Improving our tagging (redux)

Question

At the risk of suggesting more work for the stalwarts who already contribute most to this site (and of boring everyone to death as I revisit a topic I raise quite often):

Is there an approach we can adopt to systematically improve our tag wikis?

While many of them have 'excerpts' and some even have 'full wiki entries' we still have a number of problems:

• tags without any guidance at all
• tags with excerpts but not full entries
• tags with poor wiki content (including an inconsistent approach to providing the same level of guidance for similar tags).
• tags that don't have a clear purpose or overlap with other tags un-necessarily

We also have a number of exemplary tag wikis that could be used as a basis for improving others; and there have been a number of efforts to rationalise overlapping/similar tags.

Why should we improve matters?

• to make it easier for people to choose the right tags (which will reduce the effort in re-tagging)
• to steer those (few) people who will read the full tag wiki into asking better questions (by providing tag-specific guidance that we can't include in the main Help), which may reduce the workload of improving questions
• to give us tag-specific guidance to which to point posters who obviously need more guidance on constructing their question.

As a by-product of reviewing all the tags, I'd also hope to see:

• more tag rationalisation where appropriate
• potentially, generation of topics either for the regular chats or (even better) for canonical questions
• a framework for constructing new tag wikis more quickly and easily

I'm not suggesting we should put so much guidance in the tag wikis that we obviate the need for (even basic) questions -- but I am suggesting we pay attention to the advice here https://stackoverflow.blog/2011/03/24/redesigned-tags-page/ especially

[the] tags page is an essential map of what your community is, and is not, about.

---

To inform the discussion (and maybe prioritise activities), here is a snapshot analysis of our tag usage, based on a rough categorisation I've allocated to the tags... where the category is blank, I couldn't pigeon-hole the tag. And bear of mind, this is about numbers of tags (not questions, which have multiple tags).

Answer 1

My suggested approach to improving our tagging and tag wikis is: (recognising that only those few of us who are motivated will participate, and we need to minimise the effort expended versus the payback for moderators, tag wiki editors and 'ordinary users'):

1. Agree the Utility/Purpose of Tags: use https://stackoverflow.blog/2011/03/24/redesigned-tags-page/ as guidance, plus the practices of similar but older sites such as history.stackexchange.com, to define tests on whether a tag should survive or not, but accommodate convincing genealogy-specific arguments for different handling where they are made and voted above removing the tag. This will need a series of questions (one per category) to reach a consensus. I suggest we allow a month per question (hope springs eternal that people will vote on meta, in spite of past experience) and then act on the majority vote in each category (if there is no majority, the status quo should be accepted).

2. For tags that don't survive the winnowing process, set up suitable tag synonyms.

3. Agree what makes a 'good' tag-wiki -- I'm torn between doing this in a chat-room or doing it on Meta. Point 1 above should hit the big requirements, leaving the rest for a chat-room? General guidance and info from older-relevant site should also feed into this.

4. For the tags that survive the initial 'winnowing', convene a working-group (in a dedicated chat-room which anyone can join if they're willing to put in the necessary effort) to identify a 'best practice' wiki to replicate for other tags in the same category. (based on 1 and 2) above. For some tags, this will exist already; for others, it will require more effort. Effort should be expended in the order of usage of tag categories that survive the winnowing, although the 'working group' may decide to hit the 'quick-wins' first. This should also give us a template (or set of templates) for similar tags so keeping up the quality in future should be less onerous.

5. Any outputs from the chat-room that aren't agreed by a majority of participants could be verified by a meta question. I'm torn on this as well -- anyone who really cares and participates on Meta will participate in the chat-room, I suspect... So will a meta-question add anything to the process?i

I'm willing to raise the necessary meta-questions and set up/nudge the chat-room, to minimise the effort on moderators. Anyone who cares about the outcome will have every opportunity to participate...

Answer 2

I am glad to see this suggestion to re-visit the advice about tagging from Stack Overflow. It is a challenge to think about how Stack Overflow-style tagging could be applied to genealogy and family history, but IMO now is the time to do it, when the number of questions is still small.

To me, the critical thing to think about is here:

Periodically asking questions like “who would ever subscribe to this tag, and why?” can reveal a lot about the nature of tagging on your site.

I see far too many cases where someone (including yours truly) has tagged a question with things included in the question that could be easily found via search.

Imagine a day when the site is so busy that users can't keep up, and they want to subscribe to their favorite tags to make sure they don't miss a good question. I'll use myself as an example since I have been one of the offenders.

How likely is it that someone's interests are so specialized that they would subscribe to a tag because they only want to see questions about the 1840s? How helpful would it be to compare all the questions tagged 1840s, just because the question happened to center on this decade? A tag for the decade will sweep up questions from that decade no matter what geographic location is involved. Is this really useful?

See how different this is from a tag 1840-census for a question about analyzing this specific record type?

For tags about chronology (in the table, "Period"), I think it is helpful to have century tags. But if we want to have others, let's think about what would be useful.