6

EDIT Preferred form of answer is "Yes/No, because...", rather than "Because..." :)

Background:

  • All my questions here arise from my use of WPC, and accordingly initially I tagged them . I use no other WL platform and have no idea whether the questioned behaviour is shared by other platforms.

  • People have repeatedly removed that tag saying e.g. "your question was in no way specific to the use of the cloud, so I have removed it" https://archive.is/lic8h#selection-963.12-963.95 EDIT: When I restored that instance, the same person removed it again https://archive.is/UUWMA

  • Another has objected to the tagging e.g. "there is absolutely no reason for you to stubbornly tag all of your questions wolfram-programming-cloud when they have nothing to do with it." "The fact you are using it is irrelevant. If I happen to be eating when I ask a question, I don't tag it sandwich, do I?" https://archive.is/HuPrc#selection-831.327-839.7

  • When I then omitted the tag from a question, a reader added the tag, saying "Added the Wolfram Programming Cloud tag because the described behavior seems confined to that platform, and not to desktop MMA 10.2" https://archive.is/kGKZz#selection-345.0-345.131. Tags are surely parts of the question, not the answer.

  • Another user said "@ChrisJJ do us all a favor. ... If you don't have access to [the desktop version of Mathematica], leave a comment stating you have not been able to compare the two. This small extra step will help keep the peace" https://archive.is/2l3E4#selection-849.150-853.37 I take this to suggest that in this forum for "Mathematica .... as well as related products", every question from a non-Mathematica user needs a note saying the behaviour has not been tested on Mathematica.

Note: This is not a RTFM issue. WL does exhibit undocumented variance in behaviour across platforms.

Note: Edited to replace links with links to archived versions.

  • 3
    Thanks for asking this question. When there is not an established practice (which is exactly what this question should hopefully produce) we end up with mixed messages like you received. – Mr.Wizard Jul 29 '15 at 22:35
  • Nice. Exactly what was needed. – rcollyer Jul 30 '15 at 1:49
  • 1
    A quick note, stackexchange archives everything. So, referencing things via the internet archive is not necessary. – rcollyer Jul 30 '15 at 3:25
  • @rcollyer Occasionally moderators clean out the comments that are not on topic. Meta discussions in particular risk being deleted like that. I'm talking here of SE in general, I've seen it happen many times at for example workplace.SE which is more prone to off topic chatter than we are. – C. E. Jul 30 '15 at 7:03
  • @Pickett, true, but if push comes to shove, mods can see deleted comments, and we can show the original flow of conversation if need be. – J. M. will be back soon Jul 30 '15 at 10:33
  • @Pickett also, as a site, we are very lax about comments. So, they don't get deleted as often as you'd think, here. :) – rcollyer Jul 30 '15 at 11:55
  • @rcollyer I know that, but the OP has no idea how this site works (based on his criticism here and elsewhere). So just saying. – C. E. Jul 30 '15 at 15:23
  • 1
    @Pickett absolutely. just making sure we're all clear. – rcollyer Jul 30 '15 at 15:40
  • @rcollyer "stackexchange archives everything. So, referencing things via the internet archive is not necessary." Referencing things via the internet archive is necessary.to avoid links to questions showing something different, as already happened in discussions about off-topic markings that were since rescinded. – ChrisJJ Jul 30 '15 at 16:45
  • 2
    @ChrisJJ I know you believe that, and the evidence would seem to suggest it, but dig a little deeper. All the edits are archived along with who made them, which you are aware of by the links you made, and deletions can be accessed by the moderators. Also, there’s Stack Exchange Data Explorer which gives you query access to the underlying database. The info is not going anywhere. On this site, we are not adversarial (I make no claims about Stack Overflow), and we genuinely wish to help you. (Some of it is intended to be gentle instruction, like the links.) Please give us the chance to do so. – rcollyer Jul 30 '15 at 17:06
20

I feel the best way to deal with potential platform dependencies is to mention your platform in the question and don't use the platform name as tag unless you are sure the platform plays a role. We can only use 5 tags and if everyone is going to tag his OS and version number, regardless of relevance, differentiation will be hugely impacted.

If you own different platforms you can do the experimentation wrt system specificity yourself; if not, wait until community members have checked this. We do this for the "bugs" tag as well.

Related questions can be found in the links below. In the first there are not an awful lot of reactions but I'd say the gist is that system tags are useful if the issue is system specific. [I'd like to add: But only then].

  • A similar discussion about version tags here. – Sjoerd C. de Vries Jul 29 '15 at 22:46
  • Many of us used to tag bugs with the version number. This procedure had some problems so we now use a more or less standard header formalized in this Q&A and don't tag with a version anymore. – Sjoerd C. de Vries Jul 29 '15 at 22:53
  • Recommended putting all these links in the answer proper. And I'd include this one too: (1361) – Mr.Wizard Jul 29 '15 at 23:01
  • 1
    I'll do that tomorrow. One-finger editing on an iPad is not really a party. – Sjoerd C. de Vries Jul 29 '15 at 23:23
  • I'll do what I had in mind; you can revert or correct later if needed. – Mr.Wizard Jul 29 '15 at 23:25
  • "don't use the platform name as tag unless you are sure the platform plays a role" That's quite a departure from stackexchange custom, and this forum's advice mathematica.stackexchange.com/help/tagging , esp. "A tag is a word or phrase that describes the topic of the question." I'm not saying it's a bad idea, just that it sounds more like a proposal for change rather than a description of current usage. – ChrisJJ Jul 30 '15 at 0:35
  • It is in no way a proposal, @Chris. This was community consensus, agreed upon before you started asking questions here. Again, less is more. If you're not sure, don't add the tag yourself; let it be added after community vetting. – J. M. will be back soon Jul 30 '15 at 0:59
  • "This was community consensus" Perhaps it was then, but I think you'll find newcomers will go by mathematica.stackexchange.com/help/tagging "What are tags, and how should I use them?" which makes no mention of this. – ChrisJJ Jul 30 '15 at 1:06
  • 2
    IMO: WPC can only be the topic of a question if it conceivable that "Mathematica 10.2" could be the topic of a similar question. If you are asking about the behaviour of the Wolfram Language, then WPC is not an appropriate tag. Analogy: if you were asking a question about Swift, you wouldn't tag a question "xcode". If you think it's really a WPC problem (that is, a problem not with your use of the Wolfram Language but with how the WPC interprets your WL code, or a question about how the WPC works which is not really a question about the WL) then the tag is appropriate. – Patrick Stevens Jul 30 '15 at 6:36
  • 4
    @chrisjj In order for all of us to get an idea of what the community thinks about this you might consider writing a separate answer voicing your opinion about the issue at hand. In that way, people can vote about this and policy can be settled based on the results. – Sjoerd C. de Vries Jul 30 '15 at 7:03
  • 7
    @ChrisJJ, would you really tag a similar C question "gcc" or "ubuntu", because those were what you were using to compile/view the results of the program? If your mind is anything like mine, it wouldn't even have occurred to you to tag a C question with the specific compiler you were using unless you thought your question was compiler-specific. In that situation, don't you agree it would be standard to post your question about C, and subsequently for people to tell you it was actually a gcc question (if it were)? – Patrick Stevens Jul 30 '15 at 14:52
  • 5
    @ChrisJJ, I would object to that, yes. I would delete the tag, and if the author restored it, I would see if they had any reason for thinking that the issue was related to GCC. If they did not have such a reason, I would once again delete the tag. "WPC" is there as a tag so that people can search for issues related to the WPC, not so that people can search for issues which happen to manifest on the WPC as well as other places. Remember, StackExchange is there to help people in the future as well as you; having everything labelled "gcc" or "clang" is useless to people who are trying to learn C. – Patrick Stevens Jul 30 '15 at 15:34
  • 4
    @ChrisJJ: I'm sorry, but very-preliminary results suggest consensus is against that view, judging from the upvotes on my comments. Tag status is determined by the question, not by the intentions of the poster. Questions are answered on StackExchange so that the answer may help anyone with the same problem in future, not specifically for your assistance, so tagging is in principle determinable just by seeing the question, independent of the poster. I'll write this into an answer - thanks for the prompt. – Patrick Stevens Jul 30 '15 at 16:27
  • 2
    @ChrisJJ: observe the existence of the "too localized" reason for closing, which to me is clear evidence that the intent of StackExchange is to provide a place for the answering of questions in such a way as to help more than one person. I'm going to duck out of this conversation now, as I'm getting heated up. – Patrick Stevens Jul 30 '15 at 16:51
  • 3
    @chrisjj Unfortunately for some perhaps, the world is not black or white. I'm not saying yes or no, because the issue needs more nuance than that. If the issue is very specific to WPC by all means add the tag. But if it isn't or if you are not able to make sure it is, don't. – Sjoerd C. de Vries Jul 30 '15 at 18:32
10

No to the question you intend to ask; yes to the question you actually asked.

The tag is for questions which are related inherently to the WPC. If your question is about the Wolfram Language, whether or not you access the WL via the WPC, then it does not warrant the tag. If your question is about something you think is a bug in the WPC, then it should have the tag. If your question is about the WPC interface, then it should have the tag. If you would have the same question while using Mathematica (referring to the desktop program), then it should not have the tag, except under weirdly specific circumstances like a bug which manifests differently in both environments.

The reasoning behind this is that the tags are there to help filter the questions for people in the future, and for people who are currently trying to answer questions. questions are expected specifically to be about the WPC, in the same way as questions are expected specifically to be about integration; that way, people with the right expertise can find relevant questions. For instance, if I used the function Integrate as an example to demonstrate a general principle about UpValues in Mathematica (and there was nothing about my answer that inherently used Integrate), then the question would not warrant the tag , because the main idea of the question would really be UpValues. Likewise, the fact that I use the WPC to formulate my Wolfram Language programs is not enough to justify the tag. It is unhelpful to future visitors to use that tag in this context, because the question isn't really about the WPC at all.

Imagine you were an expert on the ins and outs of the WPC, how to use it to its fullest potential, and all its interfacing questions. You search for the tag with the intention of providing specialist knowledge, and are presented with thousands of questions about the Wolfram Language, almost none of which actually require knowledge about the WPC to answer. I think one, in the position of WPC expert, would justifiably be annoyed in that situation.

  • "No to the question you intend to ask; yes to the question you actually asked." The question I asked is the question I intended to ask. "If your question is about the Wolfram Language, whether or not you access the WL via the WPC, then it does not warrant the tag." Sorry, that does not accord with the stated policy of this forum mathematica.stackexchange.com/help/tagging – ChrisJJ Jul 30 '15 at 16:51
  • I may come back to this tomorrow; as commented above, I'm leaving this for now. – Patrick Stevens Jul 30 '15 at 16:53
  • And on "If your question is about the Wolfram Language, whether or not you access the WL via the WPC, then it does not warrant the tag." a point that seems to have got missed despite my saying it many times: WL on WPC is not the same as WL on other platforms. There is variance, both declared and not declared in docs. – ChrisJJ Jul 30 '15 at 17:13
  • 4
    @ChrisJJ that is mostly incorrect. The exceptions surround the need to display on a webpage vs desktop computer and to execute arbitrary code in a cloud environment. The rest is WL and is run by the same kernel (the execution) engine as found on the desktop. – rcollyer Jul 30 '15 at 17:25
  • @rcollyer What makes that statement correct is the exceptions. "The rest is WL and is run by the same kernel (the execution) engine as found on the desktop" Evidently not e.g. try SocialMediaData[]. – ChrisJJ Jul 30 '15 at 22:47
  • 3
    @ChrisJJ congrats! You found something that is different. Now, I don't know why it is excluded, but it is in essence a loaded library. Have you never worked on a platform where a library was unavailable? My post stands. The execution engine is identical with a few exceptions, and the major difference is in the display. – rcollyer Jul 31 '15 at 3:14
  • 1
    @ChrisJJ the question of what functionality is not available on WPC is a legitimate one, and should be asked. I think it would be valuable information. Please ask it on Mathematica. – rcollyer Jul 31 '15 at 12:06
  • @Chris, as rcollyer says, that would be a well-received question on the main site should you choose to ask it; just… ease up on the 'tude, okay? – J. M. will be back soon Jul 31 '15 at 19:55
  • @rcollyer "it is in essence a loaded library" So what? Such implementation details is irrelevant. What is relevant is that WL differs across platforms. "The execution engine is identical with a few exceptions" Thanks for the confirmation. – ChrisJJ Aug 1 '15 at 1:22
  • @ChrisJJ your misunderstanding makes sense now. You're coming at this from other languages where the key-words are immutable constructs, such as auto, class, and return in c++. Put this out of your mind. In Mathematica, unless the symbol is Locked it can be overridden. So, the functionality that comes with WL cannot be considered key-words in any sense remotely like any language you've likely encountered. For example, if I create System`bob it is indistinguishable from the default WL, and yet, should it be considered part of WL? – rcollyer Aug 1 '15 at 3:08
  • @ChrisJJ (cont’d) So, when I say the language is the same, except for the cloud constraints, I mean it. I’m not being facetious. Only what is allowed (in the computer system sense) has changed, but the language remains the same. As I said, I do not know why SocialMediaData is excluded, but a surprising amount is allowed, such as URLSave. I’m curious as to what else is excluded. – rcollyer Aug 1 '15 at 3:08
  • @rcollyer "So, when I say the language is the same, except for the cloud constraints, I mean it." Show a list of these "cloud constraints" and that statement might actually be useful. Until then, it isn't. "Only what is allowed (in the computer system sense) has changed" So you know what's changed? " I’m curious as to what else is excluded" So you don't know what's changed. Thanks anyway for the suggestions. – ChrisJJ Aug 1 '15 at 20:06
  • @rcollyer "your misunderstanding makes sense now. You're coming at this from other languages where the key-words are immutable constructs" Nope. WL's mutability is a matter totally separate to the variance of WL's behaviour across platforms. – ChrisJJ Aug 1 '15 at 20:30
  • 2
    @ChrisJJ I'm done bothering. No matter what I say, you're going to continue believing what you wish. Good day. – rcollyer Aug 1 '15 at 22:32
8

Is the tag appropriate for a question that is specific to the WPC?
No. That would be selfish. The intent of this site is to house a knowledge repository that would be helpful to future users. Otherwise there would be no point in, for example, closing duplicate questions. Similarly, the many, who use only plain old Mathematica don't tag every question as such.

Is the tag appropriate for a question on an issue that is specific to the WPC?
Yes. But, until you have verified that, there's no grounds to include the tag.

Edit: Let me also quote the rules, to which you appeal so.

  • If the tag can’t work as the only tag on a question, it’s probably a meta-tag. Every tag you use should be able to work, more or less, as the only tag on a question. Meta-tags, like [beginner], [subjective], and [best-practices], are not helpful by themselves – they do not communicate anything about the content of the question.
  • If the tag commonly means different things to different people, it’s probably a meta-tag. For example, the meaning of the tag [subjective] is, itself, subjective; the same is true for tags like [best-practices] and [beginner]. Best practices to whom? Beginner by what criteria? Use only tags that have a broadly accepted, objective definition.

not only is insufficient as the only tag (just like Mathematica-related questions almost never have a tag, but always have some like ), it is clearly perceived differently, than you take it to be perceived. That is, if I would for some reason click on , I would most certainly not be searching for problems of users, that happen to have access only to WPC. I'd likely be searching for subtleties specific to WPC.

  • 2
    "That would be selfish. The intent of this site is to house a knowledge repository that would be helpful to future users." - +1 just for that. – J. M. will be back soon Jul 31 '15 at 1:32
  • @Guesswhoitis. thanks for the formatting. Post my edit I realized, that there are no Mathematica tags :D – LLlAMnYP Jul 31 '15 at 6:03
  • We actually blacklisted that tag during the first few weeks of this site, yes. :) – J. M. will be back soon Jul 31 '15 at 7:09
  • @Guesswhoitis. it took that long? I thought it was within the first few hours. :) – rcollyer Jul 31 '15 at 12:09
  • @rcollyer, it was so long ago that hours and weeks seem the same. In contrast, these days an hour can sometimes feel like a week, and less often, a week can feel like an hour. ;) – J. M. will be back soon Jul 31 '15 at 12:32
  • 2
    @ChrisJJ I think you got the answers to most of your questions here even if you may not like some of them, so do read it out of the multiple answers you got and deal with it. Now specifically to address your comments: 1) Go edit the tag wiki if you aren't satisfied. 2) Because "default platform". Most questions here are from users of Mathematica proper. We do have version tags for version-specific issues. 3) I'm getting kinda' tired of constant appeals to generic sites over which we have little to no control. Go complain on meta.SE about that (cont'd) – LLlAMnYP Aug 1 '15 at 5:15
  • 1
    No, we are not building a database of every question on Mathematica (and related products). It seems to be the default "philosophy" of SE (every question about XXX should be on this site), but this community decided that we don't need a wealth of questions such as "how do I multiply a and b in MMA" and "how do I add c and d in WPC". For this reason we have the custom close reason "simple mistake or easily found in the documentation". Anyway, regarding your actual question. This site is not your personal coach (cont'd) – LLlAMnYP Aug 1 '15 at 5:26
  • 1
    You bring up an issue, the community researches the issue rather than researching you. The question is (should be) valuable without the asker. For example, we delete questions about bugs/errors that only seem to happen on the asker's machine and cannot be reproduced. Just questions and answers, no distractions... Also, from a completely different angle, but regarding your usage of tags, I might point out, that wikipedia has an article on Cerenokov radiation (cont'd) – LLlAMnYP Aug 1 '15 at 5:38
  • 1
    But it doesn't have separate articles on Cerenkov radiation in water, Cerenkov radiation in oil, Cerenkov radiation in some other optically dense medium etc. To sum everything up, it looks to me, as if the WPC-tag question has been answered. Separating that from follow-up questions, if something's not clear, feel free to ask. Regarding the follow-up questions, which all seem to be along the lines of "but this seems to conflict with the help pages which I liberally interpret as XXX" (cont'd) – LLlAMnYP Aug 1 '15 at 5:43
  • 2
    It has again been clearly stated several times, that the help is not perfect, sometimes there's some ambiguity, but the community's policy is this or that. It's all good and well to make mistakes, especially so when there's some ambiguity in some help pages. It isn't ok to flip out and say "I don't give a crap, what you mods say, because it isn't explicitly written on page XXX". – LLlAMnYP Aug 1 '15 at 5:51
  • 2
    Chris, are you capable with dealing with imperfect documentation? I'm sorry, but it's not going to change without a very serious reason. Over parts of it the mods here have no control. – LLlAMnYP Aug 1 '15 at 21:16
  • 1
    @ChrisJJ I disagree with your interpretation of the documentation. Also I disagree with thoughtless attempts to apply the standards of SE in general to any given community. Although parts of the general documentation of SE leak into the documentations of each specific community, however often the mods have no control over that. Nonetheless, every community is run not by SE employees, but by the specific community itself. This principle, IMO, prevails over whatever general approaches SE has. Thus, if the community chooses to operate in a way, that in parts conflicts (cont'd) – LLlAMnYP Aug 5 '15 at 14:39
  • 1
    with the overall principles of SE, the rules set by the community should have precedence over the rules set by SE. Now, as far as interpreting the rules, law is not a simple subject. When talking about laws of a given nation, it is sometimes unclear, what interpretation to go with. In that case, it is usually up to the constitutional court or the supreme court to publish its own interpretation, which is the only correct one. Drawing parallels to the community here, it is the community elected mods, that do the arbitrage. (cont'd) – LLlAMnYP Aug 5 '15 at 14:42
  • 2
    Whatever interpretation the mods present is final and most certainly overrules whatever you yourself have read out of the help pages. The "nice" way to go about it is to admit "sorry, I had read that differently, I'll not make the same mistake in the future, maybe you could make the wording of such and such paragraph a bit clearer". You however, have created several posts which sound like demands along the lines of "you're giving me crap, change that page immediately, otherwise I wish not to hear of what you mods have to say". – LLlAMnYP Aug 5 '15 at 14:44
  • 1
    Pointing out the exact words that have mislead you to believe otherwise would also be helpful, rather than just distributing links and saying "your opinion is crap and this page says otherwise". – LLlAMnYP Aug 5 '15 at 14:46

You must log in to answer this question.

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