Is our tagging structure fit for purpose? Do we agree what its purpose is?

Question

Recent discussions about tag synonyms prompted me to check whether we have an emerging consensus about the structure of our tagging system, or whether things are still in flux.

I tried to review all 262 tags (including synonyms) that were in use a few days ago, and some things stood out.

We tag consistently for date, place and technology

By which I mean

• Tagging to specify the time period/historical event to which a question relates, either by identifying the relevant century/decade or by specifying a well-known historical event such as US-Civil-War, WW1, WW2 or the Holocaust.

• Tagging to specify the place to which a question relates, to the level of country, state or county, city as appropriate.

• Tagging to identify a specific technology (e.g. software, website, dna, gedcom) or package/website (e.g. ancestry.com, family-tree-maker).

We're somewhat confused about tagging for events/attributes and sources

By which I mean:

• Tagging for an event/attribute associated with an ancestor (e.g. birth, imprisonment, military service, occupation, cause-of-death).

• Tagging for the sources supporting some aspect of research.

Some of us tag with the event/attribute etc. Some of us tag with the records associated with the event/attribute (e.g. birth-record, court-records, military-records). And some of us tag with a specific class of source (e.g. bmd-certificates, grave-marker, manifest, newspaper) rather than the more generic xxx-records, complicated by the fact that in some cases generic synonyms have been created inconsistently (e.g. birth-certificate resolves to birth-record, but bmd-certificates resolves to vital-records).

There's no structure that I can discern for tags about the research process

We've got a bunch of tags related to the research process and documenting and sharing the results, including standards/best practice guidelines that I haven't had the time to dig into deeply, but which incorporate concepts such as source, citation, proof, information, evidence, date name and numbering standards as well as research skills & techniques and elements of the research process such as locating-records, interpreting-records, record-keeping, sharing, transcriptions, palaeography, ... Some of these tags are high-level, some of them very granular, some of them overlap.

We've got some very very specialised tags

To take two examples at random, kataster and geneatheology

So what?

All of which is a long-winded way of saying: I think we need to take a step backwards, understand as a community what the purpose of our tags is and then identify a structure that will meet that purpose. Once we've got a structure in mind, we can create synonyms and/our prune tags to move us toward that structure.

Editing note: I deliberately haven't linked to any of the tags to make this more readable, and because I'd like to focus on structure rather than the value or otherwise of individual tags.

Answer 1

As I understand it, the 'standard' answer is that tags are intended to help us organise our content, by sorting questions into clear specific categories, and so to enable members of the community (and people who haven't joined us yet) searching for content to navigate quickly to questions they might want to respond to or learn from. (They are used for search engine optimisation.)

Additionally, the tag wiki allows us to present frequently-referenced information about the subject of the tag that isn't readily available elsewhere (e.g. in a dictionary), and guidance on how it should be used on this site.

Tags are not intended to provide information that isn't in the question, or to summarise the question, or highlight key aspects of the question (other than to categorise it). And a tag should stand by itself -- if a tag only makes sense when used in combination with another tag, it's a bad tag.

As an example of how I'd assess our existing set of tags against this purpose, I offer:

Our geography tags are useful to help people narrow their focus to areas they have some knowledge or curiosity about. (I'm unlikely to spend much time on questions tagged Russia, for example -- I have zero relevant expertise and no Russian ancestors that I know of).

Our technology tags let people navigate quickly to questions about specific of general subjects they're familiar with or interested in. (Don't ask me about Gramps, but I've an interest in GEDcom and DNA and expertise in Family Historian)

The utility of all our date tags is dubious. Use of historical events as a way of grouping questions makse sense: WW1, US-Civil-War and the like. But what exactly do the "decade" tags group together? Is anyone going to be an expert in a single decade, or search for questions on the 1910s or the 1870s? They mostly seem to be in use to summarise questions. I've mixed feelings about the "century" tags -- it seems they ought to be a useful grouping, if I could think of a reason why, but most of them are again being used to summarise the question.

(The analysis could be continued, but I'll save the effort until we have a consensus about purpose).

Answer 2

One important use of tags (for me) occurs when I suspect that a new question is similar to (potentially a duplicate of) something I have seen before.

IF the system is working optimally, hitting a tag under the suspect question should bring up whatever is lurking in the back of my consciousness for review. Obviously some classes of tags (dates, for example) are of very little use in this context (because they capture so many questions that are not similar to my target).

Answer 3

An important role for tags that I had largely ignored (since an initial burst of interest) is as the hook for email alerts or rss feeds.

My one alert is attached to a low-traffic tag (less than 2% of questions) so I am usually (pleasantly) surprised to receive the email message announcing something new. (For background, that tag falls in the technology group.)

Perhaps one approach to the how do we promote question might be to make better use of the tag-rss channel. If "avid" members who have a personal web presence were to embed the rss feed for the tag relevant to their interest (and hence not be out of place on their page), it might feed some additional traffic.

Answer 4

For some perspective on the issue of tagging, there's an older post on the Stack Exchange Blog: The Death of Meta Tags.

Jeff Atwood makes a lot of valid points in his post, I thought -- and then I read the comments, and was equally swayed by people arguing to keep certain tags because of the way they were used by the community.

In response to this point by ColeValleyGirl:

The utility of all our date tags is dubious. Use of historical events as a way of grouping questions makes sense: WW1, US-Civil-War and the like. But what exactly do the "decade" tags group together? Is anyone going to be an expert in a single decade, or search for questions on the 1910s or the 1870s? They mostly seem to be in use to summarise questions. I've mixed feelings about the "century" tags -- it seems they ought to be a useful grouping, if I could think of a reason why, but most of them are again being used to summarise the question.

I'll confess that I am one of the people who has employed decade tags -- it was largely reflex that came about because of Ancestry's filter that allows one to filter historical records by decade. Tagging by decade has no comparable utility here and IMHO they can be deleted.

I also have mixed feelings about the century tags. I personally enjoy working in the 20th century more and would like to be able to have questions from that era brought to my attention, but at the moment, our question volume is small. It is odd to be earning points toward tag badges for earlier centuries just because we have more of them; it feels a bit not authentic to earn a badge for eras where I don't feel as capable as the more modern one.