9
$\begingroup$

As Chemistry.SE has grown, we have created many tags to help classify questions. I have been responsible for some of them, and I am also very glad to see others chip in with their expertise in proposing tags, applying them to questions, and doing up tag wikis. So I would like to begin with a sincere thank you to those who have put in hard work into these.

However, as the number of tags grows, we also need to be careful about the tags we create. Tags should generally be specific and help to identify areas or subdisciplines of chemistry that are of interest to people. At the end of the day, the primary use of tags is for users to easily search for or highlight questions which they are interested in.

I am beginning to have some concerns about our tags, especially those pertaining to organic chemistry, which tends to get a lot more exposure on our site. Ultimately, the problem is that we can only place five tags on a question – this is a system-imposed restriction. If we continue to add and apply new tags, we will run into a problem where more than five tags are potentially applicable. In fact, we have arguably already ran into this problem.

Here are some perfectly reasonable questions, and potentially applicable tags, which illustrate the point.

I read about an asymmetric Strecker synthesis of amino acids, catalysed by thiourea (Jacobsen, Nature 2009, 461, 968–970). Why does the reaction select for this particular enantiomer?

  • organic-chemistry; carbonyl-compounds; amino-acids; stereochemistry; selectivity; catalysis; organosulfur-compounds; reaction-mechanism

Why does the C-4 chlorine in 2,4-dichloropyrimidine undergo Suzuki coupling preferentially? see: Selectivity in Suzuki coupling of 2,4-dichloropyrimidine

  • organic-chemistry; heterocyclic-compounds; selectivity; transition-metal; catalysis; halides; organoboron-compounds

Is there a name for the synthesis of an indole from 2-nitrotoluene and diethyl oxalate, followed by the addition of zinc dust? (There is: Reissert synthesis)

  • organic-chemistry; carbonyl-compounds; esters; nitro-compounds; heterocyclic-compounds; aromatic-compounds; terminology; redox

Hopefully you can see where I am coming from. I do not think the way forward is entirely clear, and I do not think that one single meta post will be able to solve all the problems. I think there are some particular tags that we need to look at carefully to evaluate whether they are actually useful, and in the coming days/weeks we will hopefully see more concrete and specific proposals on meta. For now, though, I just want to bring this issue to your attention, and provide some context for the discussion that will ensue.

If you have any thoughts on the matter – maybe you have an idea of a tag which should be reworked, or maybe you disagree with me that it is a problem(!) – please let us know in the answer box.

$\endgroup$
  • 1
    $\begingroup$ I guess we need stricter rules for tagging. $\endgroup$ – Mithoron Aug 17 '17 at 21:10
  • $\begingroup$ Yes. In general I think that we need to (1) narrow the scope of certain tags (2) rework or get rid of some tags (3) be careful applying tags and make sure that they capture the essence of the question. Just because XYZ reaction contains an organometallic compound doesn't mean that all questions about XYZ should be tagged as such, unless the question is asking something about the organometallic compound itself. Obviously none of this sounds concrete but I don't want to post everything here now, probably nothing will get done if I do that. $\endgroup$ – orthocresol Aug 18 '17 at 9:06
  • 1
    $\begingroup$ You don't need five more paragraphs to convince me it's a problem. I've run into this many times, and my non-elegant solution was leaving out the broader tags in favor of more specific ones. That's not a good solution either, because as the site grows to get, say, 50 QPD, people would only filter questions by broad tags. $\endgroup$ – M.A.R. Aug 18 '17 at 9:08
  • 3
    $\begingroup$ guidelines, not rules @Mith. Sheesh $\endgroup$ – M.A.R. Aug 18 '17 at 9:09
4
$\begingroup$

To offer a different perspective, I'd argue for an approach that starts at the top (broad) and works its way down to more narrow, specific tags.

This is possibly a bit of a heavy departure from the 'slight change to the status-quo', and unlikely to be practical, but heres my 0.01GBP!

1. Every question should be tagged with some label that gives an 'overview' of what are of chemistry the question is regarding.

As an example, and broadly following the way conferences/research-departments organise themselves, most (if not all) questions could be categorised into one of 5 themes (6 if we decide to add materials, but I don't really think theres the user base right now):

enter image description here

Whilst, as Jan points out, these broad tags aren't particularly 'useful', what they do allow for is a user to quickly filter out just the questions relating to their field of interest, and I'd argue strongly for keeping the broad tags as they are (if I just have 5 minutes to spare, I just click on in the side-bar and scroll through these, I'd probably check Chem.SE a lot less if I had to wade through 5 pages of things I didn't care about just to find something relevant).

In practice, this could be implemented by requiring one of these such tags for each and every question (at least I think this is possible).

2. From here, the additional tags (4 of them) are really just needed to further narrow down what the question is pertaining to. I don't necessarily think that there is a canonical way of doing this, but I would make the point that there probably isn't a one size fits all answer to the question, and I don't think we should look for one necessarily. What a question tagged with needs is much different to what a question tagged with needs.

Since all your examples are based on organic chemistry, lets work with that:

  • I'd generally favour getting rid of tags that relate to specific functional groups,I find it messy, and as has been pointed out, you can quickly fill up the full set of tags if you've got a molecule with a few functional groups involved in reaction.
  • Preferentially, I'd prefer to tag organic chemistry questions based upon the process that is taking place. To give an example (though not a suggestion necessarily), Comprehensive Organic Synthesis II classifies all organic reactions into 7 categories: Addition to C=X bonds, C-C bond forming reactions, Additions and substitutions to C=C bonds, Heteroatom manipulation, Oxidation, Reduction.
    • This in itself may be redundant, but my point is that many places can and do manage to categorise all of organic-chemistry quite neatly– theres no good reason why we can't. Failing this we could just have tags for specific classes of mechanism: etc.
  • From there, we then narrow further. Is the question asking about a mechanistic step? Is the question asking about regioselectiviy? Is the question asking about practical skills?

Why does the C-4 chlorine in 2,4-dichloropyrimidine undergo Suzuki coupling preferentially?

  • : broadly define the topic of the question. Although it involves a metal, the Suzuki coupling itself is a synthetic-organic transformation (if the question was asking about the electronics of the metal centre and the rate of OA, then would be more appropriate).
  • or something more like : defines the actual process taking place in the question
  • : defines how the reaction is taking place
  • : defines the issue of the question - why the selectivity is observed

I read about an asymmetric Strecker synthesis of amino acids, catalysed by thiourea (Jacobsen, Nature 2009, 461, 968–970). Why does the reaction select for this particular enantiomer?

  • : broadly define the topic of the question, again, more of an organic thing as we're considering it from the point of view of the molecule being made, not whatever catalyst is being used.
  • : broadly defines what takes place during a Strecker reaction
  • : defines how the reaction is taking place
  • : defines the issue of the question - why the selectivity is observed

In contradiction to myself, disfavouring tags for functional groups, I can see some benefit to having tags to distinguish name reactions (we've already added Wittig, Aldol etc.).


I'm sure there are issue with the above, but for questions about organic chemistry, I think we can pretty well define the topic and scope of a question using just 4 tags: , , , . In theory this would then allow us 1 tag free, to use for something able to be indexed, such as .

Note also that anything like this would require someone skilled in the other 4 areas of chemistry to find a similar hierarchy of tagging that worked....but also generally I feel like trying to find one answer to all questions is doomed to fail anyway.


Like I say, just some thoughts, but, if you can find an question that breaks the proposal, feel free to comment and...make me look foolish..

$\endgroup$
  • $\begingroup$ As an aside, if we have genuine issues with the current tagging system, I can see some value in trying to set up a group chat between the mods here and the mods at the other science-based sites on StackExchange....although nothing will change for chemistry alone, if multiple sites requested some modification to the system I can see it being a lot more persuasive. $\endgroup$ – NotEvans. Aug 19 '17 at 17:17
  • $\begingroup$ I haven't fully figured out what I think the solution is, and I'll think about it carefully over the coming days. But I don't think only having 5 tags is the problem. I think it is more about the implementation on our side. Actually, I'm not a fan of functional group tags either. However, I'm also not convinced by the idea of categorising reactions by "addition to C=X". I do like the potential introduction of more specific tags, though: e.g. stuff like organocatalysis and cross-coupling. These tags would allow us to capture the real essence of the question. $\endgroup$ – orthocresol Aug 19 '17 at 17:24
  • 1
    $\begingroup$ Of course nothing I say here should be taken as an indication of what we will do - I'm just sounding out ideas here. Everything, as usual, will be painstakingly discussed on meta in typical Chem.SE fashion. $\endgroup$ – orthocresol Aug 19 '17 at 17:27
  • $\begingroup$ i.e. no decision is expected to be reached before Winterbash 2018 ;) $\endgroup$ – Jan Aug 21 '17 at 3:12
3
$\begingroup$

Tag Review 2017 is now underway.

Table of contents

I. Functional group tagging - revisited

$\endgroup$
-3
$\begingroup$

I think, we already have a possible solution to this: if there are more applicable tags to a question than we can fit in, we should:

  1. Select the tags which are most important to answer the question

  2. If after applying 1. we still have a choice of tags to make, leave out the broadest tags.

For example, the Strecker synthesis question is primarily concerned about and . Also, is probably a key one. Thereby we have three tags taken. Going by what is too broad in scope, I would leave out and . This leaves the functional group tags for reactant, product and catalyst. Since the selectivity is in the product, I would choose and since the catalyst is thiourea, in comes .

Note that applying this method can lead to equally valid results.

This is, in general, the same approach taken by other SE sites. If there is a choice between more and less broad tags, go for the less broad ones. Also, tags which are more relevant to the actual question asked should come first. The system is not perfect as it is, but categorisation is not meant to be perfect here. Other sites attempt to have near-perfect categorisation which then leads to tag lines longer or larger than the post itself — not desirable here.

$\endgroup$
  • 2
    $\begingroup$ I'm not sure (2) is such a good idea. At the least I think things should broadly be defined by their general grouping (organic, inorganic etc..). Certainly when I look at the homepage, I just filter by organic questions if I'm in a rush, to avoid things I'm unlikely to care about $\endgroup$ – NotEvans. Aug 19 '17 at 16:40
  • 1
    $\begingroup$ I'm somewhat ambivalent on the broad tags, but would lean towards keeping them on the question. (See also MAR's comment on the question.) I can see reasons for not bothering with them (if the specific tags already imply the broad tags), but I also think they do have their uses. Most people in research or academia would probably primarily identify with one branch of chemistry and if the purpose of tags are to match people to the questions they're interested in, then the broad tags are essential. $\endgroup$ – orthocresol Aug 19 '17 at 17:30
  • $\begingroup$ Oh, and as I also mentioned above, please don't take what I say as some final moderator decision - for the most part I'm just thinking out loud. $\endgroup$ – orthocresol Aug 19 '17 at 17:39
  • $\begingroup$ @orthocresol Considering organic-chemistry is such a broad topic, I have not marked it as favourite or ignored it. Instead, I have favourite tags that are a few levels lower such as reaction-mechanism. Favouriting the broad one would give way too many false positives imho (but ignoring it can be good — ignoring fails to remove many questions anyway). $\endgroup$ – Jan Aug 20 '17 at 13:37

You must log in to answer this question.

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