goes: eliminating sector-specific terminology is a desirable goal but not as the expense of all other considerations, and especially not those of clarity or correctness.
Secondly, it contains a number of significant factual inaccuracies: one example that stick in my mind particularly is that twisted pair has no separate send and receive pairs at gigabit and faster.
Finally, it is far too long. A lot of the problems would disappear if the text that does not belong there is trimmed out. The lead is intended to briefly summarise the rest of the article, not replicate it in different words. No should it make points that are not built on later in the article, other than for standard things such as alternative terms and pronunciation.
These are issues significant enough that I'm half tempted simply to revert, but I'm not going to do that now. If it is going to stay in the long term, though, it needs a whole lot more work. Crispmuncher (talk) 17:02, 29 July 2011 (UTC).
Yes, I was just about to give similar comments. There are several typos introduced, and it is a bit redundant for my taste too. I woud say, for example, remove the complications of explaining CSMA/CD since that is only of historical note now. I can try reworking to a compromise; might be better than a revert. W Nowicki (talk) 17:25, 29 July 2011 (UTC)
Capitalisation
Why is Ethernet capitalised? It appears to be neither an abbreviation nor a proper name. This deviation from standard English usage should get at least a passing reference in the introduction or History section. Zipperdeedoodah (talk) 11:35, 12 October 2010 (UTC)
It is a proper name akin to a brand name. I don't think this needs to be mentioned in the article, but if you feel strongly about it, make the edit and we'll see how it goes. --Kvng (talk) 13:11, 12 October 2010 (UTC)
Well I think it might be compared to a trademark, in the sense that it really should be used as an adjective and not a noun. That is, "the Ethernet local area network" or "the Ethernet network interface" not "the Ethernet". But since no one company defended it as a trademark, sources now generally use it as a noun, but still capitalize it. So I vote to keep the way it is, Ethernet.
W Nowicki (talk) 23:59, 6 April 2011 (UTC)
Keep caps; it's a proper noun. At some point there may be sufficient history to support changing to lowercase/common noun status but I--personally--don't think we're there yet. — UncleBubba( T@C )02:14, 7 April 2011 (UTC)
Looks like we were over-ruled and it was changed back to lower case in many cases? I suppose this might be more gramatically correct so will not revert, but did not see that as a consensus. W Nowicki (talk) 20:36, 4 May 2011 (UTC)
I started to lower them all, but saw it was throughout the article, and decided to wait, then saw this section. Not sure what to do. I would think that 'ethernet' is a generic term, not a proper noun. If it is a "proper noun", I surely would like to see some evidence, but won't buck the consensus. All uses I have seen outside of here show it as not capitalized. Dennis Brown (talk) 00:59, 5 May 2011 (UTC)
Please have a quick look at Internet capitalization conventions for background on a similar issue. Your assertion that all uses outside of here are lower case is quickly disproved with a quick search. Please revert your changes. The article should be consistent. --Kvng (talk) 16:52, 5 May 2011 (UTC)
Arrogance aside, I did read it, and afterwards, I didn't see where it mandated that "E/ethernet" should be capitalized. The "Internet" may need to be capitalized as a proper noun, perhaps, but ethernet is a group of standards, not a trademark that I am aware of. At least the article doesn't claim it is, or give a reason why it is a proper noun. If you want to revert, no hard feelings, and I won't change anything else on the page, but I can't help but to think that that the name is generic and general and caps aren't required. Dennis Brown (talk) 19:45, 5 May 2011 (UTC)
I apologize if I came off as arrogant. Internet capitalization conventions does not talk about Ethernet. The part I found interesting was the discussion of proper nouns creeping towards common nouns over time fueled by widespread general use. I hope we can agree that Ethernet started off as a proper noun. Ethernet is more obscure than the Internet. If the Internet is not a common noun yet, then Ethernet certainly isn't either. Several editors including myself have chimed with support for a proper noun. You seem unsure about it. If we need to discuss it further, we will. In the meantime we need to put the article back to a self-consistent state. I have reverted your changes. --Kvng (talk) 21:38, 5 May 2011 (UTC)
I had not noticed the ongoing conversation until well after I had made the edits, and I noticed that many were capitalized after making a few changes and figured there must be an ongoing conversation, but I was short on time and didn't follow up til the next day, although I did stop decapping. Again, I won't buck the consensus and at the very least would agree there are two equally valid sides of the debate. I would lean toward the non-cap side. This might be due to having spent the least couple of decades working on networks and seeing it used without caps (granted, in more informal settings.) I also did a uspto.gov search and found that the single term "ethernet" isn't listed as a registered trademark, dead or alive. Lots of marks have the word "Ethernet" in them, but the single term doesn't appear to have ever been registered. Of course, that doesn't mean it wasn't ever a brand name, but it (to me) further indicates that the term has fully entered the public domain for all intent and purposes. I don't know any company you have to pay to use the term on your product, which further bolsters my belief that it is a generic term. I also notice many tech terms are in all caps, but some are not. ie: "local area networks" is not first letter capitalized, yet "Medium Access Control" is. Honestly, I would love to see (or participate) in a larger discussion about when to cap and when not to when it comes to tech terms, to at least understand what the "rules" really are about them. So again, I won't buck the consensus and certainly have no intention of re-editing them, but it would appear to me that capitalization isn't needed. Dennis Brown (talk) 23:57, 5 May 2011 (UTC)
There is plenty of evidence: the 1976 paper calls it "Ethernet" in caps as a proper noun, and the 1980 and 1982 specs say "The Ethernet" even more clearly a proper noun (although odd that "The" was not in the '76 paper?). Interesting that early journal papers call it generally "Ethernet local network" or "Ethernet network" as an ajective but it seems clear it was only under patent protection, not a trademark (to encourage other vendors, probably worth a mention). True that it is now a large family of related protocols, and standards bodies love to use unpronouncible acronyms and numbers (supposedly to put all nations into equal confusion). But vendors almost always capitalize. When in doubt, going with cited sources might be a good strategy; they cap. W Nowicki (talk) 19:34, 6 May 2011 (UTC)
The 802.3 standards consistently capitialise it too, and I would argue that that is authoritative. Personally I don't see why it should be capitialised (it is not a brand name) but that would be a very "I don't like it"-ish argument to maintain. Crispmuncher (talk) 20:15, 6 May 2011 (UTC).
Time for a navbox?
Most of the articles on Ethernet have a somewhat random collection of "See also" wikilinks. Perhaps it is time to create a navigation template? For example, there is one for {{IEEE standards}} the question is where to stop. Some obvious ones:
My basic impression is that there are way too many small articles while some topics are covered in various bits and pieces.
For example, one idea would be to consolidate several of the historic-only variants into an article 10 Megabit Ethernet or similar title (but keep the mainline variants separate), and rename Fast Ethernet to be 100 Megabit Ethernet. That might be more consistent and make navigation easier. W Nowicki (talk) 22:12, 8 June 2011 (UTC)
Intersting idea. But it might be huge. Is it possible to develop several templates instead? For example one for Ethernet, or perhaps wired LAN:s, one for 802.11 or wlan:s, etc?Mange01 (talk) 23:51, 8 June 2011 (UTC)
Of course, the fundamental principle of engineering: "with enough time and resources, anything can be done". The above list was meant for an "Ethernet" template, while both "LAN" "Wired LAN" or "Wireless LAN" could have their own. There already is {{802.11 network standards}} and {{Comparison of mobile Internet standards}} although they are tables that usually go in the body, not a navbox, but serve the same purpose. {{UTP Cable Standards}} already exists for cables, probably need one or more for fiber-optic articles. {{Cellular network standards}} includes WiMax but not WiFi.
But the more I think about it, the more merges make sense. For example, all the gigabit PHYs are described in the Gigabit Ethernet article, and all the ten gigabit PHYs in 10 Gigabit Ethernet. Several of these are actually still used. But obscure PHYs that nobody ever used have stand-alone articles with no or few sources? They might not survive deletion requests without demonstrated notability, so to preserve them merges might be the best option. Also several articles link to Ethernet when they really mean "the original 10 Megabit Ethernet". W Nowicki (talk) 17:11, 9 June 2011 (UTC)
I don't know how you know when it is time for a navbox. I'm personally satisfied with Category:Ethernet which I reviewed within the last year. I'm not opposed to a navbox. I'm not sure exactly what merges you are referring to. Be WP:BOLD and merge or put up merge banners where discussion would be helpful. It is true that while there are 100M, 1G and so on, there is no 10M Ethernet article. I couldn't find a way to fix that on my first pass through these articles mainly because one of the editors was set on preserving Ethernet over twisted pair. --Kvng (talk) 04:24, 10 June 2011 (UTC)
I wasn't the only one who felt Ethernet over twisted pair should have its own article. We do have a redirect from 10BASE-T, and there would be nothing wrong with including it in an eventual navbox. WP:CLN also makes it clear that categories and navboxes are not redundant with each other. --Tothwolf (talk) 07:51, 10 June 2011 (UTC)
WiFi wouldn't be included in {{Cellular network standards}} because it isn't a cellular network standard. The best starting point I've found for navboxes are categories. Once a group of articles has been categorised, those categories can be used to create subgroups for a navigational template. If the template becomes very large, it can be coded to have certain sections autocollapsed or uncollapsed and parameters can be passed to it as well. --Tothwolf (talk) 07:59, 10 June 2011 (UTC)
As for "when" to start a template, it is taste a bit but generally: when the "See also" entries at the end of articles tends to be long, but still does not include articles that are closely related. I went ahead and started one at {{Ethernet}} that has many of the major ones, but other ones will probably need to get added as it evolves. But do not plan on adding everything in the category, just ones that often appear in "see also" lists.
Agree Ethernet over twisted pair is notable enough to deserve its own article. I would merge 10/100 and 10/100/1000 in there too - they are already mentioned in the lead, just change the link to embold. The 10BASET-T in its lead is already embolded, but there is a wikilink later on in the article that might be confusing (going back to itself). I will add more merge proposals as the template work progresses. W Nowicki (talk) 20:26, 10 June 2011 (UTC)
Addresses and routing
Here's my chance to learn something. Don't routers and switches work on the *ethernet* MAC addresses? Not the higher-level (IP) addresses that might be carried in a data frame, but the lowest-level MAC addresses? If you didn't have these addresses in the Ethernet frames, a router would have to understand the protocol and format of a data frame to figure out what to do. It'd be like the Post Office steaming open each letter to read the heading on the first page, instead of the address on the envelope. --Wtshymanski (talk) 19:25, 29 July 2011 (UTC)
As far as ethernet is concerned a router is just another node on the network. When a host sends a packet to a destination to a node not on the network it sends it to the router - the destination MAC address is set to that of the router to ensure the router gets it. The IP packet within contains the ultimate destination IP address, not that of the router (ignore NAT for the minute) which is what the router uses to make its forwarding decision. However, the complete IP packet, including IP header, is simply payload from ethernet's perspective. All ethernet sees is a packet from the host (with the sender MAC to that of the host) to the router (with the destination MAC set to the router). What the router then does with that packet ethernet neither knows nor cares about. You're correct that the router then needs to understand what is in the payload of the ethernet frame, but only as far as the network layer (IP) - it needs to understand the IP header but all other contents of the packet are an irrelevance. As such the router still needs no knowledge of TCP, yet alone HTTP, email or anything else may be contained in the IP packet for basic packet forwarding.
Switching is different. That does work on the MAC addresses but only has any meaning on the local network. The primary motivation is to keep links free of traffic that known not to be needed on a certain link, by keeping track of where source MAC addresses originate from: if the switch knows a particular MAC is found out of a particular port it doesn't need to send that packet out all of the other ports. However, this is not routing, if you connect two networks up with a switch the result is a single larger network, not two networks connected by the switch. The distinction is made because of the number of nodes that need to be considered: most switches can keep track of a few thousand MAC addresses and want port they are one. The whole system would fall apart if the switch had to consider the MAC of every machine on the Internet to say nothing of the problem if a remote node is not on an ethernet network and therefore has no MAC. Crispmuncher (talk) 20:29, 29 July 2011 (UTC).
It's just occurred to me perhaps a picture is helpful here. An application layer program gives its payload to the transport layer (UDP here, it could equally be TCP) which adds its own header (source and destination port numbers, amongst other things) to the data. It is then passed to the network layer which adds an IP header (IP addresses). Finally the whole assembly is put in an ethernet frame (MAC addresses and checksum). The router strips off the ethernet frame before processing and looks at the IP header to determine where to send it. It will then re-encapsulate it in another data link layer packet for the outgoing interface. That could be another ethernet frame, or equally a PPP frame, or whatever else is used on that particular link. Crispmuncher (talk) 20:50, 29 July 2011 (UTC)
Now we're getting somewhere. Our articles on routers and switches would confuse someone from Cisco.
If I understand this, routers use IP addresses (or whatever other protocol is being carried on Ethernet frames). A router must "speak Bocce" and understand what's inside an Ethernet frame. A (n Ethernet) switch, on the other hand, is purely an Ethernet creature and only knows that MAC address N is physically attached to port X and MAC address M is on port Y, and sends a frame between the two ports when it detects those two MACs are trying to speak to each other (keeping this chatter off the segment attached to Port Z). If this is a correct restatement, then could we also add it to the swtich and router articles, adn as much of the switch part as is needed here? (Preferably with some block diagrams.) The MAC address is important for subdividing a network into physical segments, not for interconnecting networks. --Wtshymanski (talk) 21:09, 29 July 2011 (UTC)
A router is a kind of (packet) switch that operates at layer 3. A bridge is a switch that operates at layer 2. There are occasionally switch-like devices at higher layers, but they are rarely called such these days. (For example, some people used to use the word "switch" to describe software that gateways email between several different protocols.) The original "switch" in this sense was of course the telephone switch, which actually was a switch in the electrical sense. Some years ago, a number of vendors started to make marketing claims about their products being "true" swiches, or similar terminology, as a way of sowing confusion; each vendor tried to claim that their particular technology was a "real switch" and the other vendors' products weren't. But properly speaking, any device that examines the contents of a packet to decide where to send it next is functioning as a switch. 121a0012 (talk) 02:21, 30 July 2011 (UTC)
Actually our articles on computer networking probably confuse many readers, those from Cisco and many others. :-) In general, lead sections of high-importance articles might not be good places to engage in speculation. They should just sumarize the body and give some context to general readers. I think one thing missing is how Ethernet was a good fit to IP, which helped lead to both of their sucesses. Metcalfe and Boggs were Stanford students (PARC is on Stanford land) and Vint Cerf was on the faculty for example. Ethernet was designed to handle multiple protocols from the start (since PARC ran several by then) but was well tuned to IP, which meant the routing and switching advancements applied to both. I just realized the layering stuff also got dropped out of the lead. All this needs to be in the body first with citations of course, and then we can summarise in the lead. W Nowicki (talk) 22:57, 31 July 2011 (UTC)
Metcalfe was at Harvard (seconded to MIT Project MAC) when he developed what would become Ethernet when he brought it to PARC. IP was still five years or so in the future. (Metcalfe in fact built an ARPAnet host interface for the PDP-10 while at Project MAC.) 121a0012 (talk) 01:23, 29 August 2011 (UTC)
Half duplex compatibility
A recent edit introduced "...acted as half duplex links for compatibility and to allow for the use of hubs". That's not quite right because we have to assume that a network is designed for more than two hosts, and there is no Ethernet procedure to have full duplex between more than two hosts (a switch is required). Also, two typos need to be fixed: it's "separate" and "receive". Johnuniq (talk) 04:19, 30 July 2011 (UTC)
The only way a network can be composed of more than two hosts with 10base-T is to use a hub (which doesn't allow for full duplex) or a switch (which does allow for full duplex). Plugwash (talk) 00:59, 1 August 2011 (UTC)
I guess the wording is ok, but my concern was that this is an article on Ethernet and there is simply no possible Ethernet method of having full duplex other than between only two hosts, or using a switch which does various tricks to avoid the limitations of the shared medium. Johnuniq (talk) 01:20, 1 August 2011 (UTC)
Sorry it wasn't ideal wording but I was trying to fix blatently wrong info without going into a major restructuring job. The article previously claimed that full duplex switched 10BASE-T networks were common when afaict they were virtually nonexistent (afaict autonegotiation which is important for practical full duplex deployment wasn't introduced until the introduction of fast ethernet) without adding too much bloat/repitition to the article.
I've tried a slightly different wording now but IMO the REAL fix IMO is to merge the "history" section with the "evoloution" section and possiblly the "standardisation" section to provide a single unified progression on how ethernet developed but that is more work than i'm personnally prepared to do on the article. Plugwash (talk) 15:35, 19 August 2011 (UTC)
Actually I would be in favor of a more unified narative in roughly chronological order. That is, this would become more of a "history of Ethernet" article, with summaries of the specific steps along the way. Might take some time to get around to it of course. W Nowicki (talk) 21:01, 27 August 2011 (UTC)
To my eye, the only problem was that there were three paragraphs covering Evolution at the end of the History section. I have moved these. The History section is incomplete in that it does not cover much beyond inception. We need to add some details about the growth of Ethernet. I have added a reminder about that to the article. --Kvng (talk) 21:30, 29 August 2011 (UTC)
Apart from technical considerations, "Since its commercial release, Ethernet has retained a good degree of compatibility." (Introduction) is not adequate for me. Compatibility with what? Hardware, legacy networks, protocols, girlfriends... 24.27.31.170 (talk) 14:16, 30 September 2011 (UTC) Eric
Done History section now extends to present. --Kvng (talk) 20:17, 30 September 2011 (UTC)