Family Historian User Group

If I have for example, an image of an extract from Scotlands people of a birth entry and a birth certificate, what is the optimal way to handle it.

I'm thinking that a single source should be created and the two media items should be attached to the single source.

The alternative presumably is to create 2 sources each with 1 media image.

Is there a "best" way to handle this situation?

Thanks

Colin

If you have two separate documents, you have two separate sources, so create two sources each with one media item attached, and cite both sources for the event.

OK thanks for that prompt help.

Colin

There may not be a "best" way for everybody, but use whatever is "best" for you.
However, here are some points to ponder.

If the extract from Scotlands people and the Birth Certificate both illustrate the SAME details, then you may as well attach the two images to one Birth Source record, although the Text From Source transcript may be more of a challenge.

But if they are distinctly different documents with significantly different information, then create two Source records.

Remember that the Source record will typically have several Citations not just for the Birth Event but also for the parent Names and their Residence and Occupation facts, etc...

Those same concepts apply to all Source documents.

Thanks Mike for that additional help.

This is what I love (and hate) about genealogy and FH. I do understand there is no "best" way and that what works for an individual is the best in practice. But there does seem to be a lot of variation as to what best should be.

The problem is that without detailed knowledge, what works for me today may be building up problems for me in the future, maybe months and years ahead when I want to take advantage of more detailed aspects of the software.

So seeking an optimum way is always my aim so that I learn from the mistakes others have made.

I'm using AS to input most docs, so multiple Citations shd be well catered for, and I've been using Automatic Source Citations for the others.

In this case I thought a single Birth Source would be better, the extract from Scotlands people contains quite a bit more info than the shortform certificate. But what data is present is essentially identical (as it should be), although the date/ registrar data is (usually/always?) different. I do not capture Text From Source. (That is a whole different question as to whether Text from Source should be captured).

The same appears to me to apply equally to other BMD documents.

Without being too technical or fussy, it seems to me that both documents (by now of course digital images) are in relation to the same event ie the birth, marriage death etc and belong together.

Colin

Without being too technical or fussy, it seems to me that both documents (by now of course digital images) are in relation to the same event ie the birth, marriage death etc and belong together.

Which is why you cite both sources for the event and link them that way.

Combining both documents into a single source might seem simple and straightforward but... for a more general case:

1. Do two documents both provide exactly the same information in exactly the same format? If not, then you (or somebody you share your research with) might want to go back and check exactly where a particular 'claim' came from; and also, as Mike Tate say, you might want to cite one document but not both in support of other 'claims'' (facts).
2. Are they both equally reliable? Is one a transcription or abstract or index, in which case it's prone to errors being introduced when it was created. Certificates aren't immune to error either, (some of which can be introduced by you reading bad handwriting!) but separating the two sources allows you to consider their drawbacks separately.
3. Are you going to be sharing your research with other (but without sharing the media)? If so, and you combine the two sources, how will somebody else know what document to consult if they want to check on what you've said? (As they should!)

It's not often I disagree with Mike, and of course you may choose to do things the way that suits you best, but in this case I disagree categorically with Mike: genealogy best practice says that 1 document = 1 source. You will find life much easier down the track if you start off that way, whereas disentangling 'mingled sources' in the future might be a nightmare.

Yes, as I said, if the details in the two images are essentially the same and thus involve the same Citations then there is no significant problem in combining them in one Source record. But as Helen says there are issues relating to Reference id, and Reliability.

The purists might choose to use two separate Source records because they are two different documents with two different Repositories, two different Reference Id, and different Assessment Reliability, especially if it is likely you want to pass on your research to others.

However, I don't perceive any significant future problems whichever approach you take.
If problems did arise it is not a major exercise to split the few combined Sources into two.

Text From Source is useful if you wish to include those details in Reports, because images are often difficult to read.
It also comes into its own when the image is hand written and almost illegible.
Finally, you can use tools such as Queries and Plugins to search Text From Source whereas that is not possible with an image unless you use OCR.

Thanks to both of you. As it happens, I only have a small number (about 8) of these cases at the moment, and I had created 2 entries for each. I was beginning to think that combining them was the right way to go.

But as I said earlier, I'd rather try and avoid issues in the future, so if keeping them separate is (potentially) going to cause less problems then I think I'll just leave them as they are.

Text From Source is useful if you wish to include those details in Reports, because images are often difficult to read.

Yes a lot of old records are not easy to read, and ones that are, may not be so easy to read when reduced down to report size. I've never done a report yet (I want my data - or the bulk of it in place and verified first), but this is where I need to learn from those that have been before. As I'm using AS for the bulk of my sources, I already have many Text From Source entries. It will not be a huge task to enter the small number of others as a separate exercise. I will do so.

Thanks for the input.

Keeping your Sources separate is certainly the safest future proof option.

I would strongly recommend that you investigate whatever styles of Reports and Diagrams you envisage using BEFORE entering all your data. The reason being that what appears in those outputs is governed by how your data is organised.
For example, if you are keen on using Fact Witnesses, then they only appear in Narrative Reports and nowhere else, so if keen on using Individual Summary Reports you would be wise to avoid using Fact Witnesses too mich

Similar arguments apply to other tools such as Plugins and Queries that help identify Source records you have not yet found.

So do consider what administrative tools you might want to use, and adapt your data entry to maximise their features.
Otherwise, you might find yourself re-entering data again to suit the tool-set you want to use.

Regarding Text From Source, do NOT rely on Ancestral Sources automatically getting the wording correct.
In the AS help it advises that you edit the Auto-Text box so that it mimics the document more closely.
A popular example is an Occupation you probably record in AS/FH as Agricultural Labourer but in documents will often be abbreviated to Ag. Lab. which is what should appear in Text From Source. However it appears in documents, the Occupation field should consistently use Agricultural Labourer so tools can search and analyse that information more easily.

tatewise wrote:... if the details in the two images are essentially the same and thus involve the same Citations then there is no significant problem in combining them in one Source record. ...

This is something I have been trying to get my mind around as part of an ongoing verification, validation and evidencing exercise that I am doing on a number of long-standing distinct family trees.

In terms of sources, I am a "lumper", so I should have one source for say "1911 England and Wales Census".

I say "should" because I have until recently duplicated some sources depending on what I now think (after sitting down and trying to think clearly!) is a mistaken view of "repository":

"1911 England and Wales Census (Find My Past)"
"1911 England and Wales Census (Ancestry)"
Because the online images can vary in readability and the transcripts can also vary

I used to quote the image URL in the "Where within Source" (citation) - which was at least consistent, but have moved to quoting the Archive Reference such as (for 1911 E&W Census)
"RG14; PN: 8578; RG: 78; PN: 444;" - which irrespective of your online provider gets you back to the original document. The logic that follows from this is that the "Repository" is not "Find My Past" or "Ancestry" but "The National Archives, Kew, London" - where the original document "reposits"

I am now wondering if as a true Lumper I should have a single "1911 England and Wales Census", and distinguish between online providers in the citation ("Where within Source"):
"RG14; PN: 8578; RG: 78; PN: 444; (FMP)" and
"RG14; PN: 8578; RG: 78; PN: 444; (Anc)".
"Provider" of course is a field not recognised within GEDCOM or FH!

I can then save the media against the citation (not against the fact) - which means the media is properly cited.

On occasions I might cite a source twice; once via the Ancestry "provider" and once via the Find My Past "provider" - where the images may show different detail.

This feels cleaner, but cleaning up my family trees is a daunting task!

Generally speaking I'm not a lumper; however, I make exceptions for things like the GRO indices, but I still maintain different sources for different... well, sources (e.g. FreeBMD, Ancestry, the GRO themselves). Very often, the indexing is different -- you can find an entry at FreeBMD and not Findmypast, or vice versa. Plus of course now the GRO make information available that may not have been indexed elsewhere (mother's maiden name, age at death). So I always indicate which index I found a reference in -- how you do it is up to you (separate source -- which is what I do, where within source, citation note) depending on what makes sense for your way of working.

I don't lump census sources, but agree that -- if you do -- it's essential to indicate where you saw the record. Image quality varies; transcriptions vary; and you may not find the same pages at every website -- either because of scanning errors or because (for example) some providers have access to what was previously believed lost -- don't quote me, but I think FMP has some 'lost census' pages from Cheshire for example. If you're going to cite (e.g.) the same census records from both Ancestry and FMP you might find it easiest to have a separate source for each provider, so you can see what's going on in your source tab -- citing the same source twice might be confusing.

Yes, it all depends on how 'thorough' you want to be.

Like Helen, I cannot come to terms with the Method 2 'lumper' approach for Census records.
It involves repeating the same Citation fields and image links for every household member Census Event.
It gets worse if you want Text From Source transcripts (for all the reasons I gave earlier).
I know you can copy & paste and/or use Autocitation or AS, but all that replication annoys me, and if it ever needs correcting you must remember to update all copies.

davidf wrote: ... "Provider" of course is a field not recognised within GEDCOM or FH! ...

No, but "provider" is the electronic version of the publication data, I suggest.

davidf wrote: ... This feels cleaner, but cleaning up my family trees is a daunting task!

I think that you need to be very wary of having a clean look where you have either lost or hidden information so deep that it's difficult to find - or print - again. A while ago I was contemplating the Barthomley parish records - I have used them on microfilm (images of the real thing), on-line (images of the real thing again) and in a printed version (transcriptions of the original thing). The microfilms come from two places, as well - Crewe Library and Chester Record Office. As much as it might seem easier to have one source-record for "Barthomley Parish Registers", it's hiding information one layer further down than I think is helpful. If I were to lump them, I'd have (at least) three source records:

• "Barthomley Parish Registers (LDS microfilm)"
• "Barthomley Parish Registers (FMP)"
• "Barthomley Parish Registers (Audley FHS transcript)"

I say at least, because I might want to have "Barthomley Parish Registers (LDS microfilm, Crewe)" and "Barthomley Parish Registers (LDS microfilm, Chester)". These multiple source-records might sound over-pedantic but it does put the stuff that I want to see out in the open. And the other thing it does is simplify the thinking process of entering new data. No, really. The definition of source is "What did you use to get the data?" And clearly I don't actually use the "Barthomley Parish Registers" themselves - I used the microfilm at Crewe, or the FMP web-site or... It might seem that there's no real difference in this case, but there are times when I catch myself trying to think what the underlying source is. Particularly with respect to parish registers and transcripts. Being strict over the question, "But what did I actually, really, genuinely use?" simplifies the digging for the underlying source by stopping me digging!

In the end, you can rest assured that you will think hard, create a scheme, test it, use it and then (inevitably) alter your scheme. And then ... repeat!

Adrian,

I'll see your Barthelmey Parish Registers and raise you (all for Stalbridge Parish in Dorset -- and all cited in X's family tree):

• * Parish Registers (originals when they were in the custody of the parish, and later in the custody of Dorset History Centre -- who may still produce them for view (not sure))
  * Selected transcripts of Parish Registers made in pencil by spotty 13-year old X from originals in the Vicar's living room in 1959 (he still has the transcripts but not the spots)
  * Transcripts on index cards made by GJ of originals circa early 60s
  * Spreadsheet privately circulated, produced from GJ index card some time in the 80s
  * Database privately circulated , produced from spreadsheet "corrected by person or persons unknown" sometime in the 90s
  * Bishops Transcripts (originals) at Swindon and Wiltshire Family History Centre
  * Privately held un-indexed un-transcribed photographs of Bishops Transcripts (same non-spotty individual -- but my fault they're not yet transcribed)
  * Databases with images of Dorset Parish registers at Ancestry.co.uk ) based on documents provided by Dorset History Centre, including Stalbridge Parish Registers
  * FamilySearch have microfilms or PRs and BTs I believe and have indexed the PRS....

I'm know myself or have been told that, for example:

* The originals in 1959 were much more legible than the images on Ancestry (or indeed the originals more recently)
* Ancestry have missed some pages
* There are entries in the Bishops Transcripts that don't appear in the Registers and vice versa (Have verified this for myself)
* .. There will be other differences between the record sets (the stuff based on the GJ index cards is particularly problematic)

There's no way if this was my family history that I'd combine any of these sources when constructing citations to support a fact.

AdrianBruce wrote:

davidf wrote: ... "Provider" of course is a field not recognised within GEDCOM or FH! ...

No, but "provider" is the electronic version of the publication data, I suggest.

Like Adrian, I regard the "provider" (the website where I found an image of a document) as the publisher, as it is literally making the image public. So I enter the name of the website in the Publication Information field of the Source Record. (And I regard the Repository as the building where the original document is physically located). But of course that only works if you are a 'Splitter'. There is no obvious place to record the provider/publisher if it has to be recorded in the citation.

Helen,

We need to be careful about what we mean by a "source"! In respect of Parish Registers:

The Bishop's Transcript is a different source from the Parish Register - often with transcription errors but sometimes with added "colour". (As I found reading through the BTs for some North Cumberland Baptismal Records, where the person preparing the transcript had added his (I presume his) opinion of some of the mothers presenting their children for baptism - presumably he felt the bishop should know!)

The Parish Register is a Primary Source, whilst surely the BT is a secondary one. Initially I may only have access to a BT, but would seek to corroborate the BT by looking at the Parish Registers themselves.

I am not sure that I buy into Adrian's The definition of source is "What did you use to get the data?". The question is certainly valid and needs recording - which is partly why I am tempted to start including (FMP), (Anc), (FS) etc. as part of the citation.

A source is surely "where did the quoted "information" come from?"
"23rd Aug, Josie, dau of Sarah Smith" may be from my reading of the parish register
"28th Aug, Julie, bastard dau of Sarah Smith (low born mother of many bastards)" may be from my reading of the bishop's transcripts.
They are two difference sources - even though we know (once we have compared both) that most of the later comes from ("is sourced from") the former. If we are using the BT, unless we have checked against the former we don't know; and therefore the BT is the source of the "information" being quoted - which may be subject to transcription "variations"

I would treat any transcript as a separate source. So Scottish Census record in Ancestry (which last time I looked were only transcripts - prepared by Ancestry?) are a distinct secondary source. The same would be true of Ancestry's English Census records unless you check the transcript against the image. Sometime in a census transcription it will say under Place of Birth just "Somerset" - but the image may also give a town name (but the US based transcribers could not make out what it said - "Williton" just reads as a whole lot of up and down strokes!); on the other hand the transcript may add "England" even though that is not on the image!

There is then the further consideration coming from Adrian's definition - If I was looking at an image of the original parish register was I looking at:

1. the actual parish register dug out of the strong room at Cumbria Archives
2. the Cumbria Archives microfilm of those registers
3. the LDS microfilm of those registers

This is worth recording if you later end up questioning the reading of an image. It's not that the actual registers will necessarily be more reliable than an image (sometimes the later can have enhanced contrast), but for someone to be able to say to you "have a look at the FMP image and I think you may change your mind", is useful - if you have recorded that you were looking at the Ancestry image.

In addition I strongly suspect that original BTs may be labelled by some online providers as "Parish Records" (The implication being that they are the Parish Registers not the Bishops (Diocesan) Transcripts; hopefully in error - rather than as a means to claim higher coverage of "Parish Records"). By recording the "provider" of the Images, it is later possible to take account of the realisation that what you thought were parish registers were actually bishops transcripts and therefore subject to transcription errors.

With all software (other than that which we have written for ourselves) we always run into the problem that certain data entities are either not recognised at all or not realised in the way that we want. Hence we compromise and try to accommodate our compromises within what is available. If you are a "lumper" you may have cited text duplicated; if you are a "splitter" you will have sources duplicated (in as much as each page in a source is effectively treated as a separate source).

The Bishop's Transcript is a different source from the Parish Register

Agreed, which is why I referred to records for the parish rather than the Parish Registers. I have BTs that record burials that never appear in the PRs:

From the Llanfair nant y gof Bishops Transcripts, 31 December 1824 to 31 December 1825
Burials
Not any Buried by me, but two children of Joseph Francis of Cwmwyntell & Martha his Wife were buried by those who attended the funerals before my arrival at the Church Gate whose names were John Francis aged 20 and Anne Francis aged 24. The mother of them is a Baptist & it’s customary in this neighbourhood (that the parents of those) to name their children themselves, they were not baptized in the Chapel of Llanvair as there is no account of them in its Registers & I have reason to believe & think that they were not baptized by any licensed Preacher of their own Sect.
This Circumstance happened on the 9th of October 1825 a day of as violent Rain as happened before or since, the Body of the Chapel was then unroof’d to be put in repair, & they – the Congregation assembled, pleaded as an excuse, that they had no place to shelter themselves from the Inclemency of the Weather.
William Edwards Rector of Letterston with the Chapel of Llanvair annexed.

Re

The definition of source is "What did you use to get the data?"

, that is very different from

"where did the quoted "information" come from?"

(prone to transmission errors,) so I'm with Adrian on this -- send people to the place I looked at, not the place where the 'original' data lies. And record exactly what you looked at -- don't assume the stuff you didn't consult says the same thing!

This discussion has grown a bit.

It does strike me that we (and here I mean new/ less experienced users) all need a single document with clear definitions of what is a repository, what is a source, a citation etc etc. Many of the help files and the Calico book contain bits of this, but nowhere that I've found is a simple clear definition.

What is obvious is that there are some pretty large variations as to what different people would consider those answers to be.

However, I'm a great believer that I (and many many other users) will probably never use some of the advanced techniques that appear here. I regard myself as a run of the mill user who simply wants a good quality family tree with good graphics and (in due course) reports. The endless degrees of personalisation that are possible, will probably not be of huge interest to me. I don't want a web site, I want to share my tree with my family, very few of them have a serious interest in the subject. They want base information well presented.

So definitions that suit the 80% of "everyday users" are what interest me. I do not have the time, or the interest or the funds to research to the degree that many of you appear to do. I fully accept that this means I will have gaps and errors in my database/tree. But I'm willing to bet that with all the extra detail you have, you also have errors and gaps. Perhaps less of them, but maybe not.

It does strike me that we (and here I mean new/ less experienced users) all need a single document with clear definitions of what is a repository, what is a source, a citation etc etc

Well, yes, but most of the concepts you mention are not specific to FH, but are generic genealogical concepts. So, you perhaps ought to be looking for best practice elsewhere. This site is about how to use a particular product, not about how to 'best do genealogy'.

What you need (perhaps) is a document pointing you towards info and training (like https://www.futurelearn.com/courses/genealogy) about best practices. And then you can ask about how you should use this product to support the way you've decided to work.

And while I'm on the topic, you'll get a lot of advice here about what you can do with the product but (no names, no pack drill) not all of it accords with best genealogical practice. Which may not (and should not) bother you if you not intending to share with experienced genealogists, or don't care about what others think about your research, or just want to have fun (obligatory Cyndi Lauper video https://www.youtube.com/watch?v=PIb6AZdTr-A)

However, there are people here who will share the practical pitfalls or otherwise of doing things a particular way and it's worth listening to them.

ColeValleyGirl wrote:
Re

The definition of source is "What did you use to get the data?"

, that is very different from

"where did the quoted "information" come from?"

(prone to transmission errors,) so I'm with Adrian on this -- send people to the place I looked at, not the place where the 'original' data lies. And record exactly what you looked at -- don't assume the stuff you didn't consult says the same thing!

I think we are actually almost together here, but possibly saying it in different ways!

If the "information" you are entering comes from "transcript by spotty 13 yo", that is the source - "Spotty 13yo's transcript" and the "information" will will be subject to his transcription (and your!) errors.

If the "information" comes from an image of the original document - and that is the answer to Adrian's question: "an image of the original document"*, I think we are then debating whether the image counts as the primary record "the original document", or a secondary record "an image of ...", in the same way that "a transcript of ..." is a secondary document.

* rather than "Ancestry" - which is how I initially interpreted a possible answer to "What did you use"

If I have looked at an image of the Parish Register (and I am happy that is what I have looked at - and recorded how I accessed it), I would rather researchers trying to validate my research went "to the Parish Register" (however they access (images of) "it") rather than necessarily repeat my access method (and consequently be prone to the same errors that I had been subjected to); they might find something different - which would be interesting even if my research consequently failed to validate!

An analogy (that possibly undermines my position):
I might cite as a source: "Miles, W; 1938; Official History of the Great War, Military Operations, France & Belgium 1916, Volume 2, Committee of Imperial Defence"
I am actually reading a fascimile reprint of the original published in 1938 - and there is a note in the flypages "In reprinting in facsimile from the original, any imperfections are inevitably reproduced ...";
Have I stated my source? Captain Miles would recognise his work, even though it is now in a softback binding. Or am I being "intellectually dishonest" in claiming that I have got information from his book?

Or is that getting too close to "angels on a pin"?

ColinMc wrote:I do not have the time, or the interest or the funds to research to the degree that many of you appear to do. I fully accept that this means I will have gaps and errors in my database/tree. But I'm willing to bet that with all the extra detail you have, you also have errors and gaps. Perhaps less of them, but maybe not.

Probably more not less - certainly a greater awareness of the fragility of previous research.

However I do know the feeling! But what I am finding - as the amount of data that I am seeking to manage expands (now well beyond "one headfull") and the capabilities of the software increases - that I regret earlier sloppiness/ignorance.

I was sloppy with places & addresses, which used not to matter that much as I could recall other records that were in the same place. Now that I have thousands of records and finding that specific knowledge of places and address is useful (and the mapping functionality is available) I wish all my places were in standardised formats so that I could use software functionality to answer questions like "who else was in Dunster, Somerset, England"?

When entering an address, if it starts to auto-fill I am alerted to the fact that that address is already used - which is useful for spotting neighbours or even duplicates. But to work you need to enter addresses in a standardised manner.

When I started hard-drive capacity was limited and expensive; it is now considerably cheaper and the software has improved handling of media. This means that downloading or scanning images of sources is now my default workflow. I am now trying to retro-fit media - particularly census images - because I have realised that neighbours can turn out (later) to be interesting as in-laws, executors, business partners etc. I want to manage my media in a way that is both compatible with how I manage my data and with the sort of questions that I might want to ask. It is therefore helpful to debate with others different ways of doing this - sometimes fairly rigorously to drive out nuances that I have missed.

But you are right that, it is horses for courses; for most of us it is a hobby after all, so why should we not do it as we wish. I am acutely aware that I do not use reports (which will horrify some people) and probably my current workflow will not support producing "good reports". I live with that.

I take what you are saying, But I couldn't care less about best genealogical practice. I made my choice to work with FH, and that's what I want to learn to use to the best of my ability. If some of that goes against best genealogical practice, I have no problems with that.

So I care what terms mean within the program I am using. If it doesn't fit with the international definitions of ged files again I have no issues, so my questions do relate to FH, not to genealogy. I'm doing my Family history for fun. It will never go near an experienced genealogist. I suspect I am not alone.

I haven't had a look at the best practices video yet I may do when I have time, but I expect it will not be of great interest. I may watch Cyndi tho

I do appreciate the help tho. It is appreciated.

ColinMc wrote: It does strike me that we (and here I mean new/ less experienced users) all need a single document with clear definitions of what is a repository, what is a source, a citation etc etc.
...
I take what you are saying, But I couldn't care less about best genealogical practice. I made my choice to work with FH, and that's what I want to learn to use to the best of my ability.
...
So I care what terms mean within the program I am using

That being the case the actual labels "repository", "source, "citation" actually do not matter that much. They are merely data entities* that within FH have relationships with each other. The question for you is then how to make best use of those relationships to do what you want to do.

* Chunks of similar data held in a standardised way - On the main FH records screen, the major data entities have their own tab.

At the core of what you are wanting to do is to manage the relationships between individuals and their families* and to record information about those individuals and families. The "managing the relationship" is key; without that requirement you could just use a word-processor document.

* A family is defined as a couple (husband/wife, father/mother)

Managing information is a lot easier if the information is recorded in a standardised way - which is why we use facts (either events or attributes) rather than just massive notes for each individual or couple (although that is possible). It is helpful to get your mind around which facts relate to couples and which to individuals - so a birth fact relates to an individual and a marriage fact to a couple. Parentage is managed through the relationship of the individual to their family (learning to understand the contents of the "All" tab and how the data is structured is useful).

Beyond the core functionality there are a series of data entities to help answer the question (if desired) - how did I work "that" out? Sometimes this is a reasoned argument best held as a bit of text attached to the fact (the "that" above) in a note. More often you rely on something external that has a level of recognised "authority" (a recalled memory, a family note, a signed document, a government record etc.). These are loosely referred to as "Sources". Being a data entity it has a structure (standard bits of information - not all of which have to be used)

Using data entities in an internally consistent way enables you to use the software to answer questions through things like queries - if you want to.

It gets complicated because FH allows you to manage the relationship between facts and sources in variations on two different ways. The relationship (the link) is referred to as a citation - a term that is in major use in academia. It is a data entity but because its primary use is as a link, it is not visible on things like the records page. In the property box (for individuals and families) you can see the space for Sources (top Right) and Citations (bottom right - although it is not labelled as such). [Click the "Show Sources" icon if necessary to show the RHS of the property box] You cannot enter citation details until you have entered a source (because the citation is a link to a source).

In database systems (and FH is a database system) links are important because you get different types of links; where links are one to many (or many to many) they need to be defined. Thus
- A fact can have many sources
(e.g. birth sourced to a certificate, a birth index, a newspaper announcement, etc.)
- A source may support many facts
(e.g. a census can support "place of residence", "occupation", "date of birth" etc.)
A citation defines each of these links

In an academic document:
You might have an Event Fact for an Individual "Born: 21 April 1926 in London".
A source for this might be "Pimlott, B, 1998, The Queen: A Biography of Elizabeth II , Wiley"
(This is an academic form of the structured definition of a book: Author/Date/Title/Publisher)
Your citation to link this fact with this source would be "Pimlott, 1998, page 2" - which provides a (hopefully) unique pointer to your source (Pimlott, 1998) and an indication of where in the source (page 2) the supporting information would be found.

This works fine where information comes from printed books and similar and is being incorporated in academic papers etc.

Outside this use it gets more complicated because it is not clear what a "source" is.
For instance

• is "the 1881 Census" a "source", or is it the collated volume of census returns for a particular enumeration district (the physical thing that the National Archives could produce if requested), or is it the specific sheet on which the individual is recorded?
• is Wikipedia a source, or is the specific webpage a source?

The more detailed the source, the less detail is required in the citation.

What you use in the great scheme of things does not matter, although a level of consistency makes it easier for you. However it fuels the "lumper" / "splitter" debate because the pros and cons of each approach are of different priorities to different people.

Repositories I don't think are that controversial possibly because they are not particularly useful to most people! They might be thought of as places where sources are kept. Thus they would have been important to the Church of Latter Day Saints - who have a major influence on the GEDCOM definition.

I don't know if this gives "a single document with clear definitions of what is a repository, what is a source, a citation etc".

Thanks, I've had a quick read through . Ill read it in more detail tomorrow night.

Colin

I couldn't care less about best genealogical practice

Most people don't and that's fine -- but many aspects of best practice arose because of painful experience with doing things in other ways. Such as discovering that unpicking the inevitable errors in a tree is a long and tortuous process because of the way you've 'simplified' recording stuff. Or finding that you've repeated a great chunk of research and wasted a lot of time (and sometimes money -- hands up anyone who bought two copies of the same certificate in their early days?) Inevitably, you're going to be dealing with a lot of information -- including a lot of information that you decide isn't relevant to your tree. As you've already realised, you need an organised and consistent system of working with it all, including the stuff you discard...

You've taken that single statement out of context from the whole discussion. It was you that raised the whole topic of "best genealogical practice"

My point is I want best practice within FH so that I try to minimise

because of painful experience with doing things in other ways. Such as discovering that unpicking the inevitable errors in a tree is a long and tortuous process because of the way you've 'simplified' recording stuff.

Within that context, yes I couldn't care less about best genealogical practice. But only because my target is to have

best genealogical practice within FH