11

I appreciate @dionys's recent assiduousness in creating tag wikis. But recently one for the tag came through the review queue that I voted to reject. The reason was that it seemed inaccurate currently and perhaps not reflecting the focus that the community would like for the tag. The questions currently under the tag exhibit too broad a range of topics: primarily either Mathematica as a software system (system files, system options, packages, contexts, etc.) or anything called a mathematical system (usually a system of some sort of equations); plus a few oddballs. I think the lack of a good tag wiki has contributed to the incoherence. At the time I voted, I had no guide for judging how well the description captured its intended meaning. I felt it probably could have been improved, but I didn't know how.

Later I found some discussion of the issue in Do we really need that "system" tag? but it does not seem that a consensus was reached. One problem is that in my mind any question about Mathematica functionality in some way relates to the system. Defining appropriate boundaries would be difficult for me.

Can the tag be defined appropriately? What should be included and excluded? I don't think defining it would be too controversial, but in case of disagreement, it seems the standard procedure is to post competing descriptions and vote on them.

  • "post competing descriptions and vote on them" ... is the intention that one should upvote agreeable suggestions and downvote suggestions one disagrees with? – dionys Aug 16 '15 at 23:52
  • @dionys First, I don't think up/down votes on meta affect rep points, and using them as a poll has happened before. I would only downvote a proposal I thought was harmful to the site and upvote all proposals I would approve (somewhat strongly). I just wouldn't vote either way on proposals I didn't feel strongly about but did not prefer. If downvoting, one might add a reason, if it were not already in comment. Some might be willing to alter their proposal to address concerns. (There aren't really rules on this, btw.) – Michael E2 Aug 17 '15 at 0:04
  • @dionys I might add, that not much interest has been shown in this concern. I have an idea that, like before (see link in my Q), this will fade into oblivion. – Michael E2 Aug 17 '15 at 0:21
5

The following occur to me as possible interpretations of

  • mathematica systems -- e.g., non-commutitive algebras
  • Mathematica's evaluation system -- the kernels and how they work
  • Mathematica's functions for interacting with the operating system it runs on
  • Mathematica's own installation structure -- the file architecture of its many components

I think this indicates that is too vague and should be removed. I propose each of the above interpretations should get its own tag. Saying so puts me on the spot, requiring me to come up with some new tags.

I invite comments offering additional interpretations or improvements to these.

Aside: Should and be merged with one becoming a synonym?

  • Thanks for your response. As for the aside, I think evaluation is an important and complicated topic, and probably deserves its own tag. There are relatively few questions tagged with both currently, although I have not examined the questions for mis-tagging. – Michael E2 Aug 5 '15 at 14:03
1

It might help the discussion if I posted a counter-proposal (to mgoldberg's proposal to eliminate the tag). I think my command of the appropriate jargon is weak, so I welcome criticism, edits, and other constructive action.

I think one thing is clear: The tag was meant to indicate questions related to the CS notion of "system" (if that term is well-defined) and not systems of equations and other mathematical meanings.

Current tagging environment

Possibly related tags (the system-xxx tags seem quite closely related):

There are some that I felt were not close; but I also felt others might confuse them or possibly think they should be subsumed by :

Tag wiki proposition

Excerpt:

Questions about the design and structure of Mathematica as a computational system.

Description

Questions on the relationships of the component subsystems of Mathematica, including system options, distribution files, libraries, external resources (local and network), and hardware dependencies.

Discussion

Feel free to amend. The key to me, although I could be off base, is the relationship between components. Questions with a narrow focus within one component, say, only about CUDALink or only about LibraryLink, should not be tagged . Questions about the FE-Kernel relationship, incorporating external resources and so forth might be appropriate for the tag.

Most System*, but not Systems*, functions might be appropriate topics; however, most individual options in SystemOptions[] are narrowly focused on some functionality within the kernel. Some of the topics in Some Notes On Internal Implementation are similarly focused. These might not be appropriate.

I've tried to make this as attractive an option as possible. I am not yet convinced either way whether the tag should be kept or eliminated.

  • 1
    This makes sense. We could maybe also use another name for this, like mma-system or mma-system-design and maybe make system a synonym (I am also not sure what I prefer). I think it would be nice to explicitly mention the System` context as on topic for this tag. – Jacob Akkerboom Aug 7 '15 at 9:40

You must log in to answer this question.

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