Talk:Reliable byte stream

	This article is within the scope of WikiProject Computing, a collaborative effort to improve the coverage of computers, computing, and information technology on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.ComputingWikipedia:WikiProject ComputingTemplate:WikiProject ComputingComputing articles
Low	This article has been rated as Low-importance on the project's importance scale.
	This article is supported by Networking task force (assessed as Low-importance).

Deletion/merge proposal

User:Oli Filth suggested this article to be deleted at WP:TEL and in the template at the top of the article, because of the following concern:

Non-notable terminology; I don't think it exists outside the incidental combination of the adjective "reliable" with the noun "byte stream". No references to show otherwise.

I agree with the critisism, but instead of a deletion I suggest a merge with Byte stream, or perhaps Reliability (computer networking).

Please have in mind that the following articles are linking to this article.

User:Jnc
TCP/IP model
Reliable byte-stream (transport page)
CYCLADES
User talk:203.122.11.133
Reliable stream (transport page)
- Internetworking
- Xerox Network Services
PARC Universal Packet
EFTP
Retransmission (data networks)
Wikipedia:WikiProject Telecommunications
User:DumbBOT/ProdSummary

Mange01 16:33, 17 September 2007 (UTC)[reply]

Pardon me, but this term is very well known (and used) in the computer networking field. And it's not an occasional use of "reliable" along with the well-known term "byte stream" in text: try doing a Google search for "reliable byte stream" and you get like 20K hits - something a random concatentation of "reliable" with some other term (try, e.g. "reliable packet switch") doesn't produce. It gets used in slide titles (e.g. this page), and places like that. Etc, etc. I have restored the page, and will try and add some references. Looking for them now... Noel (talk) 02:34, 21 September 2007 (UTC)[reply]

Still trying to find the best reference (plus source of the term). Interestingly, the TCP spec (RFC-793) doesn't use the term, instead referring to an octet stream (since that was in the days when not all computers used 8-bit bytes; the then-common PDP-10 supported variable byte size, and six-bit bytes were used a lot on them). The earliest RFC to use the term "reliable byte stream" in its now-common meaning appears to be RFC-826, from November 1982. Sigh, I think it's one of those terms that, to all in the field, needed no explanation - as soon as you heard it, you knew what it meant. I can find plenty of papers that use the term, but none that define it, alas. But it is the most common term used to describe the service provided by TCP (e.g. W. Richard Stevens, TCP/IP Illustrated: Volume 1 - The Protocols; Alberto Leon-Garcia, Indra Widjaja, Communication Networks, etc, etc). Noel (talk) 05:33, 21 September 2007 (UTC)[reply]

I don't doubt that it's a widely-used phrase with a fairly well-understood meaning, but I'm still not sure it's any more notable than an incidental combination.

As an example, we all know what the phrase "fast internet connection" implies, and a Google search returns 501,000 hits. Does that mean that "fast internet connection" is a notable term in its own right? I don't think so; I think it's the incidental combination of "fast" with "internet connection".

If you're able to find some references which give an explicit definition for the term (even "a stream of bytes which is reliable"), then that's great, and I don't think I'd have a problem with the article staying. If not, I'll put it up for AfD at some point. Oli Filth^(talk) 08:17, 21 September 2007 (UTC)[reply]

Another point; most terms that engineers, etc. consider notable in their own right (as opposed to just an incidental combination) end up being shortened to an acronym; however I can't find anything relevant for "RBS". I know that this isn't a particularly formal metric for assessing notability, but I think informally, it's quite a strong hint that the term isn't particularly relevant in it's own right. Oli Filth^(talk) 08:26, 21 September 2007 (UTC)[reply]

My original reasoning for the sourcees tag was that it's only a term. Is there a theory or wider background for it? Remember, Wikipedia is not a dictionary. This is a problem for many many WP:TEL articles. I've been working on moving such articles to Category:Telecommunications terms, until I decide what to do to them.

If the article only can give a definition for a term, it should be moved to a dictionary or merged to somewhere else. I suggest merging to the article about TCP if that is where this term is used in this form, and explaining it there. Alinja 14:30, 21 September 2007 (UTC)[reply]

First, TCP isn't the only protocol that provides a reliable byte stream service paradigm, merely the most well-known. In addition to the articles on other protocols (some admittedly historic, but coverage of historical stuff is an important ancyclopaedic function) which use this term - as well as some that should but don't - a quick Google search reveals more things that are spoken of as providing a "reliable byte stream" service, e.g. Java, SOU, etc.

A reliable stream isn't the only service paradigm which reliable protocols provide; others (e.g. SCTP) provide a reliable message stream, i.e. the data stream is divided up into discrete objects, the boundaries of which are visible to the consumer of the data. (In fact, the article ought to mention that; I'll add it.)

It would also be a bad idea to add it to the TCP article because it would just bloat that to an even longer length.

Second, if this article was just a definition (i.e. what it means), it might be a candidate for a dictionary, although dictionaries are usually for words, not terms, and a lengthy description of a non-trivial technical concept (e.g. "mass-energy") isn't really grist for a dictionary. However, it also does more than that; it describes how most protocols which provide a reliable byte stream service paradigm do so.

Finally, I simply don't understand this need to get rid of a perfectly fine, albeit not lengthy, page. WP:PAPER, right? This isn't a vanity page about some obscure band, it's about a key technical paradigm of one of the most-used pieces of technology on the planet.

Yes, I wish I could find a written source which defines it (although my extensive collection of networking stuff isn't long on introductory books, which is where one would expect to find one), but it seems to have entered the lexicon without anyone bothering to do so, because to the technical community which created it, it's self-explanatory. However, Wikipedia is for everyone, and a term like "reliable byte stream", which might seem perfectly self-explanatory for a long-time computer networking geek, might not be so self-explanatory for Fred the waiter or Sally the gardener. Let's remember we are writing for others, not ourselves, yeah? Noel (talk) 05:04, 22 September 2007 (UTC)[reply]

I understand what you are saying. It's very common in computer and telecom field that there are no single definitions for terms. One term may mean slightly different things in different contexts and completely different words can refer to the same idea from different point of view (eor example Reliability (computer networking)). Also, most people learn terms from school or learn them from common usage - and everyone organizes the terminology differently in their minds - so there may not be a book source available at all. This makes some implications for a WP:TEL article.

The article seems to have evolved into the right direction already. I think we can justify an article about this concept if the scope is kept wide enough and the subject covered widely enough. This is a central theoretic concept in networking after all. It could become a good general audience article without too much technical details about specific implementations. It still needs a lot about the background what it purpose it servers, other approaches to the same goal etc, about terminology in different contexts, etc. Also retransmission is one implementation, FEC is another which should be at least referred to in the article. Just brainstormed some ideas.

Good that some kind of discussion emerged. I think there are way too many closely related short articles that need to be merged into "real" articles. This is some sort of problem because the concepts are usually relatively narrow and thus don't have the potential to become good, complete, well-presented articles. Alinja 07:19, 24 September 2007 (UTC)[reply]

Like Noel, I understand "reliable byte stream" as an important term in computing. In fact "reliable byte stream service" is the service provided by TCP, just as "datagram service" is the service provided by IP. For that reason alone, it deserves its own page. Which raises the disturbing point that "datagram service" doesn't appear in wikipedia. Hmmmm. However... I added two references that use the term in a prominent way (in fact, Peterson and Davie's text devotes an entire section to "Reliable Byte Stream (TCP)." On the other hand, as I was looking for more references, I found various authors using similar terms that clearly mean the same thing. For example, Comer talks about Reliable Stream Transport Service and doesn't define either that or reliable byte stream in his glossary; reliable byte stream doesn't appear in the glossary in Peterson and Davie; and although reliable byte stream service is discussed in quite a number of books, it only seems to appear in a small number of networking books (at least, of those digitized by google, so I may be looking at a rather unrepresentative sample - the Peterson and Davie edition, for example, is quite old). Ngriffeth (talk) 21:26, 10 May 2009 (UTC)[reply]