14
$\begingroup$

The idea:

I'd noticed quite a few sites on StackExchange have adopted occasional challenges (in the style of the Programming Puzzles and Code Golf site on StackExchange) as a way of getting more people in the community engaged.

Chemistry.SE often has the same problems as the other sites who've had a go at this, namely that the bulk of questions are at an incredibly basic level and don't require multiple answers from different viewpoints or perspectives (which is possibly a topic of conversation in itself).

The idea of is that it encourages people to get involved by providing an open-ended (but well defined) question that can be answered by users of different levels, hopefully encouraging both engagement and retention (and as @Ortho pointed out, anything that raises the bar scientifically is probably a good thing).

The first challenge:

The first challenge I put up was to devise some route towards the molecule below, sodium Fluvastatin.

enter image description here

Fluvastatin was chosen since it was complicated enough to need more than a cursory glance (i.e. it was challenging enough not to bore the actual chemists here), but hopefully not so complex as to be off-putting. The mixture of heterocyclic chemistry with asymmetric chemistry was also appealing (for rather selfish reasons).

Initial results:

In the space of 10 days, the post received 23 up-votes (+1 down vote), 12 favourites, and 439 views. This is far above average for a typical question posted here (excluding featured, general-interest questions).

More importantly, the question also got 5 (4 ignoring my own) good quality answers, submitted by users with a variety of 'experience' (take that as length of time on chem.se or otherwise).

Future directions:

Whilst I think this 'experiment' worked well, there should probably be some kind of discussion amongst the community here as to whether or not theres any benefit to continuing with this. One issue is with the frequency/format of any synthesis-golf type questions, as I'd be the first one to argue that if the homepage gets spammed with 5 of them a day then the whole thing becomes awfully counter-productive.

$\endgroup$
5
  • 3
    $\begingroup$ Even though it is unlikely I will ever join such a quest, I fully support the incentive to brighten up the community. I guess a good schedule would be to run it once every month and maybe think about featuring it twice, which would give it a nice two weeks on, two weeks off schedule. I am willing to throw in some rep points since the bounty dance is somewhat meandering away right now. $\endgroup$
    – Martin - マーチン Mod
    Commented Jun 1, 2017 at 7:00
  • $\begingroup$ One question is whether or not these types of challenges can apply to challenges outside of synthesis. For example, provide a mechanism for this reaction. Though one might imagine this could be abused for homework purposes. $\endgroup$
    – Zhe
    Commented Jun 1, 2017 at 14:01
  • $\begingroup$ I quite liked synthesis as there's necessarily some creativity to it — there might be two or three plausible mechanisms, but not a vast number of possible answers (it would also likely be hard to actually 'agree' on whats right $\endgroup$
    – NotEvans.
    Commented Jun 1, 2017 at 15:06
  • 3
    $\begingroup$ Synthesis is the obvious choice for such a challenge. I personally can't think of anything else that would be suitable, but I do like the idea of broadening it to include other stuff. // I am biased, but I think even once a fortnight is plausible - feature it for a week. As I mentioned last time I like things that raise the level of the site (which in turn hopefully attracts more experts). However, that's just my opinion. I think once a month is very reasonable too. $\endgroup$
    – orthocresol
    Commented Jun 1, 2017 at 22:04
  • $\begingroup$ I like this idea a lot! $\endgroup$
    – Sidharth Ghoshal
    Commented Apr 21 at 3:13

2 Answers 2

Reset to default
7
$\begingroup$

The proposal:

Based on the comments so far, I propose the following (comments very much welcome, as above).

  • We do synthesis golf once a month, starting on the second Monday (I actually agree with @Ortho that more frequently could work, but lets start slow to avoid burning out).
    • Each challenge lasts for the duration of the month.
  • The 'answer' to the 'question' will be awarded to the solution with the shortest step count (to be defined).
    • As people post their solutions, I'll edit the original challenge question with a ToC and summary of total step count/longest linear sequence.
    • To encourage participation, bounties will be started at several points throughout the challenge. At least one of these will come from the rep gained from the question (such that I'm not unfairly benefiting from doing this), and hopefully (as Martin suggested), the BountyDance could also provide additional bounties. (These to be awarded by whoever starts them for whatever reason they see fit: most creative, most precedented, prettiest figures etc etc etc).
  • I'll create to keep track of these posts, and will write the tag-Wiki to explain the general premise and point, as well as the 'rules'.
  • I'll also create (as per Martin's suggestion) an FAQ page to be linked to
$\endgroup$
1
  • 3
    $\begingroup$ I think it would be wise to create a faq post about synthesis golf in addition to the tag and the wiki entries, which can be referred to in every post. It might also be a good idea to keep a TOC there for easier finding which molecules have been done. Obviously, after creating the tag, it should be removed from the title, and I think it would be better if it is the only tag to make it stand apart. $\endgroup$
    – Martin - マーチン Mod
    Commented Jun 2, 2017 at 5:47
1
$\begingroup$

Just to play devil's advocate here...

In the current format, I don't think these questions are in line with the question/answer format of Stack Exchange and seem much more suitable for a forum site (eg Reddit). They could be closed both on 'too broad' or 'primarily opinion based.' The suggestion by @NotEvans to award the correct answer with the shortest step count attempts to address the latter, but rigorous step count is not a good judge of synthetic strategy.

If I might make a suggestion to bring them more in line with the Q/A format, perhaps the question could be presented as solving a particular challenge that must be solved to achieve the synthesis of the target compound. For example, for fluvastatin, the question could be "How can the stereochemistry for the 1,3-dihydroxy moiety be established in fluvastatin?" Or "What indole synthesis could be used toward fluvastatin?" I think phrasing questions this way would still engage the creativity of the community and provide educational benefit to the audience, while narrowing the scope of the question to a manageable level and reduce the necessity for opinion in answering.

$\endgroup$
10
  • $\begingroup$ Keep in mind, I say this as a practicing synthetic organic chemist who has made a career of thinking about these types of challenges. $\endgroup$
    – jerepierre
    Commented Jun 2, 2017 at 22:44
  • 1
    $\begingroup$ +1. I actually debated voting to close the Synthesis Golf question as too broad, but held off since I was interested to see where the idea would go. $\endgroup$
    – hBy2Py
    Commented Jun 3, 2017 at 2:19
  • $\begingroup$ That sounds interesting, and I agree that step count is a superficial way of judging the synthesis, but how would you assign the "best" answer to your proposed questions? (Or perhaps there is no need to assign a best answer...?) (nb: I agree that this does fall under "too broad", but I decided to make an exception the first time it happened, so it's good that this is being brought up in meta.) $\endgroup$
    – orthocresol
    Commented Jun 3, 2017 at 13:22
  • $\begingroup$ @jerepierre. Thanks for commenting! I agree that it was broad, as for too broad, maybe, its possibly quite subjective. The issue as I see it, is that to get a range of interesting answers the question has to be broad enough not to be answerable by a single SciFinder search, hence the creativity aspect (which I also agree is difficult to judge). "How can the stereochemistry for the 1,3-dihydroxy moiety be established in fluvastatin?", whilst an interesting question, doesn't leave much scope for a variety of answers other than directly copying from a simple web search. $\endgroup$
    – NotEvans.
    Commented Jun 4, 2017 at 11:47
  • $\begingroup$ @orthocresol The person that asks the question has the discretion to accept the answer they consider "best". Having multiple good answers to questions is not unusual on this site. The criteria for accepting one strategy over another will hinge on how it fits into the natural product suggested and how likely it is to work based on literature precedent and our chemical intuition. $\endgroup$
    – jerepierre
    Commented Jun 5, 2017 at 16:24
  • $\begingroup$ @jerepierre There are criteria defined in the question right? Fewest steps? If it wasn't defined, we could ask the asker for more info, or close as unclear. But too broad? No way! $\endgroup$
    – Pritt says Reinstate Monica
    Commented Jun 5, 2017 at 16:30
  • $\begingroup$ @NotEvans. "... doesn't leave much scope for a variety of answers other than directly copying from a simple web search." Maybe... It will still be a more complex question than those currently filling chem.SE. If it's really a challenging natural product target, then a web search is not going to give a satisfactory answer. There may be many ways to address a synthetic challenge, but considering the constraints of the natural product target forces a chemist to evaluate those options and rationalize their choice for the best. $\endgroup$
    – jerepierre
    Commented Jun 5, 2017 at 16:32
  • 1
    $\begingroup$ @PrittBalagopal If there was a question by a beginning chemist that asked "(1) How do I synthesize and indole? (2) How do I selectively prepare a trans-alkene? (3) How do I synthesize a 1,3-diol with stereoselectively?", the person would be asked to split that into three questions, and the original would be closed as too broad. People spend PhD's developing syntheses of natural products. I fail to see how such questions are not "too broad". $\endgroup$
    – jerepierre
    Commented Jun 5, 2017 at 16:36
  • $\begingroup$ @jerepierre. Sorry, what I thought you meant by your initial comment was to truncate the structure so much that all would be left was the 1,3-diol portion, which would indeed by a simple SciFinder search. To give a counter example, would you accept, for instance **suggest how one might control the stereochemistry and olefin geometry in the C1-C11 fragment of Lejimalide?" as a 'narrower' question, rather than choosing an entire drug/natural product(I think this is basically what you were suggesting). I would be totally on board with that as an acceptable level of complexity $\endgroup$
    – NotEvans.
    Commented Jun 6, 2017 at 19:09
  • $\begingroup$ I entered this comment elsewhere in meta but I'd be interested to entertain the most commercial synthesis where cryogenics and silyl protecting groups are avoided, for example. Also, some agrochemical or flavour & fragrance targets. Something that might bring Synthesis Golf more on topic is to explain an issue that necessitates exploration of the route or specify the starting material(s) to ensure parity between attempted answers $\endgroup$
    – Beerhunter
    Commented Jun 12, 2017 at 22:13

You must log in to answer this question.

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