I recently asked this question on the main site. For it I created a tag , because I'm not asking the community to put their effort in developing the algorithm (it may be a hard task), but to report already done efforts in the problem.

However, I understand the tag could be misunderstood as a GimmeTehCodez allowance for "How could I program the Simpson's rule iteratively?" kind of questions.

Should we keep the tag? Are there better suggestions for its name?

Edit

The proposals so far are (please read the answers):

-> Jens (posted as comment)

-> Ajasja (but with a twist, read the answer)

-> R.M (my personal favourite, but does not cover requests for FE code, for example)

  • 1
    How about "Code-resources"? – Jens Jun 24 '12 at 21:09
  • @Jens I think your name is better than mine. Let's wait for other people's opinion – Dr. belisarius Jun 24 '12 at 21:59
  • @Jens Would you like to post it as an answer, so other people can vote on your proposal? – Dr. belisarius Jun 26 '12 at 11:36

I don't think we need a tag. My reasoning is as follows:

  • Implicitly almost every question is a code request, since the solution involves at least some lines of code :)

  • GimmeTehCodez are not highly regarded and such a tag would only encourage them

  • Nobody will probably want to browse through so it would only be useful if somebody wanted to block this kind of questions.

For your original question perhaps instead one could use , since you are asking for existing work on the topic of shadow removal.

PS: I don't think questions of the type "Has anybody implemented X in mma" are bad, as long as the poster has done at least a basic google search. Quite the opposite. If X is already implemented, then the answer saves the OP a lot of time, if not the OP can answer his/her own question after solving the problem.

  • 2
    I think your answer is a sensible one, but a "reference request" is somewhat broader than a "code request". I could regard as a reference request for example "what books should I read to understand and properly use the Wavelet functionality in Mma?" – Dr. belisarius Jun 24 '12 at 23:01
  • 1
    I agree with this answer, especially with the first point. If anyone is interested in an algorithm, we already have the tag algorithm, so it can be combined with reference-request so there is no need for any new tag to be introduced. I would vote for less many but more general tags that can be combined freely than for more tags with more specific coverages. – István Zachar Jun 25 '12 at 16:52
  • +1. I agree that in many cases some helpful references and a few suggestions or hints on a putative Mathematica implementation would be sufficient for a good answer. Nonetheless, I think R.M's suggestion of algorithm-implementations would be a better name given that this is Mma.SE, so references alone aren't really acceptable. – Oleksandr R. Jun 25 '12 at 18:43
  • On the other hand, István does make a very good point. Maybe it isn't so wise to try to come up with a specific tag for every situation after all. algorithm + reference-request is probably 90% of the way there as far as specificity goes without contributing as much to "tag bloat". – Oleksandr R. Jun 25 '12 at 19:21
  • @OleksandrR. I understand your point, but unless the question is very clear, if I see the algorithm tag, I will take for granted that the OP wants an algorithm – Dr. belisarius Jun 25 '12 at 23:40
  • @belisarius You're right "reference-request" is broader. But having thought about this some more I see another problem: "code-request" is really reminiscent of a meta tag (ie. a tag that can not stand on it's own). And the advice on meta tags is: "meta tags, tags that cannot stand alone as the only tag on a question, are not allowed" (taken from here) Both "algorithm-implementations" and "reference-request" might also verge on being meta tags. – Ajasja Jun 26 '12 at 8:06

Rather than a tag, which sounds like gimme-da-codez, I suggest using . Along with this, some possible requirements when asking such questions:

  • The problem should be presented clearly.
  • The OP should present their efforts into researching existing algorithms, such as journal articles or implementations in other languages.

Both of these together indicate that the OP has a firm grasp of what they want and what they've done. Your question had both of these and additional references (and MATLAB/C implementations) were also available in the linked dsp.SE question.

If users ask questions with this tag along the lines of "Does anyone know of Mathematica code to do X?" and nothing more, we should close it as NaRQ or NC. Not that it isn't really answerable, but the motivation to answer that question is missing.

  • I like this name best out of all those offered so far and can't come up with anything myself that I'd prefer. The other requirements are definitely consistent with belisarius's intentions, but on balance I think I'd favor Ajasja's looser interpretation: an experienced asker may be willing to take on the algorithmic development themselves if given a few helpful references and hints (see e.g. many of J.M.'s excellent answers), or alternatively an answerer might draw on their own experience and produce an implementation without directly referring to previous work, as nikie did in this case. – Oleksandr R. Jun 25 '12 at 19:14
  • I definitively like this name! – Dr. belisarius Jun 25 '12 at 23:37

How about "Code-resources"? There isn't always room in an answer to post code for larger problems, but that doesn't mean that larger questions need to go unanswered. In that case, you'd point to resources that people can look up and download elsewhere. I think we already have some questions where links to existing library archives or commercial products appear, and those could also get such a tag.

  • I like it. It's both obviously distinct from reference-request and specifically excludes "gimme teh codez". To be clear, I like all three tags: algorithm-implementations, reference-request, and your code-resources. Unless these can be synonimized using a combination of simpler tags I'd be happy to see any or all of them in use in the appropriate situations (and not interchangeably, given that they refer to distinct concepts). – Oleksandr R. Jun 27 '12 at 4:07

You must log in to answer this question.

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