This problem is a poll (and a little bit of "white hat hacktivism" as well) regarding Cody users sentiment about the use of regexp to reduce the size/score of Cody solutions.
Your function should return one the following strings, reflecting your opinion on the question: would you like the regexp (?@cmd) functionality to be banned in Cody?
'no!' 'no' 'yes' 'yes!'
note1: the score of your solution will be 10 for 'no!' 20 for 'no', 30 for 'yes' and 40 for 'yes!' (so that a quick glance at the solution scores should give you an idea of the poll responses). If you want to solve this problem but do not have an opinion on this question -or if you are really interested in getting a low score-, just return an empty string (and you will get a 0 score).
note2: the question regards only the use of dynamic expressions of the form (?@cmd), it does not ask whether you would like regexp to be banned entirely)
regexp '' '(?@a=1)'
is equivalent to
Hiding the functionality inside an eval or a regexp string allows Cody players to achieve a very low score regardless of the true complexity of the evaluated syntax. While the function eval is banned in Cody, the function regexp is too useful to be banned entirely, so it is not uncommon to see the leading solution to many Cody problems using this regexp trick.
There are, nevertheless, relatively simple ways to ban the use of dynamic expressions of the form (?@cmd), while still allowing the rest of the functionality of regexp. This poll asks your opinion about having such a ban in Cody.
This problem testsuite includes a little hack that temporally disallows the use of dynamic expressions of the form (?@cmd) in all of Cody (not just this problem), while allowing the rest of regexp functionality to work normally. And your vote decides whether to implement or not this temporal fix:
There are typically around 5 Matlab servers running Cody solutions, and these are reset every hour approximately, so each "hack" will have a relatively short lifespan, and it will also only apply to a small proportion of all Cody solutions (but note that the effect is cumulative, answering ~5 'Yes!' questions is going to disallow the use of (?@cmd) dynamic expressions for all solutions sent to Cody for ~1 hour; answering ~5 'No!' questions is going to revert to the normal behavior).
This allows Cody users to not only express their opinions but also assert some form of 'decentralized' control over Cody behavior (which might, or might not be after all, a good idea). For now I am planning to leave this behavior only for a short time as a small 'hacktivism campaign', and revert this problem to being a pure poll (no longer affecting solutions to other Cody problems) after that. Feel free to take a look at the testsuite to explore how this is implemented (and please use responsibly), and/or leave your thoughts in the comments.