4

We currently have a number of questions about interpreting a single document:

Deciphering Lunacy Asylum case notes about patients previous history

Deciphering Lunacy Asylum case notes about nurses entry?

Deciphering Lunacy Asylum case notes about administering Paraldehyde

Deciphering Lunacy Asylum case notes about escaping through a window

Deciphering Lunacy Asylum case notes about symptoms of inflammation

There are 12 questions in total.

The OP asked for advice about whether to raise one question or multiple ones.

I argued that it would be better to keep them together to maintain the context and because they were all essentially about interpreting the writing in a single document, so there's not a lot of value in splitting them out to provide more detailed answers. If it were separate documents, or the questions were about interpretation and not just reading handwriting, I might take a different view. I also worry about splitting to artificially increase question volume and/or game the reputation system.

Somebody else suggested that they should be split out... but I think the debate should be held here and not in the comments trail of a question.

Screenshot of comment trail

| |
  • 1
    I was the original OP (still have a few more extracts to sort out - step by step!). I agree it is better to turn those comments into this meta question because it was detracting from actual answers. :) I see both points of view IMHO. – Andrew Truckle Aug 15 '19 at 8:19
3

My concern in splitting up these questions is that we are treating them differently to other questions.

Essentially here we now have 5 duplicates. Yes, each of them is slightly different, but they all boil down to "how to read this document?"

If someone were to ask:

  • where is John Smith on the 1881 census?
  • where is John Smith's wife Jane on the 1881 census?
  • where is John Smith's son James on the 1881 census?
  • where is John Smith's son Henry on the 1881 census?

We would probably close these as duplicates. The answers to all the questions are more or less the same. They may or may not have been in the same household on the 1881 census, so the answer might be a different reference, but the process for finding them is the same.

I do not see why palaeography questions should be different. A good answer to such questions might include sources for improving the asker's ability to decipher further text of this kind, rather than just giving the answer.

We do have chat as well, which might be a better place for getting assistance for transcribing long documents, when there is just the odd word that is hard to read.

At the very least, we need to make the question titles unique. This will make it less likely for users to consider them all duplicates. Numbering them is not particularly helpful. I was trying to work out which ones I had already looked at and tried to read, but ended up opening the same ones several times because they have the exact same title.

| |
  • 2
    I would argue that palaeography questions are different from questions about how to search a particular record set in that they frequently require a skill-set that will be unfamiliar to many users (whereas most people probably have some experience with searching databases these days). However, I think you are right that chat may be more appropriate where the question is about a single word in a document, and I certainly agree with your point about the titles! – sempaiscuba Aug 15 '19 at 15:00
  • Perhaps you can suggest examples of how you would title them. Since I have a few more to do yet. Just an example to give me an idea. – Andrew Truckle Aug 15 '19 at 17:41
  • 2
    @sempaiscuba agree with Harry -- we ought to be answering these questions with guidance on how to solve them, not doing the work for the OP -- although I'm as guilty as many. A good answer, as Harry says, would include a pointer to a suitable resource for learning how to do your own deciphering, and maybe also a crib sheet for words frequently seen in a particular context -- causes of death and other medical conditions come up very frequently, for example. 1/2 – ColeValleyGirl♦ Aug 15 '19 at 18:14
  • 2
    I worry that with these sorts of questions we always end up doing people work for them (giving them a fish) rather than helping them learn to do it themselves (teaching them to fish). And I'm not convinced they're questions we want a huge amount of -- they're rarely useful to other people, just the OP, especially if the answer doesn't address how as well as what. – ColeValleyGirl♦ Aug 15 '19 at 18:15
  • 1
    @ColeValleyGirl Honestly, I;m not sure how much 'teaching them to fish' we can do on a question & answer site like Stack Exchange. There are some online palaeography resources, but they are all pretty superficial (for comparison Birkbeck College's Introductory course lasts a full term). If we are going to accept palaeography questions then we also need to accept that they are different from questions about how to find or search particular types of record. Yes, we can explain how we go about reading a specific text, but that in itself may not be applicable more generally. – sempaiscuba Aug 15 '19 at 18:24
  • 1
    @AndrewTruckle I’d like to see something in each title that helps distinguish them rather than a number. Perhaps “- Extract 3” could be “about administering Paraldehyde”. – PolyGeo♦ Aug 15 '19 at 20:09
  • 2
    I broadly agree with this answer. I don't think a swarm of "what's this word" questions is particularly useful, and I'd rather see them unified, especially if they are from the same document and/or written in the same hand. If people can't answer all of the question, I don't have a problem with multiple partial answers, they can be consolidated later if required. – AndyW Aug 16 '19 at 8:47
  • 2
    @Sempaiscuba, we can do a lot of 'teaching them to fish'. We can explain how to approach transcribing a document, the value of understanding what sort of document it is, whether there are any formulaic elements to that type of document at that period, what resources might help them unfamiliar words in a particular context, what common abbreviations they might encounter. Exactly the same sort of fishing lesson we want to see in the best answers to all questions -- that's what SE is about IMO. We don't have to teach an in depth course, but we can do better than just provide the transcription. – ColeValleyGirl♦ Aug 16 '19 at 12:01
  • 1
    @ColeValleyGirl I think we will have to agree to disagree on this one. By all means, direct them to an online resource on the subject (such as the one ay the UK National Archives, for example) and suggest that they come back with any specific questions they may have after that. Otherwise, the answers are likely to become far too vague and open-ended, which is absolutely not what SE is about. – sempaiscuba Aug 16 '19 at 17:41
  • @sempaiscuba The last thing I want to encourage is link-only answers -- about as much as I want to encourage 'these are the words you want' kthxbye There's a good argument for a very well-thought out canonical question on the subject... supplemented by pointing people to chat for odd words. – ColeValleyGirl♦ Aug 16 '19 at 18:35
  • 1
    @ColeValleyGirl You misunderstand. I'm not suggesting a link-only answer. I'm suggesting that we transcribe / transliterate as requested, and append a suggestion to the effect of my comment above. However, I agree that a well-constructed canonical question/answer (or perhaps more than one) on the subject would be beneficial. – sempaiscuba Aug 16 '19 at 18:55
  • @sempaiscuba genealogy.stackexchange.com/a/4215/6485 might form the basis of a canonical answer -- is was written in response to a person who asked for a transcription of a six page will. – ColeValleyGirl♦ Aug 17 '19 at 16:24
2

Personally, I think questions like these should contain only one extract per question. There are several reasons for this.

Firstly, it allows potential transcribers to attempt only those extracts that they feel comfortable with. A question with a single paragraph of text is far less intimidating than one containing half-a-dozen!

Secondly, where there are multiple possible readings of a word or abbreviation in a passage, these can all be put forward as answers (or comments) to a single question with much less risk of confusion.

Thirdly, it avoids any potential confusion should a transcriber decide to skip one or more extracts in their answer. This is probably not an issue in this case, but I have seen several wills, for example, which contain multiple extremely similar paragraphs.

Finally, limiting questions to a single extract from a document should also help avoid overly-long, and potentially unwieldy, answers (something I am often guilty of myself!)

However, we should also be clear that links to related questions containing more examples of text written by the same person. are essential.

One of the key difficulties in reading handwriting is often recognising the idiosyncrasies in the style of particular individuals. The more text we have by that individual, the more examples we are likely to have of each letter in their hand (in this case, the way the letter "p" is written offers a good example of what I am talking about).

| |
1

The question I keep circling back to is "Why are we here?" If we want the site to have lasting value, the questions and answers should be written or edited with an eye toward helping more than just the original person.

To my mind, we have three classes of questions which can be problematic, because they're tied too much to the specific case, and not likely to be helpful to others. First, the endless "what's this DNA relationship likely to be" set. Second, dating photos, and third, the "what's this word?" questions. I'd like to see this site be more than a "do my homework for me" destination -- I have plenty of "what the heck is this word" problems of my own to puzzle over without coming here to see more.

I'd rather see the extracts from the same document kept together in one question, if they are written in the same handwriting. Separating them out destroys the context and scatters the evidence over several different pages. What if we want to look at more than one fragment at a time to compare letter shapes? It's annoying and cumbersome to have multiple browser sessions open to compare fragments from different questions.

I don't want to propose a hard-and-fast rule about how many words one can ask for help with at once, but practically speaking, if there are a lot, it might be a sign that someone isn't trying hard enough to solve their own problem before coming for a second opinion. Remember how we had the question where someone plopped down the six-page will and said "help I can't read this!" with no concept of how long it would take to transcribe the entire document?

One big problem with questions of the "help, I'm stuck reading this writing" variety is how seldom the Q shows research effort. Has the person writing the Q made a letter table, consulted references or tutorials on handwriting of the period, etc. ? Very often, people don't say.

| |
  • 1
    genealogy.stackexchange.com/q/4214/6485 is the question you refer to. – ColeValleyGirl♦ Aug 17 '19 at 6:31
  • 2
    I certainly would not want a hard-n-fast rule. In my case a single answer would have been overwhelming. It is also not just about not doing your own home work, as if you are lazy. Peoples reading abilities and eyesight differ. And their experience differs. This is why we come together. Anyone who stumbles on my questions via a search will now have the benefit of my asking for help. It just might help them. Either way, I am grateful for the valued help I received. – Andrew Truckle Aug 17 '19 at 19:11
  • 1
    @AndrewTruckle Sorry, but I'm not convinced that there's any value in your questions for other people -- there's nothing about the title that will make them stand out in a search, and the questions and answers are so specific to that one document that there's nothing more widely applicable that can be learned form them. – ColeValleyGirl♦ Aug 21 '19 at 9:17
  • @ColeValleyGirl. Fair enough! – Andrew Truckle Aug 21 '19 at 14:18
  • 1
    @AndrewTruckle Here's one of the issues we need to discuss. You said: "Anyone who stumbles on my questions via a search". Questions here should be crafted so people can easily find them via a search, not by "stumbling upon them". We have an ongoing difference of opinions about how question titles should be crafted, which you've stumbled into the middle of and aren't to blame for. My two cents: I'd like to see titles which "pop" in a search and highlight the general nature of the question rather than being ultra-specific. – Jan Murphy♦ Aug 21 '19 at 16:43
  • I've given this some thought, and concluded that my answer to the question "Why are we here?" is simple: "To help people overcome hurdles in their research". If the answers can help a wider community, then that is great, but first and foremost I will be trying to answer the question in front of me. In terms of the specific issue of questions about palaeography, I'd say that yes, we could do with a few canonical answers, but very specific questions should also be on-topic. – sempaiscuba Aug 29 '19 at 15:46
  • As for the argument that "separating them out destroys the context and scatters the evidence over several different pages", I could not disagree more. When reviewing any handwriting examples, I will always either open the images in separate tabs to view them at full-size, or use an external graphics program. Whether the snippets are grouped in one question or separated out makes no practical difference there. Where it does make a difference is in how they are answered, as I described in my answer below. – sempaiscuba Aug 29 '19 at 15:54
  • 1
    I've added a note at the end to clarify that my problem with a lot of the paleography questions is that they don't show research effort. – Jan Murphy♦ Aug 29 '19 at 16:52
0

I lean towards splitting rather than lumping but by far my preference is for right-sizing, and in these particular questions I think the one extract per question works well. To me they are not too little/granular (only one word to decipher) and not too much/coarse (3 pages with 50 words to decipher would be offputting). Focusing on an extract gives answerers enough room to not only decipher the words but also to provide any context that they can supply beyond what is written there.

I am also going to focus on your:

... worry about splitting to artificially increase question volume and/or game the reputation system.

Like it or hate it there is a degree of gamification in the design of Stack Exchange sites. It is a game we can play well and lose our Beta label sooner, or play poorly and wait longer to do so. We recently Missed it by that much( and we know that the criteria for leaving Beta are changing but not yet to what. It seems like questions asked per day may no longer be one of them and certainly they seem to be going away from seeking 10 per day. Nevertheless, I always feel far more comfortable when I see our site stats at https://area51.stackexchange.com/proposals/43502/genealogy-family-history looking a little healthier, like they do today:

enter image description here

I think the recent spike in our question numbers (from about 0.8 to 2.4) per day will help rather than hinder our chances to lose our Beta label sooner rather than later. I also think that @AndrewTruckle having "right-sized" these questions has made a significant contribution to that.

| |
  • 1
    I'm not hung up on the Beta status, especially as it now seems that losing it is a 'management whim' and not an achievement, and the Area51 stats are increasingly irrelevant. – ColeValleyGirl♦ Aug 28 '19 at 5:56
  • 1
    WRT "gaming" the stats, splitting these questions increases questions per day but arguably decreases answers per question, as a "lumped" multi-segment transcription might attract several one-segment answers. So increasing QPD this way isn't a 'pure win'. But since those targets are arbitrary and now shown as meaningless, I'm not too worried about it. I'm ok with us being "the senior (and bestest!) beta site" without gaming the question system. – AndyW Aug 29 '19 at 9:28
  • 1
    @AndyW To stay amber we only need one answer per question. That stat is MUCH easier to maintain than even a low QPD. I don't share the understanding that the stats have become meaningless for losing our Beta label. It's clear they will be changed but, except for QPD, I think it will be more of a tweak than a re-write. – PolyGeo♦ Aug 29 '19 at 10:01

You must log in to answer this question.

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