28
\$\begingroup\$

In the interest of having an explicit strategy on component identification questions, here is a "poll" question for community opinion on the matter. Historically, these questions have been allowed, but this is an existing strategy, not necessarily the current opinion of the community. This question is designed to be a simple binary question that will be used to shape a more nuanced policy later. I will include the pro and con reasons that I can think of, if you have another reason that I did not include, feel free to leave a comment.

TLDR: Should we allow component identification questions?

Please note, only upvotes will be counted.

\$\endgroup\$
  • 6
    \$\begingroup\$ To reiterate at the risk of being redundant, downvotes will not be counted. All they do is make it so I have to look at the vote totals instead. You don't get a "double vote". \$\endgroup\$ – W5VO Sep 4 '18 at 14:45
  • \$\begingroup\$ I was just about to post a comment about the reason I believed a component discussion is in order here, then you delete my whole post... Likely this whole thing was triggered about identify this blue plastic blob question, which is 1) highly uninteresting, 2) of no value to future readers and 3) probably not even answerable, as two answers point out. However, if someone asks to identify this circuit it would be fine and on-topic. \$\endgroup\$ – Lundin Sep 6 '18 at 14:28
  • \$\begingroup\$ @lundin his comment on your post said that a second post regarding what rules we the community want, will follow this vote. Not sure you have the rep to see deleted posts yet, or saw his comment in time. \$\endgroup\$ – Passerby Sep 6 '18 at 14:50
  • \$\begingroup\$ @Passerby Yes and that's completely backwards. Without defining what a "component" is, this whole post is pointless, as it could be anything from a resistor to a lawn mower. \$\endgroup\$ – Lundin Sep 6 '18 at 14:54
  • 2
    \$\begingroup\$ Please show one question where someone asks what a lawn mower is. \$\endgroup\$ – Passerby Sep 6 '18 at 14:57
  • \$\begingroup\$ Your focusing on extreme edge cases where no one who sees it will be confused about it being off topic. \$\endgroup\$ – Passerby Sep 6 '18 at 14:58
  • 3
    \$\begingroup\$ @Lundin You're describing the nature of the beast. Most of these questions will be "Identify this part with insufficient/vague markings". These questions not being interesting is your personal opinion. The value to future users is inherently limited. There is a good possibility that a component identification question isn't (practically) answerable. What does it mean if you have a valid answer on an "unanswerable" question? \$\endgroup\$ – W5VO Sep 6 '18 at 15:11
  • \$\begingroup\$ @W5VO A question about "identify this lawn mower" is uninteresting because it is off-topic, I came here for electrical engineering, not gardening. Now I'm sure that some engineer out there is a passionate gardener and loves to give input, but that doesn't make it any more on-topic. And so personal opinions don't matter, just where we draw the line of what's EE and what isn't. \$\endgroup\$ – Lundin Sep 6 '18 at 15:19
  • 3
    \$\begingroup\$ @Lundin Do you want to have a separate poll about lawn mowers? I think your example is too hyperbolic for useful discussion. I think the rough guidance will be connectors, components that are a single line on a BOM, no PCB assemblies. \$\endgroup\$ – W5VO Sep 6 '18 at 16:08
  • \$\begingroup\$ The only problem with part identifying questions is that there could be a possibility of duplicate questions. I don't have a problem with part identification questions themselves but it could lead into problems if people want to identify the same thing. Because StackExchange can't organize uncertainty, I feel like we would see duplicate questions but we would not able to find the duplicate questions easily. What's preventing someone from asking a part identification of the same thing that was asked, say, a few years ago? How would the user be able to find that question? \$\endgroup\$ – KingDuken Sep 7 '18 at 15:08
  • \$\begingroup\$ i.e. "Identify this connector type". There's probably hundreds of questions asking for connector type. Anyone who wants to find a particular connector type, for an example, has to dig through a lot of questions to find what they're looking for. Should we edit the question and say what connector type was found in the post's headline so that we don't receive duplicate questions? How can we organize these questions so that users can easily find what they're looking for? \$\endgroup\$ – KingDuken Sep 7 '18 at 15:18
  • \$\begingroup\$ @KingDuken Again, that's kinda the nature of the beast. It can be difficult to find duplicates on regular questions, let alone these identification questions. Right now I'm not suggesting that we fix that, or that it even is something that can be fixed. \$\endgroup\$ – W5VO Sep 7 '18 at 17:01
  • 1
    \$\begingroup\$ Apparently people didn't get the upvotes message. \$\endgroup\$ – laptop2d Sep 7 '18 at 19:14
  • 1
    \$\begingroup\$ @KingDuken Generally not done network wide to edit the question to reflect the answer. \$\endgroup\$ – Passerby Sep 7 '18 at 21:34
  • 1
    \$\begingroup\$ Hey, what's up with deleting my answer to this question? Maybe it wasn't the answer you wanted to hear, but it was my genuine answer to the question. 4 others agreed, and 1 disagreed. As far as I can tell, no rules were broken. What the...? \$\endgroup\$ – Olin Lathrop Sep 10 '18 at 12:47
44
\$\begingroup\$

Yes:

Component identification questions should be allowed (This does not excuse other close reasons).

  1. People answering these questions enjoy answering the questions.

  2. We've probably all at some point tried to scavenge parts and wondered what a particular one was.1

  3. Some community members enjoy answering these questions.

  4. The questions hitting the Hot Network Questions tab improves our site visibility.

  5. Reverse Engineering is a legitimate design activity. You have to know what the parts are before you can figure out how the circuit works.

(add comments below for additional reasons)

\$\endgroup\$
  • \$\begingroup\$ Folks, could you please post a quick comment about your reason for voting "yes". If the reason is already on the list, say which items. Or, add your own. \$\endgroup\$ – Nick Alexeev Sep 4 '18 at 6:16
  • 1
    \$\begingroup\$ I voted yes because of point 1, mainly (when I'm the one answering the question). Also because I often learn things when other people answer these questions. Also because some parts are super hard to identify, and only people who already seen similar things will be able to answer. Chat does not leverage enough participants to hope getting relevant answers in these cases. \$\endgroup\$ – dim Sep 4 '18 at 7:21
  • 9
    \$\begingroup\$ I'm reluctantly voting yes, mostly because of point 2, but I agree with most of the reasons for no. I want component identification questions, but I want some stricter rules: Must include where it is placed on the board or where it was found, must include all visible text in the question, and should include a short textual description of the item. Anything else should be closed as unclear. \$\endgroup\$ – pipe Sep 4 '18 at 7:34
  • 3
    \$\begingroup\$ I agree with pipe on clearer guidelines. I am voting yes because there are occasions to reverse engineer something and finding out what a part really is can be incredibly useful. I would add that if it is on a circuit board (as is often the case), I want to see at least an attempt at a schematic for context. \$\endgroup\$ – Peter Smith Sep 4 '18 at 11:23
  • 3
    \$\begingroup\$ I voted yes because I sometimes enjoy answering these questions (Point 1 and 3?). But mainly because I know how hard it sucks to be unable to find that damn connector. I'm in for stricter guidelines. Not sure how hard it is to get that based on the use of the identification tag. Some sites have a special notice on the site when a tag is used. \$\endgroup\$ – Arsenal Sep 4 '18 at 14:04
  • 1
    \$\begingroup\$ @pipe close voting immediately is a death sentence for those questions. Op would have no time to edit that in, and once the question is closed it's unlikely to get reopened. \$\endgroup\$ – Passerby Sep 4 '18 at 14:35
  • 3
    \$\begingroup\$ @Passerby I find that to be a site-wide flaw though, which is not specific to this type of questions. I think everyone who voted to close should get an automatic notification if the question has been edited. \$\endgroup\$ – pipe Sep 4 '18 at 14:37
  • 3
    \$\begingroup\$ To paraphrase Tim Post's community post in the last year or so: Stack exchange is about helping people, even if it's just ONE person, the OP. A question doesn't need to help others, and that's why too localized was removed as a close reason. Identification questions help people. Period. \$\endgroup\$ – Passerby Sep 4 '18 at 14:39
  • 3
    \$\begingroup\$ @Passerby As a recent example Raspberry Pi SE reconsidered a policy to accept product recommendations. Now they are explicitly off-topic. Not that I'm against helping people, but stricter rules sometimes help the community more than more lax rules would. \$\endgroup\$ – Dmitry Grigoryev Sep 4 '18 at 15:28
  • 1
    \$\begingroup\$ I vote yes, because of reasons 2 and 3.The problem I have with these questions is sometimes your 'passing the buck' to someone with more google-fu (because you just don't know how to search), which I don't feel is the intention of the site. The other problem is that users generally post a pic, but most often they need more information in the question, or to get out a meter and figure out how the part functions. This generates a lot of comments and useless discussion. I still vote yes, as long as they don't turn into repair questions. \$\endgroup\$ – laptop2d Sep 4 '18 at 15:37
  • 7
    \$\begingroup\$ I voted yes because sometimes (often?) these questions give raise to exceptionally good answers that are excellent learning material. This is true especially when the component to be identified is not just a weird connector, but a strange/old/unusual component. Maybe it looks like a cap or a transistor (e.g.), but it is not, and then the illuminating answer(s) kicks in, perhaps showing interesting niche application circuits or vintage electronics techniques/schematics involving that component.... \$\endgroup\$ – Lorenzo Donati Sep 7 '18 at 15:50
  • 6
    \$\begingroup\$ ... It is also fun: it prompts some detective work from the google masters in our community and the results are often interesting. I agree with @pipe that we should require the questioner to show effort by providing the maximum amount of context possible (I'm OK with "I just found it in my old uncle's junk-bin", but I'm not OK with "I've no time to shoot and post a photo of the PCB where I found it") \$\endgroup\$ – Lorenzo Donati Sep 7 '18 at 15:53
  • 1
    \$\begingroup\$ I agree with Lorenzo, but I do see the lack of text-based search terms as a major problem. Requiring more description would be a good idea. \$\endgroup\$ – Adam Haun Sep 7 '18 at 18:34
  • 3
    \$\begingroup\$ I'm voting yes because it's neat to read the logic behind identification questions. Admittedly, a question that's merely a picture and then an accepted answer that just says what the part is is pretty boring; I mean, that's basically just a part-name-and-picture combo, which could easily been seen in bulk on a sales website. But when an answer explains the logic that went into identifying the part, it makes for an interesting case study. (Also, I agree with most of the already-listed reasons.) \$\endgroup\$ – Nat Sep 13 '18 at 19:23
  • 1
    \$\begingroup\$ 1-5 are all valid reasons I voted yes. Additionally and personally, they serve as a good entry point to bring people to the site and allow the less highly trained members of the community (including myself) to still feel like they can contribute. I'm a tech and hobbiest just starting towards an EE, as I imagine many of us were at some point. I might not be here if I didn't think I could ask these questions as I tinker. Without them, I definitely wouldn't have this rep and would be highly discouraged by how few questions I can answer as I learn. \$\endgroup\$ – Phil C Sep 14 '18 at 0:29
9
\$\begingroup\$

Yes, but only with some conditions:

  1. There must be good closeup picture of the part with these requirements:

    1. The part must be in focus.

    2. The part must be properly lit.

    3. The picture should usually be rotated so that any writing on the board can be read. In some cases it may be more useful to orient the picture so that the writing on the part can be read.

    4. The picture must be properly cropped. This allows most of the pixels to be actually showing the part in question.

    5. The picture must be big enough to show what it needs to show right in the question. This site allows pictures up to 640 pixels wide. That should be more than enough if the above rules are followed. Don't make it big just to make it big, but make it big enough to clearly show what it needs to show.

    Basically, don't just take a phone picture and dump it on us. It should be obvious that when you're asking a bunch of people for a favor that you should take a little care to present your information as clearly and accessibly as possible. However, experience has shown us that this must be mentioned.

    If you can't meet the picture requirements, don't post the question. No, we're not going to tolerate badly lit and out of focus cell phone shots just because that's all you have to work with. It's not our fault you don't have the right equipment, or know how to use the equipment you have. Annoying pictures are annoying regardless of the cause.

  2. There should generally be a second zoomed-back picture showing the environment of the part, how it's mounted, how it fits into the larger unit, etc.

  3. If there is any text on the part, that text should also be written out in the question body.

  4. Explain what the overall unit is and what it's supposed to do. For example, knowing that a part is from a 1960s television versus a 2014 motherboard could help significantly.

  5. Show the schematic if you have it, or trace the schematic and show us that (after being properly formatted, of course) if the board allows for that.

  6. Make sure we can discern the size of the part. For example, just mentioning that the surrounding parts are 0805 surface mount could be good enough. If possible and appropriate, include a ruler or something of known and recognizable scale in the detail picture. Mentioning that the pin pitch is 0.5 mm, for example, might be all that's required.

    Explicitly giving the dimenions in the question text is good, but try to provide visual context too when reasonably possible.

\$\endgroup\$
  • \$\begingroup\$ I'm in the process of writing a guidance post for ID questions (and guidance for users on closing and scope). I think I agree with most points, except maybe the strength of 1 and 5. The number of pictures a new user can post is limited, so requiring multiple images is a good way of setting people up to fail. In the end, some questions will effectively be unanswerable, either through incomplete information provided by the user or deliberate manufacturer obfuscation. \$\endgroup\$ – W5VO Sep 8 '18 at 22:44
  • \$\begingroup\$ @W5VO: I didn't mean 5 to be a requirement. Include the schematic when you have it or the board can be reasonably traced. In some cases, of course, that doesn't apply at all, like a connector on a battery you want to find the mating part for. \$\endgroup\$ – Olin Lathrop Sep 9 '18 at 13:46
  • 1
    \$\begingroup\$ @W5VO: Point 1 is important. Without a good picture, pretty much all identification questions are impossible. Most people have access to good enough equipment for taking macro shots. The problem is people being too lazy to use it right, or too lazy to learn to use it right. That's not our problem. Low quality is low quality regardless of how it got there. A low quality picture in a identification question makes it difficult to answer, and therefore a bad question. Such questions should be closed as unclear. \$\endgroup\$ – Olin Lathrop Sep 9 '18 at 13:49
  • 1
    \$\begingroup\$ Olin, I agree that #1 is important. You obviously spend a lot of time with cameras and your pictures I've seen look great. My worries with your list and your personal criteria is that these are setups for an excellent magazine-grade photo, and that you will attack a perfectly legible photo with maybe not the best framing. If you can't read the labels on a part, and that's all you have to go by, then yes it should be closed as unclear. \$\endgroup\$ – W5VO Sep 9 '18 at 19:57
  • \$\begingroup\$ Hey, do you have your photography tips rolled up somewhere? I would like to link to it instead of writing my own. Maybe you could post your heading #1 under the new guidance post, and I'll link to it? Thanks! \$\endgroup\$ – W5VO Sep 9 '18 at 21:02
  • \$\begingroup\$ I substantially agree, except for the following points, which could be a bit too much for a newbie. 1. requiring a trace of the schematic is a bit too "reverse-engineering-heavy"; it is not always easy, and even simple circuits could lead to wrongly traced circuits. I'd prefer just to "mildly suggest" the OP to try to trace the circuit, than to make it a strong requirement.... \$\endgroup\$ – Lorenzo Donati Sep 12 '18 at 8:46
  • 1
    \$\begingroup\$ ...2. Let's try to be gentle while suggesting all those picture steps. Requiring a focused picture is "mandatory", but all other steps should be requested "gently" (i.e. providing links to free image editing programs). Not every electronics hobbyist is necessarily computer savy. A bit of education here would be useful. \$\endgroup\$ – Lorenzo Donati Sep 12 '18 at 8:47
  • \$\begingroup\$ @Lorenzo: Note that I didn't say the reverse-engineered schematic is mandatory ("if the board allows for that"). If you part is on a simple single-sided or some double-sided boards, then it might be possible to trace the schematic. In many cases it's not possible. \$\endgroup\$ – Olin Lathrop Sep 12 '18 at 11:24
  • \$\begingroup\$ @Lorenzo: As for the pictures, it's the results that matter. We need good pictures to be able recognize the part. That doesn't change just because some user is too lazy to produce them. We often don't even get focused pictures. Cropping pictures to the relevant parts is very easy to do in multiple ways. If someone is not computer savvy enough to know how to do that, then they don't get to post here. \$\endgroup\$ – Olin Lathrop Sep 12 '18 at 11:26
2
\$\begingroup\$

No:

Component identification questions should be automatically closed.

  • Due to a lack of meaningful text-based search terms, the questions are unlikely to be useful to other/future users
  • These questions frequently hit the "Hot Network Questions" list, but are little more than click-bait.
  • These questions frequently are just to enable guess-based part replacement troubleshooting, and do not display the requisite theory knowledge required for a repair question
  • Enough questions start off as poorly asked questions to merit banning the category completely (e.g. blurry pictures, illegible text, poor picture framing).
  • The questions can annoy community members

(add comments below for additional reasons)

\$\endgroup\$
  • \$\begingroup\$ Folks, could you please post a quick comment about your reason for voting "no". If the reason is already on the list, say which items. Or, add your own. I'm voting "no" because of item 3 (and 2 to a smaller extent). \$\endgroup\$ – Nick Alexeev Sep 4 '18 at 4:23
  • 2
    \$\begingroup\$ @NickAlexeev According to my data, identification questions are indeed high up on the list of last year's popular questions. It also seems that they used to be less popular 2-3 years ago. Still, we're talking about 5-6 questions here. \$\endgroup\$ – Dmitry Grigoryev Sep 4 '18 at 14:18
  • 3
    \$\begingroup\$ I'm voting "no" because I feel that identification questions became a disguise for repair/shopping questions. It's true that identification is part of engineering (a pro argument I once heard), but so is procurement, and identification questions are just as localized and quickly outdated as shopping ones. Many answers to identification questions include product links which give incentive to ask them in lieu of shopping recommendations. \$\endgroup\$ – Dmitry Grigoryev Sep 4 '18 at 14:20
  • 1
    \$\begingroup\$ @DmitryGrigoryev I'm having difficulty interpreting the data from your query. It looks like it's only finding HNQ candidate questions, but I have trouble identifying which ones are "identify" questions, let alone what "high up on the list" means in this context. \$\endgroup\$ – W5VO Sep 4 '18 at 14:42
  • 5
    \$\begingroup\$ the questions are unlikely to be useful to other/future users that was removed as a requirement over a year ago. No more "too localized" close reason, and stack overflow management made it clear that questions that help even one person, OP, are welcomed. \$\endgroup\$ – Passerby Sep 4 '18 at 14:44
  • \$\begingroup\$ @Dmitry the answer having a link is not a reason to remove the question. And most identification answers do not include a link to shopping sites any more than any other question type that has link to digikey or other sources. \$\endgroup\$ – Passerby Sep 4 '18 at 14:46
  • 4
    \$\begingroup\$ @Passerby "questions that help even one person, OP, are welcomed" - What about shopping questions then? \$\endgroup\$ – Dmitry Grigoryev Sep 4 '18 at 15:03
  • \$\begingroup\$ @W5VO The query outputs a list of tags with a number of popular questions in each tag over the last year or so. "Identification" comes in line 18 with 4 questions. \$\endgroup\$ – Dmitry Grigoryev Sep 4 '18 at 15:05
  • \$\begingroup\$ @DmitryGrigoryev Thanks for the explanation. So if I've hacked the query to return the total HNQ question count correctly (I get 362) (Wasn't correct, using 135), then that means explicit identify questions are 3.0% of HNQ (4/135) \$\endgroup\$ – W5VO Sep 4 '18 at 15:23
  • 1
    \$\begingroup\$ @W5VO Your query seems to count each question as many times as the number of tags it has. The correct count is 135. Also, "identification" seems to be far more popular than "identify". I just suggested a tag synonym. \$\endgroup\$ – Dmitry Grigoryev Sep 4 '18 at 15:53
  • \$\begingroup\$ @DmitryGrigoryev I thought I had hacked it enough to make it not count tags, but clearly not. We can do some tag synonym making to help with the statistics. \$\endgroup\$ – W5VO Sep 4 '18 at 16:01
  • \$\begingroup\$ Voting no because 1. Questions not useful to future readers. 2. component identification isn't practically useful for electronics design. It may be useful for reverse engineering, but there is a reverse engineering Stack Exchange for that. \$\endgroup\$ – The Photon Sep 9 '18 at 3:56
  • \$\begingroup\$ Voting no, because automatically infers that community moderation should not apply. There are few situations ehere that should be true, and this isn't one of those. \$\endgroup\$ – Scott Seidman Sep 9 '18 at 14:52
  • 1
    \$\begingroup\$ Didn't understand that. We don't quite automatically close lawnmower questions - we vote them closed, like any other case on the close menu. I suppose you might be asking what reason to hit when we close such questions, if they need closing, or maybe suggesting a new close option \$\endgroup\$ – Scott Seidman Sep 9 '18 at 20:17
  • 1
    \$\begingroup\$ Upvoted this answer (meaning I vote no), because it doesn't make much sense. We're also going to google the same text as they are that is shown on the IC. Yet somehow we find the result, probably because we're actually googling / trying to find it. - To me it just sounds as if OP is very lazy and use us as googlers, we don't need those kind of questioners here. - Beyond that, point 1,3 and 5 is also a part of my vote. \$\endgroup\$ – Harry Svensson Sep 10 '18 at 7:48
-3
\$\begingroup\$

Yes:

But with the term component properly defined and narrowed down. Proposal:

A component is:

  • Anything that is soldered or mounted onto a PCB.
  • Connectors, cables and cable assemblies.
  • Electromechanical parts, such as relays, switches, encoders etc.
  • Optoelectronics including LEDs, light bulbs etc.
  • RF electronics and antennas.
  • Power electronic components such as battery cells, fuses, high-voltage parts.

A component is not:

  • A mounted PCB.
  • A whole stand-alone product, or a sub-system of one.
  • Home electronics, or whole subsystems of home electronics.
\$\endgroup\$
  • \$\begingroup\$ Which excludes plenty of things which electronic engineers use as components with this pigeonholed definition. So no. -1. \$\endgroup\$ – Passerby Sep 6 '18 at 12:24
  • \$\begingroup\$ Not has anyone been asking those type of identification questions. \$\endgroup\$ – Passerby Sep 6 '18 at 12:25
  • \$\begingroup\$ @Passerby Notably, this is posted as community wiki, so if something is missing, simply add it. \$\endgroup\$ – Lundin Sep 6 '18 at 12:33
  • \$\begingroup\$ @Lundin I suppose mods will post an official question about identification rules pretty soon, where you'd be able to suggest rules and vote. \$\endgroup\$ – Dmitry Grigoryev Sep 6 '18 at 13:47
  • \$\begingroup\$ @DmitryGrigoryev Everything on meta is official, that's how Stack Exchange works. Community consensus (and SE the company...) dictate the rules, not moderators. Moderators enforce the rules. \$\endgroup\$ – Lundin Sep 6 '18 at 13:53
  • 1
    \$\begingroup\$ @Lundin My point is not to mix two topics in one post. Of course, you're free to start a discussion about rules yourself, just post it as a separate meta question please. \$\endgroup\$ – Dmitry Grigoryev Sep 6 '18 at 13:56
  • \$\begingroup\$ @Lundin We'll get to the rules in another post. Apologies if it's not moving as fast as you would like, but I have a limited amount of time I can spend on this. \$\endgroup\$ – W5VO Sep 6 '18 at 14:24
-6
\$\begingroup\$

Yes, but only with some conditions:

  • Questions about identifying replacement parts shall be treated with the guidelines for repair questions.

In an addition to other conditions already proposed.

\$\endgroup\$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .