# Opinion on puzzle-series tags

We have had several discussions on puzzle-series and common names in titles, the most recent one here.

Now, I would like to propose that we introduce series-xxx tags for puzzle series. The main reason, why puzzle-series have "common names" is to allow easier identification and search, but I think this could be achieved much easier and better with tags

For example, I really think that the "Ernie" themed puzzles deserve a "series-Ernie" tag to have them easily listed next to each other and even do things like RSS feeds or news-alerts should a new one be posted.

I think such series-xxx tags should be used sparingly and only if at least 3 puzzles really belong together (and not only by the same author), but if used with common sense, I can only see them doing good. They all should be series-xxx with xxx being series-identifying so that they can be easily spotted.

Opinions on that?

Since meta-tags are bad but searchability is good, any writer who wants can put in a searchable line after each puzzle in the series. For example,

This is part of the Ernie series of puzzles.

with a link to the search result.

• Good suggestion, I like that. Jan 12 '15 at 20:05

No, that would be a meta tag. It describes certain aspects of the puzzle, not the content of the puzzle itself.

It's also worthless in categorizing and promoting the discovery of questions, which is the entire point of tags in the first place. Let's say there's a puzzle called a "Steve puzzle" in which the protagonist Steve has to solve various puzzles. What is the purpose of having a tag which, to a new user, describes exactly nothing about the post's content, rather than a much more useful tag? How could one possibly be an expert in these "" puzzles without also being an expert in , and vice versa? These types of tags would add no practical value to the site.

As the response to Doorknob's answer would be too big for a comment, I'm putting it here.

Doorknob, I simply don't think I fully agree with your statements.

Let's first deal with your example question:

"...Let's say there's a puzzle called a "Steve puzzle" in which the protagonist Steve has to solve various puzzles. What is the purpose of having a tag which, to a new user, describes exactly nothing about the post's content, rather than a much more useful tag?"...

1. First, I would propose that the appropriated question tagging would be and . It should of course be possible to find the puzzle when looking for logic-grid puzzles, but it should also be possible to find all puzzles of this ingenious Steve-series which I know is my level of hardness, has unique but fair twists and is just what I'm looking for.

2. Have you noticed that Puzzle SE is different from say Earth Science SE?
(No, this is not meant as a provocative statement, but to highlight the argument below.)

The questions rarely have an object they are about and which is tagged. Instead, most of them are challenges which are of a certain category, which is tagged. Hence tags like are a sort of meta-tag already! They describe a primary property of the question not a property of the object the question is about.
(The currently most used tag is used in questions which are cryptic-crosswords and not if they ask about cryptic-crosswords.)

The proposed tag would also describe a property of the question, namely a 2nd (3rd, 4th...) 'category' it belongs to. The difference being, that this 'category' is not generally known outside the site. Hence the mandatory series- in front of it, to make it such. Granted, series tags are useless outside the StackExchange, i.e. it is not a tag one would Google for, but that is not the purpose of tags, is it?

3. "...These types of tags would add no practical value to the site."

This is the statement I most severely disagree with. The whole purpose of this question here is , that it came about because it (might) solve an existing problem - which I would define as adding value.

So lets put prejudices against "meta"-tags aside for a moment and examine what could be done with the proposed tags. Afterwards examine the "dangers" associated with them compared to the other tags we already use.

We undoubtedly have content on this site which by one or the other means should be grouped together for easier reference. I think the best serving example (also because of the outstanding quality) are the Penguino's Ernie-themed puzzles. Looking at them, is there any doubt that they should be grouped together for easier reference/finding?

Right now, the grouping in done by a common "Ernie and the..." title. I think it has already been agreed on that grouping-by-title is problematic. "Ernie and" might be a rather unique catch-phrase, but searching by title-parts is surely not a good thing in general. It is too easy to get results which do not belong to a series, and if multiple questions use identical titles (except for numbers), it becomes even worse.

Alternatively, one could only search author, but this is a very poor substitute also. No guarantee that all puzzles of a series are done by one author, or that one author only has questions belonging to a series.

So, as an argument against a tag, how else could we solve this issue?

If there is an elegant solution, I'm the first to abolish the idea of series-tagging. But if there is no other elegant solution, I think this tagging would add value and it rather remains a question of how should it be done than if it should be done.

Now were are the problems? We do not want to promote meta-tagging. I think, this is what it boils down to. Doorknob's reaction is very much understandable if one considers that he's a moderator and therefore in constant battle with bad meta-tagging. So this series-tagging should be introduced in a way which minimizes/eliminates bad side-effects.

We clearly do not want:

• Authors tagging all their puzzles as
• Series-tags as disguise for meta-(meta-)tags like
• Hundredths of 'mini-series' of 2 or 3 posts.
• ... ?

I think we can 'battle' this from the start by enforcing some simple rules for tags:

• Series tags are created & edited in after a series has been identified by the community.
As such, a series needs to have at least 3 existing and approved of (positively scored) member questions.

• There has to be a real communality within the series on top of a common puzzle category and/or authorship.

• No other tagging could achieve the same grouping effect.

• General tag requirements (Wiki , scrutiny by community etc.) apply

• I think there is doubt that questions should be grouped in series. If one question builds on an earlier one then it should link back to the earlier one in its text. If the questions are independent and linked only by the fictional setting (~= the text which could be stripped away without losing the actual puzzle), I don't see any benefit to grouping them. This is a puzzle site: the relationships between puzzles which should matter to the site are those which are inherent to the puzzles: i.e. the mechanics. Jan 12 '15 at 16:18
• Clear statement, thanks. (I differ in my opinion though.) Jan 12 '15 at 16:21
• In particular, your first point (which I summarise as "I want to find puzzles from the same series because I know they're the kind of puzzle I want to solve") only makes sense if the link between the puzzles in the series is mechanical, and in that case a tag which describes the mechanic is both beneficial and sufficient. Jan 12 '15 at 16:22
• @PeterTaylor Not true. I like f.e. the Ernies because I know they are well made and even if complex, worth solving. (They are not all of the same mechanics) Or: I might like some riddles while I don't like all riddles. Jan 12 '15 at 16:23
• But how do you go about guaranteeing that anything tagged with series-ernie has the hard-to-define properties you like? Isn't that really a case of knowing that you like Penguino's puzzles? It may be the case that there's "no guarantee that all puzzles of a series are done by one author", but I don't think you've made an argument to support the idea that in that case there's a genuine non-mechanical commonality that's worth tagging. Do you have any examples of series by multiple authors which you think need a non-mechanical tag? Jan 12 '15 at 17:00