10
$\begingroup$

I suggest that we rename the tag to for the following reasons:

  1. The word graphs is ambiguous.

  2. looks more professional.

The tag graph might have legitimate uses, but the question tagged with graphs so far are questions in graph theory.

Please express you agreement/disagreement with the suggested change by voting on the question. You can take back or change your vote in case you thought that votes on the question were not meant to be for/against the suggestion.

$\endgroup$
  • $\begingroup$ How would you tag questions about graph algorithms? Graph theory is, for me, the mathematical studies of graphs. I don't see how that applies to most questions we have with graphs. $\endgroup$ – Raphael Mar 25 '12 at 15:30
  • 3
    $\begingroup$ @Raphael, by algorithms and graph-theory tags. Graph theory is not just the mathematical study of the graphs, see the wikipedia article or the standard textbooks on the topic like "Intro. to Graph Theory" by D.B. West or by J. Bondy. Graph algorithms is a large part of graph theory. $\endgroup$ – Kaveh Mar 25 '12 at 15:33
  • 4
    $\begingroup$ I agree, and favor making a synonym graphsgraph-theory. $\endgroup$ – Gilles 'SO- stop being evil' Mar 25 '12 at 15:34
  • $\begingroup$ We've got quite a few new questions in these tags, do you still think they should be made synonyms? $\endgroup$ – Gilles 'SO- stop being evil' Apr 29 '12 at 20:45
  • $\begingroup$ @Gilles, yes, I think the points still apply. $\endgroup$ – Kaveh Apr 30 '12 at 1:13
  • $\begingroup$ @Gilles: Is the proposed synonym a good idea, given that graphs is ambiguous? Do we expect questions about other graphs than those with nodes and edges? $\endgroup$ – Raphael Apr 30 '12 at 17:06
  • $\begingroup$ the other meaning can be stated more explicitly like function-graphs and graphics so they are not ambiguous. $\endgroup$ – Kaveh Apr 30 '12 at 21:48
  • 1
    $\begingroup$ Are you serious about “[graph-theory] looks more professional”? [graphs] does look equally professional to me, although I completely agree with you that [graphs] can mean graphs of functions and therefore graphs as in graph theory should be tagged as graph-theory. $\endgroup$ – Tsuyoshi Ito Jun 10 '12 at 12:17
  • $\begingroup$ @Tsuyoshi, I felt so, but it might be just my personal bias. $\endgroup$ – Kaveh Jun 10 '12 at 19:56
  • 1
    $\begingroup$ @Raphael Just thinking out loud... hould we go ahead and make a decision on this? Are we going to get ride of graphs and use graph-theory, then? $\endgroup$ – Patrick87 Jun 25 '12 at 4:10
  • $\begingroup$ @Patrick87: We could, but the "voting" is not very clear (+2/-1 vs +0/-2), igboring the question itself; apparently not many people care (which we can use as indicator that most are fine with either decision?). Any opposition to the mods deciding? $\endgroup$ – Raphael Jun 25 '12 at 8:09
  • $\begingroup$ @Raphael, the number of votes on the question shows quite well that people care and have expressed their view about changing the tag on these questions to graph-theory. The reason I personally haven't voted on the answers is that as I said before I don't have an opinion on keeping or removing the graphs tag, that is a side issue. Others might have similar opinions. $\endgroup$ – Kaveh Jun 25 '12 at 17:05
  • $\begingroup$ @Kaveh: Your question text does not talk about retagging, but about "renaming" graphs to graph-theory. If that is supposed to mean anything (there is no such function for non-staff) then I have to read it as "Get rid of graphs". This is equivalent to Patrick's answer which has barely been voted on, which supports my concern that votes on the question can not be considered votes on the matter. $\endgroup$ – Raphael Jun 25 '12 at 18:34
  • $\begingroup$ @Raphael, renaming the tag doesn't mean getting rid of the tag forever. I think I have replied to your concern before and no one seems to have changed their vote since the edit last week. $\endgroup$ – Kaveh Jun 25 '12 at 21:10
  • $\begingroup$ @Kaveh: That's a weak argument; based on general meta activity, I doubt that most voters have even noted the edit. $\endgroup$ – Raphael Jun 26 '12 at 7:41
3
$\begingroup$

I'll throw a possibility into the mix. Why not keep and lose ? I don't personally have any hang-ups about using and for algorithms about graphs, or and for questions about representations or whatever.

I think that reasonable people can take off their technical hats and accept a definition of along the lines of "the study of those things with vertices and edges." No need to mince words.

$\endgroup$
  • $\begingroup$ I am against attaching the -theory suffix without reasonable cause. Practitioners are bound to be confused by academic distinctions between theory and others, anyway, and telling them "Well, as soon as you use graphs you are doing theory." won't help. $\endgroup$ – Raphael Apr 30 '12 at 17:08
  • 1
    $\begingroup$ What's the point of the -theory suffix then? Why would I use graph-theory and not create the more evidently applicable graphs if my question is not about theory, say a question about a graph operation in a compiler (variable interference graphs)? (I've kind of changed my mind as to which tag should be the main tag if we merge. In part because we've got several graph-theory questions that could be graphs, but questions like this don't look much like graph theory.) $\endgroup$ – Gilles 'SO- stop being evil' Apr 30 '12 at 19:02
  • 1
    $\begingroup$ @Raphael It's sort of a situation similar to the one with "integrity" and "data-integrity": "graph" is a common English word and, even in the context of CS, could reasonably mean a variety of things, not all of which have anything to do with "graph theory". This is less of a problem with other terms, like "tree", which - while having a meaning other than the technical CS one - would (usually) be expected not to come up during discussions of CS. I mean, any chart or plot counts as a "graph", and we can reasonably expect to see plenty of those. $\endgroup$ – Patrick87 Apr 30 '12 at 19:09
  • 1
    $\begingroup$ @Gilles Without reading the description of the tag, new users shouldn't be expected to know whether "graphs" refers to pictures or two the things with nodes and edges, even if they are experienced with both kinds of things. IMHO, that makes "graphs" a bad tag. "graph-theory" is much less objectionable on these grounds, since it typically (in my experience, always) refers to the things with nodes and edges. $\endgroup$ – Patrick87 Apr 30 '12 at 19:11
  • $\begingroup$ @Patrick87 My point is that if we don't have a graphs tag, it's going to be created by someone who won't think that graph-theory applies to his question. $\endgroup$ – Gilles 'SO- stop being evil' Apr 30 '12 at 19:14
  • $\begingroup$ @Gilles While that may be true, we could at least temper that with a good tag wiki, and routine retagging isn't something I'd imagine to be out of the norm on SE sites. $\endgroup$ – Patrick87 Apr 30 '12 at 19:17
  • $\begingroup$ @Patrick87 A tag wiki can clarify the meaning of an existing tag. Someone who looks for the graphs tag and doesn't find it isn't going to be helped by a tag wiki on some other tag. If you're going to be systematically retagging graphs into graph-theory, there's a tool for that: tag synonyms. $\endgroup$ – Gilles 'SO- stop being evil' Apr 30 '12 at 19:19
  • $\begingroup$ @Gilles, IMHO, that question doesn't need a graph tag at all. It is asking for a data-structure, I don't see how it is related to graphs other that OP not wanting DAGs. I also went over all question tagged graphs and the graph-theory tag would apply to all with the exception of that one. Even if we want to keep the graph tag these questions should be merged to graph-theory. $\endgroup$ – Kaveh Apr 30 '12 at 21:42
  • $\begingroup$ @Kaveh Ok, bad example, but I still disagree with removing graphs altogether. I'd prefer making graph-theory redirect to graphs but would be ok with the other direction. $\endgroup$ – Gilles 'SO- stop being evil' Apr 30 '12 at 21:43
  • $\begingroup$ @Gilles, I not suggesting removing the graphs tag altogether, I don't have an opinion on that. The questions tagged graphs so far are all graph-theory questions and IMO that should be the tag for them. $\endgroup$ – Kaveh Apr 30 '12 at 21:53
0
$\begingroup$

I just want to summarize discussion from my point of view, since seems this is still an open problem here.

is not suitable to express a question in graph theory, i.e graph is a general term not scientific name and can be used for charts or even in graphic design, functions, ..., which are not related to computer science. e.g how to draw this graph with tikz is not a graph theory or computer science related but is graph related. Because of three main reasons I want to propose to use tag synonym for this:

  1. Everyone who works on graph-theory expects to see for questions which are related to algorithms, complexity classes, ... of graphs. e.g there is no in CSTheory but there are specific tags like , ... which are more specific and is not comparable with general term graph which causes to misinterpreting.

  2. Currently all questions that I saw them and tagged by graphs are related to + something or without graph theory just by using other tags. Generally, if some question tagged with , and is not possible to change it to + other tags, means is not computer science problem, it's math, graphic, painting, ... , by writer's understanding and knowledge about CS.SE, we don't have any evidence against this.1

  3. New people in graph theory or computer scientist who does not work on graph theory related topics, may call as . So for fixing terminology, I would suggest to use tag synonym to change graph/s to .

1. If anyone thinks there is a question with graph tag and cannot be expressed by other tags, is better to share it here.

$\endgroup$
  • 1
    $\begingroup$ Your ill-advised editing spree aside, I still don't follow your reasoning. The claim that there "is not [a] scientific name" is ridiculous; the object graph theory studies is unequivocally named graph. As a counter-argument to your conclusion (merge the tags), consider this question (there are others). It's clearly not a graph-theory question, but it is as clearly about graphs. How would you tag it? $\endgroup$ – Raphael Sep 23 '13 at 7:07
  • $\begingroup$ Regarding term overloading: this happens all the time and does not bother us at all. Automata, machine, languages, proof, etc -- there are many words that have other meaning in other fields (or the real world). Some even have multiple meaning inside CS, e.g. simulation. If that were to disqualify the use of any of these terms, we could hardly talk about anything. But, luckily, the association with "graph" is quite clear (afaik) in computer science, namely nodes and edges. So there is no real problem. $\endgroup$ – Raphael Sep 23 '13 at 7:12
  • $\begingroup$ By the way, it's true that some newcomers use graphs wrongly, but that's not not exclusive to this tag; they use tags wrong all the time. Lucky for us, that's easily remedied. $\endgroup$ – Raphael Sep 23 '13 at 7:12
  • $\begingroup$ Saeed: Please be nice. $\endgroup$ – Gilles 'SO- stop being evil' Sep 23 '13 at 9:27
  • $\begingroup$ @Raphael, huh? again you talk in a rude manner? Still you didn't learned that here is not a hospital or school? huh? $\endgroup$ – user742 Sep 23 '13 at 15:16
  • $\begingroup$ @delete aka Saeed: Neither do I see where my comments are rude, nor why you would think it would be acceptable to be rude in a school or hospital. Feel free to contact the community moderators via team@stackexchange.com, though. Good luck. $\endgroup$ – Raphael Sep 23 '13 at 17:26
  • $\begingroup$ @Raphael, "ill-advised editing spree aside" is rude. may be is not rude in your opinion, this is just by our personal definition of rudeness, but I think either you didn't think about what you said (childish->school), or you thought about it, and you know what you say (rudeness--> instead of madhouse I wrote hospital to use a little bit polite format). By the way this is totally annoying conversation, and not constructive at all. Just look at your totally nonsense start of first comment, are you series? this is an answer to my post? by the way is not important at all ... $\endgroup$ – user742 Sep 23 '13 at 18:20
-2
$\begingroup$

So far, only 7/25 questions tagged with are also tagged with , and 7/14 in the other direction.

I still feel there is a difference. I would not tag questions like Where to get graphs to test my search algorithms against with .

For most questions, I can not give a clear criteria of distinction, though. Therefore, I can not speak out against the merge. However, I suggest retaining as the main tag, with the synonym, as appears to be more inclusive.

$\endgroup$
  • $\begingroup$ Downvotes and no comments? Come on, guys, you are better than that. $\endgroup$ – Raphael Apr 30 '12 at 17:04
  • 2
    $\begingroup$ I would tag the particular question you linked to with [graph-theory]. Honestly speaking, if Dijkstra’s algorithm and depth-first search in graphs are not graph theory, I do not know what you call graph theory at all. $\endgroup$ – Tsuyoshi Ito Jun 10 '12 at 12:14
  • $\begingroup$ @TsuyoshiIto: But that question is not about the algorithms but where to get graphs. $\endgroup$ – Raphael Jun 10 '12 at 21:43
  • 2
    $\begingroup$ The question asks for data for benchmarking of graph algorithms. It does not asks for algorithms, but that does not mean that the question is not about graph algorithms. $\endgroup$ – Tsuyoshi Ito Jun 10 '12 at 21:46
  • $\begingroup$ @Raphael, I think we should get rid of graphs tag, as other mentioned, this is not appreciated tag in most cases, and as you can see makes confusion. $\endgroup$ – user742 Sep 20 '13 at 10:08
  • $\begingroup$ @SaeedAmiri This is the first time in over a year that I heard anyone complain about the tag or its use, so I don't quite follow your reasoning. What I see is that one user, you, disagrees with its use. This one data point is of little use, unfortunately. See my other comment; if you want a new discussion, please open a new thread. $\endgroup$ – Raphael Sep 20 '13 at 10:22
  • $\begingroup$ @Raphael, Actually I already read all of comments, and e.g Tshuyoshi Ito's comments to your answers are clarifying that, this is not only me who see some question believes that the question is about graph theory (not general term graphs), but if you think I missed some point I want to hear it. By the way I'll try to prepare more precise answer to your possible questions. $\endgroup$ – user742 Sep 20 '13 at 11:16
-3
$\begingroup$

In order to get a clearer voting picture:

Let us keep and . We will have to clarify which questions go where, but they describe different things.

$\endgroup$
  • 1
    $\begingroup$ Contrary to your intent, this post is making the voting picture rather obscure. Kaveh is not proposing to remove [graphs] permanently; he admits that [graphs] may have some legitimate use. His point is that the questions tagged with [graphs] currently should be retagged with [graph-theory] (which I have not confirmed). $\endgroup$ – Tsuyoshi Ito Jun 10 '12 at 12:12
  • $\begingroup$ @TsuyoshiIto: Kaveh's post is not an answer. If any "voting" is to take place, then amongst the answers. My other and Patrick's answers proposed to get rid one tag but none got a positive score, so I thought there should be an answer to vote on if you want to to keep both. $\endgroup$ – Raphael Jun 10 '12 at 21:45
  • 1
    $\begingroup$ “If any "voting" is to take place, then amongst the answers.” I do not care whether you call it voting or not, but it seems clear to me that currently many people like Kaveh’s proposal that graphs as in graph theory should be tagged as [graph-theory] (no proposal on whether [graphs] tag should be used for something else). $\endgroup$ – Tsuyoshi Ito Jun 10 '12 at 23:23
  • $\begingroup$ @TsuyoshiIto: Not at all. Upvoting the question can also (and should!) mean: "Good question, thanks for bringing that up!" Note that we don't have policy-voting conventions as those on meta.cstheory in place here. Patrick proposes the same solution and only got one vote -- go figure. $\endgroup$ – Raphael Jun 10 '12 at 23:27
  • 2
    $\begingroup$ As I repeatedly said, Kaveh did not propose anything about possible use of [graphs] tag. Patrick87’s proposal includes not to use [graphs] as a tag. They are different. $\endgroup$ – Tsuyoshi Ito Jun 10 '12 at 23:29
  • $\begingroup$ @TsuyoshiIto: Well, then this answer is equivalent to Kaveh's proposal (w.r.t. tag removal). $\endgroup$ – Raphael Jun 10 '12 at 23:40
  • 1
    $\begingroup$ Sigh. “No proposal” is different from both a proposal to keep and a proposal to remove. Is this logic that difficult to understand? $\endgroup$ – Tsuyoshi Ito Jun 10 '12 at 23:49
  • 1
    $\begingroup$ @Raphael, I think when a question clearly proposes a change the votes are not for "Good question, thanks for bringing that up!" Note that this has been discussed already in several places. (but for clarity next time I will explicitly state that the votes on question will be considered votes for/against proposed change.) $\endgroup$ – Kaveh Jun 18 '12 at 17:50
  • $\begingroup$ @Kaveh: I don't think that is a good idea. You should rather post your own answer. Anyway, as it is here we have no way to interpret the voting pattern. $\endgroup$ – Raphael Jun 18 '12 at 17:53
  • 1
    $\begingroup$ @Raphael, I disagree. If you search meta you will see that this has come up several times, e.g. here. My goal in posting this question wasn't starting a new discussion but to get votes to get it done. $\endgroup$ – Kaveh Jun 18 '12 at 17:54
  • $\begingroup$ @Kaveh: I don't see where graphs vs graph-theory has come up. Can you refer me to the posts? We certainly don't have an established way here about how to put policies/tag changes to a vote. What is the trouble with creating an answer? That would allow people to vote on the question as "I like this question" and would clearly put all possibilities there on the same level. Your way causes unnecessary restrictions (and confusion, as evidenced by this thread). $\endgroup$ – Raphael Jun 18 '12 at 17:57
  • 1
    $\begingroup$ @Raphael, the problem is that this is stated as a suggestion and people have already voted on it over a month. I can edit the question and make it explicit and that will give people who would like to retract their votes a chance to do so. $\endgroup$ – Kaveh Jun 18 '12 at 18:00
  • $\begingroup$ I think one of a reason that graphs question has low accepted answer rate or fewer answers than other is the kind of tagging, If somebody tag it as graph-theory, then question will be answered normally in good fashion, but if tag it with graph, is not visible to many members search, and they should be lucky to get good answer. $\endgroup$ – user742 Sep 20 '13 at 9:37
  • 1
    $\begingroup$ @SaeedAmiri It's an unwritten rule which everybody abides by (besides you, apparently), even mods. It's also common sense: action should only be taken if the community wants is (with some exceptions that are covered by rules); you did not bother to check for that necessary condition. Yes, I could just revert your edits at the risk of starting an "edit war" (requiring me to protect posts and/or ban you temporarily), or I could spend my time trying to explain to you what the proper process is. Apparently, that's not a worthwhile use of my time. $\endgroup$ – Raphael Sep 20 '13 at 13:11
  • 1
    $\begingroup$ @SaeedAmiri Beyond it being common sense, there are threads about flooding the front page on Meta Stack Exchange. Regarding who answers, graph-theory has twice the followers of graphs. I don't understand your purpose with these retags though. Why did you retag those questions? If you see a distinction between the meaning, you need to make it clear and get the community to agree with that distinction on meta first. If there is no distinction, then the tags should be merged, which does not involve editing questions. $\endgroup$ – Gilles 'SO- stop being evil' Sep 21 '13 at 14:18

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