Category talk:Sampling and basic statistics in R

Some preliminary thoughts

So, I'd like to avoid the painstaking process of re-structuring articles after they have already been written; Lutz and me learned about this the hard way when we tried to structure the current QGIS tutorial in a way that is not terribly confusing to the students -- after the tutorial was already filled with content; additionally, we struggled with articles in the older version of the tutorial.

This is why I gathered a few thoughts on how to structure the tutorial. The problem I figure we'll always have is the conflict between the wiki structure, which is not meant to provide any chronology among single articles. Mediawiki is not exactly the right tool for providing tutorials alongside with theoretical treatment of similar subjects. It's a bit like hammering a screw into the wall. An actual wiki is an encyclopedia, and as such aims to keep redundancy to a minimum; consider the template issuing a request to merge similar articles on wikipedia, which appears there on a regular basis.
I know, I keep saying that. In my view, what we need to do in order to avoid chaos in the future is to

• Be sure to uniquely identify articles and categories with respect to the subject. Otherwise we may run into problems in the future with articles treating a similar subject.

  • Example: there is already an article Stratified sampling covering the theory of stratified sampling. And who knows, perhaps in the future we may want provide a QGIS exercise on stratified sampling. This is why, in my preliminary scheme below, I have named articles in a "Resource assessment exercises: 'subject name'" pattern. I understand that article names may appear a bit clumsy this way. But to me, it's the only really safe way to avoid confusion in the future.

• Uniquely identifying articles also has to be seen in a versioning context. Unlike articles on theory (= 'Encyclopedia articles'), tutorial articles may depend on software development and may have to be heavily edited on as much as a yearly basis. We have experienced during the QGIS wiki update, that building a new version of a tutorial is not as simple as simply editing already existing articles -- which will render the current version useless as long as editing is in progress. A new version of a tutorial may consist of a mixture of (more or less heavily edited) old articles and completely new articles. When structuring the QGIS wiki, Lutz and me decided on the semester (Year) as the version number of the tutorial (e.g. QGIS tutorial 2013/14). As I see it, there's two options for version handling in the future:

  • Adding the semester to each article name. This would make article names really long, like "Resource assessment exercises 2014: stratified sampling" Of course, this would make the tutorial structure really robust and enable us to create a new version of the tutorial while keeping the old one fully intact.

  • Handling the version via categories only. This is what we decided to do in the QGIS tutorial. Categories and sub-categories are used like "Tags", as known from platforms like Stackexchange or Twitter. Each article can exist only once. If a new version of the tutorial is created, the respective category is added to the footer of the article. This would require careful assessment of the current version of a tutorial and deciding on what to use in the new tutorial. During the process of writing a new tutorial version, articles may end up being part of different tutorial versions if their content didn't change much. Or they may eventually be removed from the old tutorial. This may eventually render older tutorials useless, as more and more articles become missing.

  • Then what do we do? I would aim for a compromise. Basically I would vote for the second option because
  1. I think that it would force us to carefully assess the status quo of a tutorial instead of starting to hack a new script into the wiki without considering what we already have (I sort of made this mistake in the QGIS tutorial)
  2. We would avoid clumsy article names.
  3. This would be the compromise: We could still keep an intact older version of an article by moving it to an "'Old article name - Version'" article. The old tutorial would contain clumsy names, but it wouldn't hurt that much. Also, marking deprecated articles with a version number would easily enable us to completely delete old tutorial versions without deleting up-to-date articles by accident. Deleting articles is not the "Wiki-way", but I guess it's not a bad thing to keep the option.

Categories and articles

(Main category) "Sampling and basic statistics in R"

• I think we should rename it to Resource assessment exercises 2014 as Sampling and basic statistics in R is less specific and in the future may conflict with other R-based articles.
• Category page contains
  • links to articles in chronological sequence (manually entered and maintained)
  • links to sub-categories in alphabetical order (automatically generated by the wiki)

Sub-Categories & articles

• Articles should contain links to
  • Main category page (for article sequence)
  • Links to preceding and following article
  • Optionally a fancy content tree I would always keep the simple article structure on the category page intact, in order to have a fallback option if the content tree crashes because of an update (we experienced this before)
  • The actual self-learning exercises contained in the PDF

(sub-category) Introduction to resource assessment exercises (2014)

• (main article) "Introduction to resource assessment exercises"
• Contains all information from chapter 1 of the PDF

(sub-category) Resource assessment basics in R (2014)

• (article) "Resource assessment exercises: loading data"
• (article) "Resource assessment exercises: basic descriptive statistics in R"
  • Including chapters 2.2 and 2.3 from the PDF
• ... further articles on finite population correction, sample size
  determination etc. I'll make them up as I come along.

(sub-category) Response designs in resource assessment (2014)

• (article) "Resource assessment exercises: fixed area plots"
• (article) "Resource assessment exercises: nested fixed area plots"
• (article) "Resource assessment exercises: k-tree sampling"
• (article) "Resource assessment exercises: Bitterlich sampling"
• (article) "Resource assessment exercises: cluster plots"

(sub-category) Sample designs in resource assessment (2014)

• (article) "Resource assessment exercises: systematic sampling"
• (article) "Resource assessment exercises: stratified sampling"

- Levent (talk) 09:39, 28 April 2014 (CEST)

Discussion

Please discuss the points stated above here! - Levent (talk) 09:59, 28 April 2014 (CEST)

Article Titles

I would avoid trying to unequivocally identify articles by their title. Different articles with the same title are not at all a problem, if they are in different categories. If you search for a key word or title that exists multiple times, Mediawiki will give you a list of choices (with information about the category). Therefore two articles on systematic sampling might be meaningful. Lutz Fehrmann (talk)