Skip to main content
replaced http://mathoverflow.net/ with https://mathoverflow.net/
Source Link

DOIs are in some sense just like URLs, but with some differences. In particular, a DOI can be transformed into a URL by prepending http://dx.doi.org/.

  1. A DOI is intended to be a permanent name, that will not change meaning over time. On the other hand, a URL may become obsolete (for example, if MathOverflow was forced to move to a different host name, or changes to the software required us to change the structure of URLs).
  2. A DOI is perhaps accepted in some places, and not others. For example, because of the permanence issue, some forms of scholarly communication may object to using a URL to cite a reference, but allow a DOI.

I think it's very reasonable for MathOverflow pages to have DOIs, with two essential caveats:

  1. The URL that the DOI resolves to must be under the mathoverflow.net domain, not at a third party site (e.g. the Winnower).
  2. The URL should resolve to a page that reflects a snapshot in time of the MathOverflow question.

At present there's just no good way to achieve the second requirement, because there's no functionality for snapshots. (There are plans afoot, however.)

In any case, if someone wanted to assign DOIs of the form 10.XXXX/<prefix>/QQQQ/TTTT (where QQQQ is the question number, and TTTT is a timestamp), which for now just resolved to http://mathoverflow.net/questions/QQQQhttps://mathoverflow.net/questions/QQQQ, and could later be adjusted to resolve to http://archive.mathoverflow.net/questions/QQQQ/TTTT, that would be fabulous.

DOIs are in some sense just like URLs, but with some differences. In particular, a DOI can be transformed into a URL by prepending http://dx.doi.org/.

  1. A DOI is intended to be a permanent name, that will not change meaning over time. On the other hand, a URL may become obsolete (for example, if MathOverflow was forced to move to a different host name, or changes to the software required us to change the structure of URLs).
  2. A DOI is perhaps accepted in some places, and not others. For example, because of the permanence issue, some forms of scholarly communication may object to using a URL to cite a reference, but allow a DOI.

I think it's very reasonable for MathOverflow pages to have DOIs, with two essential caveats:

  1. The URL that the DOI resolves to must be under the mathoverflow.net domain, not at a third party site (e.g. the Winnower).
  2. The URL should resolve to a page that reflects a snapshot in time of the MathOverflow question.

At present there's just no good way to achieve the second requirement, because there's no functionality for snapshots. (There are plans afoot, however.)

In any case, if someone wanted to assign DOIs of the form 10.XXXX/<prefix>/QQQQ/TTTT (where QQQQ is the question number, and TTTT is a timestamp), which for now just resolved to http://mathoverflow.net/questions/QQQQ, and could later be adjusted to resolve to http://archive.mathoverflow.net/questions/QQQQ/TTTT, that would be fabulous.

DOIs are in some sense just like URLs, but with some differences. In particular, a DOI can be transformed into a URL by prepending http://dx.doi.org/.

  1. A DOI is intended to be a permanent name, that will not change meaning over time. On the other hand, a URL may become obsolete (for example, if MathOverflow was forced to move to a different host name, or changes to the software required us to change the structure of URLs).
  2. A DOI is perhaps accepted in some places, and not others. For example, because of the permanence issue, some forms of scholarly communication may object to using a URL to cite a reference, but allow a DOI.

I think it's very reasonable for MathOverflow pages to have DOIs, with two essential caveats:

  1. The URL that the DOI resolves to must be under the mathoverflow.net domain, not at a third party site (e.g. the Winnower).
  2. The URL should resolve to a page that reflects a snapshot in time of the MathOverflow question.

At present there's just no good way to achieve the second requirement, because there's no functionality for snapshots. (There are plans afoot, however.)

In any case, if someone wanted to assign DOIs of the form 10.XXXX/<prefix>/QQQQ/TTTT (where QQQQ is the question number, and TTTT is a timestamp), which for now just resolved to https://mathoverflow.net/questions/QQQQ, and could later be adjusted to resolve to http://archive.mathoverflow.net/questions/QQQQ/TTTT, that would be fabulous.

Source Link

DOIs are in some sense just like URLs, but with some differences. In particular, a DOI can be transformed into a URL by prepending http://dx.doi.org/.

  1. A DOI is intended to be a permanent name, that will not change meaning over time. On the other hand, a URL may become obsolete (for example, if MathOverflow was forced to move to a different host name, or changes to the software required us to change the structure of URLs).
  2. A DOI is perhaps accepted in some places, and not others. For example, because of the permanence issue, some forms of scholarly communication may object to using a URL to cite a reference, but allow a DOI.

I think it's very reasonable for MathOverflow pages to have DOIs, with two essential caveats:

  1. The URL that the DOI resolves to must be under the mathoverflow.net domain, not at a third party site (e.g. the Winnower).
  2. The URL should resolve to a page that reflects a snapshot in time of the MathOverflow question.

At present there's just no good way to achieve the second requirement, because there's no functionality for snapshots. (There are plans afoot, however.)

In any case, if someone wanted to assign DOIs of the form 10.XXXX/<prefix>/QQQQ/TTTT (where QQQQ is the question number, and TTTT is a timestamp), which for now just resolved to http://mathoverflow.net/questions/QQQQ, and could later be adjusted to resolve to http://archive.mathoverflow.net/questions/QQQQ/TTTT, that would be fabulous.