Talk:RS-232

From Wikipedia, the free encyclopedia
Jump to: navigation, search
WikiProject Computing / Hardware (Rated Start-class, Mid-importance)
WikiProject icon 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.
Start-Class article Start  This article has been rated as Start-Class on the project's quality scale.
 Mid  This article has been rated as Mid-importance on the project's importance scale.
Taskforce icon
This article is supported by Computer hardware task force (marked as Mid-importance).
 

Pinouts[edit]

The DE 9 pinout still isn't in the standard. It shouldn't be in this article, which is about the EIA/TIA standard. The article has lots of pointers to the IBM AT pinout. --Wtshymanski (talk) 15:11, 12 March 2014 (UTC)

The archived discussion is here. Consensus was to not have the DE-9 info in this article. DE-9 serial ports are certainly related to RS-232, accordingly, the DE-9 info is in the Serial port article, which is linked from here by an {{about}} at the top of the page. Jeh (talk) 15:44, 13 March 2014 (UTC)
Two people does not make consensus. Since there are obviously at least two people that disagree with your assesment, (https://en.wikipedia.org/w/index.php?title=RS-232&oldid=599295052), That said, I was under the impression that the one with the DB-9 connector was the one that had been up for a while. It is not, so I was mistaken to restore it untill the discussion had been had.
I beleve that having the DB-9 info on this page is valuble as it is a defacto standard and it is not missleading to have it there. CombatWombat42 (talk) 15:56, 13 March 2014 (UTC)
What you would find valuable is an WP:ILIKEIT argument.
de facto standards are one thing. The Serial port article covers the ones that are relevant, including DE-9, in great detail. It's linked from here. This article covers the actual RS-232 standard. You can't possibly verify that the DE-9 connector is described in the RS-232 standard.
Regarding consensus, consensus on WP is not reached by voting, and a discussion does not have to have some minimum number of people in it for the result to be considered "consensus." Consensus is reached when valid arguments based on WP policies (like verifiability) are presented, and the other side stops arguing and the article edits on that point stop... which is what happened before. Ergo, consensus was established. You are not presenting any arguments here that were not presented before. Jeh (talk) 16:33, 13 March 2014 (UTC)
Consensus does require people agreing on your intrprtation of wikipdeia policy, which they clearly do not. you want a wikipdia policy to support my point: WP:IAR, will wikipeida be better with the DB-9 connection listed, YES.
Jeh, you could stand to be less arrogant, it would make me more likely to listen to you. CombatWombat42 (talk)
The DE 9 serial port pinout may be useful and valuable information, but it's not part of the RS 232 standard. Articles should be about their topics. Any associated or peripheral facts are only a mouse click away. --Wtshymanski (talk) 17:45, 13 March 2014 (UTC)
DE-9 is already mentioned in the connectors section, with a second WL to the Serial port article. Why is this not sufficient? Jeh (talk) 18:00, 13 March 2014 (UTC)
As for WP:IAR, please read the essays linked from there. IAR is not there so you can ignore consensus and other WP policies and get your way despite everything else. Consider that even if you invoke IAR, nobody else here is obligated to do so. Jeh (talk) 18:22, 13 March 2014 (UTC)
What a bloody stupid bunch of reasoning. You two aren't worth the typing to argue with. Andy Dingley (talk) 18:35, 13 March 2014 (UTC)
Andy, an article title like RS-232 is extremely specific. It's the name (or the WP:COMMONNAME of a standard. The article lede says the article is about the standard. I see nothing wrong with constraining its content to be equally specific. (By my opinion it should really be called TIA-232-F, but WP:COMMONNAME overrules. Rather than invoke IAR, I simply make sure that such pages exist as redirects to here.) It would be misleading to have a description of the DE-9 serial port pinout in an article about the RS-232 standard, as that information is not in the standard. But that information is not being excluded from Wikipedia! Indeed, we already have the article CombatWombat42 wants; it's called Serial port. Again I ask: Why is that not sufficient? At one point this article had essentially the same table (except for edits since then) of "every serial port pinout ever invented that anyone can find a reference for, and some others too" that the Serial port article does. Was that really an improvement? Jeh (talk) 19:35, 13 March 2014 (UTC)
I'm going to agree that the DE-9 information should not be in this article, — Preceding unsigned comment added by EE JRW (talkcontribs) 22:45, 14 March 2014 (UTC)
(Oops, pressed the save button instead of the preview above) I'm going to agree that the DE-9 information should not be in this article, RS-232 is for a D-sub, B sized shell 25-pin connector and an alternate 26-pin connector. Putting the 9-pin connector information in this article implies that the DE-9 is part of the standard. I have no problem with the DE-9 information being in the catch-all serial port article, but I think that the proper place for this information is the TIA-574 article, (which does not currently exist). That would document the information where it is actually defined, not in a place where it is not defined.
A couple of my personal peeves (and yes, I know that these make me an anal retentive personality) are calling the DE-9 connector a DB-9 connector. While you'll find references to the DB-9 all over, the DB-9 is actually a mythical creation. (some day, just for fun, I'm going to generate an image of what a DB-9 would really look like) No manufacturer makes such a screw balled contraption. The second peeve is calling the 9-pin serial port of the IBM PC-AT an "RS-232" port. The voltage levels are typically interchangeable and the connector can be adapted to work with an RS-232 port, but they are not the same thing. A third peeve is the assumption that a "RS-232" port has to use async start-stop ASCII (AKA a UART). The bit protocol is not part of the standard regardless of how many people think it is. A consensus is meaningless for this. Get a copy of the standard and tell me where it talks about the 9-pin connector or bit protocols and I will humbly apologize, but I've read the standard numerous times and have never seen either. Having information in the RS-232 article implying an incorrect assumption simply furthers the folklore surrounding RS-232. EE JRW (talk) 23:15, 14 March 2014 (UTC)
I do have a copy of the standard. I made the point about async character framing, bit ordering, etc., let alone details of character encoding like ASCII, not being in the standard, here. Re connectors, I think there is nothing wrong with mentioning that alternate connectors exist (indeed, the standard only "recommends" DB-25, it is not mandatory), and we already do that. But as that is the only connector whose pinout can be referenced to the standard, that is the only one whose pinout should be listed here. Jeh (talk) 23:23, 14 March 2014 (UTC)
I am not surprised that you have a copy of the standard since you're arguing to keep the article based on what's in the standard. While I don't chip in very often, I've been following several of these "RS" standards pages for a while. I realize that most of this has already been discussed.
I goofed and clicked the save button instead of the show preview button, then madly typed the rest of wanted to say instead of taking my time and making what I wanted to say clear. My apologies if I implied that no one else had made the point about the bit protocol. Actually my apologies to anyone who did not like what I wrote. I am not looking to start a fight, I was just adding my opinion. In hindsight, I should have reverted and taken my time re-typing my 2 cents worth, but I didn't and what's done is done. I also note that you are using the term DE-9 and not triggering one of my pet peeves. I also have the same problem as you with the title of RS-232 instead of TIA-232, but I've had that discussion on another "RS" talk page. I'm basically agreeing with you and Wtshymanski about the DE-9 connector. The main point I was trying to make is that the DE-9 connector is a standard (TIA-574, which I do not have). The connector should be documented under TIA-574. This would require someone (preferably one of those wanting to add the "DB-9" connector to the RS-232 article) to create a TIA-574 page. Then, they can add whatever they desired to that page instead of adding it to a page where it does not belong. For any one interested in creating this page, finding information about this standard from TIA can be difficult. Try going to http://www.tiaonline.org/standards/buy-tia-standards and searching for TIA-574. There is a description which would be a good starting point.
Finally, the connector. You say that standard only recommends the DB-25 connector. Again, I'm not looking for a fight, but I don't see this. Section 3.1 states "A 25-position connector is the normal connector specified for all interchange circuits. An alternative 26-position connector (ALT A) is specified for use when a smaller physical connector is required." Can you tell me where the standard says the DB-25 connector is only recommended? EE JRW (talk) 01:59, 15 March 2014 (UTC)
Re DE-9, I was one of the chief arguers who got these articles to stop using the "DB-9" term, with several cites to manufacturers' and suppliers' catalogs. (I think it was in the talk page for D-subminiature but it's not important enough to go check right now.)
Re "recommended", I confess I didn't look that up when I wrote that. Can't remember where I got it.
I doubt we really need an entire page just for TIA-574. That standard is referenced from Serial port, I think that is sufficient.
Would you mind using conventional paragraph formatting? br's create a wall-of-text effect. Your post here was very difficult to read because of that. Jeh (talk) 02:36, 15 March 2014 (UTC)
A page for TIA-574 may not be needed, but my opinion is that the standard for a "9-Position Non-Synchronous Interface Between Data Terminal Equipment and Data Circuit-Terminating Equipment Employing Serial Binary Data Interchange" is the proper place to document the 9-pin serial port connector. On the other hand, I'm not going to create another page until I purchase the standard and get enough free time to work on it. That's not likely to be soon.
Using your post as an example, I'm using multiple colons instead of br's. I don't see the difference with Firefox, but if that's the convention then that's what I'll do. I'm still learning this whole Wikipedia thing :). Thanks for the pointer. EE JRW (talk) 14:41, 15 March 2014 (UTC)

"Recommended Standard" 232[edit]

I've heard that the RS stands for Recommended Standard but didn't see anything about that in the article. If this is true, it'd be neat to provide a little background on why it's still only "recommended". — Preceding unsigned comment added by Niedzielski (talkcontribs) 05:23, 7 August 2015 (UTC)

The standard is not simply "Recommended". RS-232 was obsoleted many years ago, (by the EIA in 1987[1] - date and reference inside these parenthesis added by Gg-labz 20160229). The proper name for this article is TIA-232-F, but as is discussed above by Jeh, WP:COMMONNAME overrules. What it comes down to is that everyone calls it "RS-232" even though that title isn't correct. This has been discussed in other RS articles Talk:RS-485#RS_vs._EIA_vs._TIA and the agreement has been to keep the RS name as the title.
RS does stand for Recommended Standard, and the EIA link has this information. Other RS articles discuss this in the article. RS-485, RS-485 Technical Manual This RS-232 article does not.
So the question becomes, should this information be added to every RS article? That answer I don't have.
EE JRW (talk) 14:35, 8 August 2015 (UTC)
The evolution of the prefix is described here. Jeh (talk) 09:48, 7 March 2016 (UTC)
I think it's probably not correct. Please see also the discussion on the talk page of that article: Talk:Electronic Industries Alliance#What does "RS" stand for?. --Wosch21149 (talk) 21:54, 7 March 2016 (UTC)
RS does stand for Recommended Standard. I added some information (sorry no references) to the EIA talk page, as that seems to be a more appropriate place to discuss this. EE JRW (talk) 01:42, 29 March 2016 (UTC)
Do we have consensus here that RS standards for Recommended Standard? The article currently has "RS-232 was first introduced in 1962 by the Radio Sector of the EIA.[1][2]" The first source never mentions the words "Radio" nor "Sector". The words "Recommended" and "standard" show up but never in a useful context. The second source is a 404 and does not appear to be a WP:RS.
Unfortunately, Google Books search results are thin. A search for "Radio Sector" in documents published from 1900 through 1979 gets 18 results meaning it was easy to inspect all of then. There are zero mentions of that term in in a context that's relevant to us. Doing the same search for "Recommended Standard 'RS'" finds two results from 1979: EIA Recommended Standard RS-313-B, EIA Recommended Standard RS-313-B, and Reference Standard RS 17-2 (for City of New York building codes). I don't consider two hits for what look like the same set of words for something published in 1979 to be definitive.
Attempting "Recommended Standard" "EIA" adds to the two found above with: EIA Recommended Standard RS-349 (1968) and the same wording in another document.
That's it - zero mentions supporting "Radio Sector" and two for "Recommended Standard" when hunting through Google Books for 1900 through 1979.
I would have hoped that had either "Radio Sector" or "Recommended Standard" been what we are seeking that there would be hundreds or even thousands of hits. "RS232" and "RS 232" also get very few hits and so it's likely we are simply dealing with an era and type of material where very little has made it into Google Books.
We could try adding the 1980s or even 1990s but the explosion of people writing computer books started in the 1980s and most of those were poorly sourced meaning we'd be getting a lot of noise from unreliable sources. This forum thread includes a comment "that while the connector was introduced in 1962, the original EIA-232 Standard was published in 1960. It looks like in May of that year. Does that make sense? "EIA-232 - Complete Document Revision / Edition: 60 Chg: Date: 05/00/60 INTERFACE BETWEEN DATA TERMINAL EQUIPMENT & DATA"". That looks like something worth chasing back to a WP:RS. For example, "EIA 232" "May 1960" finds Standards for computer aided manufacturing - NIST Page which has on page 18: EIA RS-232-C, August 1969, title: Interface Between Data Terminal Equipment and Data Communication Equipment Employing Serial Binary Data Interchange, August 1969. Previous standards are: RS-232, May 1960. RS-232-A, October 1963. RS-232-B, October 1965. RS-232-C is expected to be gradually (ten years) replaced by EIA SP-1194A.
That's the sort of history we should have in this article. The downside of that NIST document is that they may have made an effort to standardize the wording. For example, they called all of the early standards "RS-232-..." but we don't know if the EIA used "RS-232-..." at the time they were published nor what "RS" stands for. --Marc Kupper|talk 06:35, 12 July 2016 (UTC)
The NIST document saying "recommended standard" is good enough for this completely minor claim. We regard NIST as a pretty damn RS here; "NIST may have made an effort to standardize the wording" seems to me to be carrying skepticism and doubt much too far. Jeh (talk) 17:21, 12 July 2016 (UTC)

Character frames and data rates[edit]

I'm a bit surprised that this article does not mention the start bit(s), character (or data) bits, parity bit, stop bit(s), and whatever may be on a line when nothing is being transmitted. It also does not mention the word "baud" at all nor the common data rates. Are these things not part of the RS-232 standard? If so, I suspect the article should say that those items are not part of RS-232 and to refer people to the articles that do cover those topics. The Serial port article has a list of common speeds though fails to go into the details of what a character frame looks like. --Marc Kupper|talk 04:18, 12 July 2016 (UTC)

No, none of that is part of RS-232. We discussed this on the talk page awhile back - it's in the archives. See for example the end of this thread: https://en.wikipedia.org/wiki/Talk:RS-232/Archive_2#RS232_level_question I can tell you that start and stop bits, etc., date back to 1916 or a bit earlier. Details are in Asynchronous serial communication. Jeh (talk) 04:38, 12 July 2016 (UTC)
Thank you for the quick response Jeh. I also just saw the "Scope of the standard" section of this article. --Marc Kupper|talk 04:57, 12 July 2016 (UTC)

dubious tag[edit]

for one, WIKI:NOR. Another thing - the statement "not at all dubious. A native serial port has a completely different host-controller interface than a USB-to-serial adapter; in fact the latter has no HCI of its own, only that of the USB HC)" should go here (on the talk page), not in the edit note. Opening a discussion in edit notes is a direct attempt to start an edit war IMO.

third thing - you're wrong. That statement is pure OR, built on weasels. There ain't any "native" RS232, regardless of the amount of handwaving you do, in the context of the phrase in question. Can DB25 port connected to MAX232 output driven by UART 16550 attached to a uC be described as a "native" device? What the hell is "native" anyway? Native as what? Most serial-to-USB adapters are just UART chips, with no regard to neither RS232 standard, signal lines, voltages, wires etc. That still doesn't prove anything; they ain't called "RS232 adapters", because they ain't. OTOH, there are devices built to comply with RS232 standard explicitly. Throwing them to one bucket ain't a good idea IMO.

Another thing - there ain't a USB-to-RS232 adapter. There are many USB-to-RS232 adapters out there. You're treating them as one - without any reference, and without any logic behind it. "Devices that convert between USB and RS-232 do not work with all software or on all personal computers." - the statement is as vague as possible; which devices (PL2303? FT232? uCs?), what is all software and all personal computers? That's an insane amount of weasel words for one statement IMO.

Poponuro (talk) 18:10, 19 September 2016 (UTC)

Well, it wasn't a "direct attempt to start an edit war" in my head, so your O is mistaken. I've participated in many edit series, and watched many many more, that were closed successfully because of explanations in edit notes. It is true that extended discussion in edit summaries is frowned on, but at the same time, a rv's edit summary is supposed to explain the reverter's rationale.
And my, you're being awfully bristly. Care to tone it down a bit?
No, my statement isn't "wrong". Probably it could have been worded better. I'm not talking about mistakes in the RS232 implementation (voltages, etc.). What I'm talking about are the programmatic differences between a PC legacy serial port (you know, the sort of thing that exposes eight I/O ports typically starting at 0x3F8 for the first one), vs just about anything else. Some PCI-to-serial adapters have, or default to, an HCI that's compatible with that... but they're increasingly rare. Some present a very different HCI (look at MOSchip MCS9901, for ex.) and these will not work with software that is specifically looking for a legacy serial port.
Nor will ANY USB to RS232 adapter. They cannot possibly present a similar, let alone identical, HCI as a legacy serial port. The only HCI that shows up for them is that of the USB host controller.
So: Legacy code that talks directly to a legacy serial port (bypassing OS APIs, drivers, etc.), expecting an HCI identical to that of a legacy serial port, will flatly not work with any USB to RS232 adapter, nor with some of the PCI adapters.
Now, you may be thinking, what apps talk directly to a legacy serial port any more? Heck, in the NT family that's not even permitted from user mode unless you're in a 16-bit app (where "virtual device drivers" intercept the exceptions that occur when you try to access the serial port and turn them into IOCTLs to the serial port driver). Don't most all apps these days use the OS's APIs and drivers? Yes, they do. And don't those APIs and drivers hide all these differences in the HCI?
Well, almost. They try. But there are a few fundamental differences in function. USB-to-serial adapters are built to the USB Communications Device Class spec. And that spec simply does not allow Data Set Ready from the RS232 connector to be communicated to the host computer. This might be ok as long as the USB-to-serial dongle is used only to talk to modems, but for some of the much more generalized uses of serial ports, where your app really does need to monitor DSR, you're kinda hosed...
Given these points, it is not necessary to present exhaustive lists of legacy-compatible and -incompatible USB to RS232 adapters. It is only necessary to understand a few details about how legacy serial ports vs. all the others work. Remember, a single example of incompatibility is all that is necessary to defeat the claim that "all are compatible".
And btw "some" (or "not all") is not a weasel word. (See WP:WEASEL. We don't use that term to mean what you think it means.) It is simply an indeterminate between "all" and "none".
No, I don't have a reference. Right now. But you can't say that there's no logic behind the claim. Jeh (talk) 14:24, 20 September 2016 (UTC)
Well, it wasn't a "direct attempt to start an edit war" in my head, so your O is mistaken.

Just because you didn't intend it to be, doesn't mean it is not. The very idea of edit wars begun with people bending (first soft, then hard) rules because "I am right, I intend no harm, my edits are good, so it's OK".

No, it isn't. The 1st content rule is: NOR, cite, ency. Another content rule is: clear, unambiguous, related. Rving a request for clarification and citation is going exactly the other way. Providing a faulty reason in edit summary only highlights it.

And my, you're being awfully bristly. Care to tone it down a bit?

You, OTOH, are being awfully ad personam. Care to back down a bit?

No, my statement isn't "wrong". Probably it could have been worded better.

If the wording makes it impossible to understand correctly - it's wrong, by the very definition of the word "wrong".

Now, going directly back to the matter at hand: 1. After you replaced the word "native" with "legacy", I can *finally* try to understand what you meant. 2. I can agree that no USB-based (and other types, as you outlined) COM port will provide legacy support for e.g. asm IN/OUT etc., because the hardware is simply not there. Chipset is not aware of it, BIOS is not aware of it etc. - still, 3. Using "some" in encyclopedic content is permissible iff there is an enumeration showing at least one example. Otherwise it's a weasel, by definition. You're a seasoned editor, and you should be well aware of it. 4. That being said, it's completely unrelated to the very dubious statement. I never said "Devices that convert between USB and RS-232 'do work with all software or on all personal computers." I only said, to reiterate: a) give an example of such devices, b) describe what kind of software/PCs are covered by this statement, c) give a source citation to back it up. It should be either rewritten, rephrased, sourced/backed with citations, or removed altogether. — Preceding unsigned comment added by Poponuro (talkcontribs) 22:29, 20 September 2016 (UTC)

As someone just passing through, I would like to say that I read that sentence and thought 'wow what a uselessly vague statement'. Then I clicked on the weird little links after it and found that you all know that it is uselessly vague and yet it's still there. No wonder people make fun of wikis. — Preceding unsigned comment added by 75.144.8.253 (talk) 16:27, 27 January 2017 (UTC)

  1. ^ "RS232 Tutorial on Data Interface and cables". ARC Electronics. 2010. Retrieved 28 July 2011. 
  2. ^ Metering Glossary Landis + Gyr Tutorial (see EIA)