My question https://softwarerecs.stackexchange.com/q/7035/903 has so far received 4 close votes. At the same time, the tag exists, and I see two others questions of IMHO the same kind without any close votes:

So, are dictionaries on- or off-topic?

  • I tend to answer this is "Software Recommendations" (not "Readers Digest"). If the question where e.g. for an Android App supporting offline dictionaries, that would be clearly on-topic (as the app is requested, not the content) – and here the mentioned tag would match. Hard to say. But besides, thanks for mentioning the others so we can VTC :) Honestly: those might have slipped through below the radar. Maybe edit your Q and add some functionality requirements (search etc), to make it clearly "software" and drop off the border-line?
    – Izzy Mod
    Jun 18, 2014 at 14:06
  • (+1) Good to get this nice and clear. Jun 18, 2014 at 14:33

2 Answers 2


This is closely related to the distinction between software and data sources, which is currently unresolved.

All four dictionary questions listed in this meta question 1 2 3 4 have a dual aspect: they cover both the software to access the dictionary and the dictionary content. Three of them have an offline requirement that make them heavy on the data source aspect; without an offline requirement, an answer could be “use this application which is able to query data sources using one or more standard or proprietary protocols”.

https://softwarerecs.stackexchange.com/questions/7035/matlab-numpy-scipy-matplotlib-dictionary is different because it's purely about the data source. There's no non-trivial requirement that the data be queried with a particular workflow or within a particular environment. I think this makes it firmly off-topic, but we haven't settled this.

  • Thanks, so if I had asked to a Matlab <-> NumPy/SciPy/Matplotlib dictionary offline for Windows or Kindle then it would have been on-topic? I feel that adding this kind of software requirements could be a way to "circumvent" the data source issue. I do agree that drawing a clear line between software and data sources is tricky :/ Jun 21, 2014 at 21:44
  • 2
    @FranckDernoncourt My personal feeling is that we're too tolerant of data sources, because data and processes are very different things and we aren't equipped to deal with data sources. I think that all questions here should solve a process problem, and that processes that require data sources (such as looking up a word in a dictionary) should make data sources out of scope of the question (either specify a data source and require interoperability, or request support for some common protocol). But we have no community consensus on that. Jun 21, 2014 at 22:16

I would say a off-line dictionary or Encyclopaedia (Eg Microsoft Encarta) can be on topic. Where they exceed what would be expected of a normal document.

A language dictionary is more than documentation, such a a dictionary program could provide features that would not be found in a mundane one. For example a dictionary software may have a in built audio for pronunciation. A PC Encyclopaedia could include video.

However if there is no requirement in the question, or in the answer "program" does not have such features, and is more or less a document, perhaps with indexing etc, would not software, it is a document.

For example this is the Mathematica Manual. It is a fairly complex website, tons of styling and hyperlinks and pictures, but it is basically static. Its a document. The Help functionality built into Mathematica that will let you (for example) right click on a function and load its Manual Page, is a Feature of the Mathematica Software. Even the inbuilt Help which is the same general content of

This is a borderline, for sure. In general: If it has a author rather than a developer, or if it is written rather than programmed, then it is a document, not software.

You must log in to answer this question.

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