* Additional Untitled Sources

[ColeValleyGirl]

Right at the top of the programme is a text menu with File, Edit, View, Add.

Click that Edit and you'll see an Undo option.

[tatewise]

Evelyn, you know about the Edit > Undo menu option that we have discussed regarding undoing Plugin actions.

[LornaCraig]

Evelyn, by the way, did you spot my post earlier in this thread where I replied to your comment about not being able to open an existing source without an existing citation? I think perhaps you have not been choosing the right option, because you can certainly select an existing source without any citation details.

[AdrianBruce]

...
That is the problem, we didn’t create them [empty sources] in FH6. ...

Well I did - and every so often, still do if I press the wrong button. And I'm on v6 still. But I think that's just loss of concentration on my part. I am wondering if the increase in input forms on the v7 screens might lead to a loss of orientation - what looks like an empty text box turns out to be an empty source? I've no idea but the only way forward is to actually track it down. If closing an empty form is indeed the issue, as Lorna suggests above, then that would be a sensible thing to look at providing warnings for, or something similar. I certainly don't recognise any normal usage Save buttons in my current v6 that would be lost in the upgrade to v7, certainly not around citations. But I'd certainly support a pop-up that said "Err, do you really want to create an empty Source?"

Some people are saying that they want explicit confirmation of every save - some that they absolutely don't want that. Balancing off those two desires is a design issue where the software guys have every right to say: "X is feasible but Y isn't". And I wonder if wanting confirmation of every save is really the requirement? Or whether it's confirmation of problematic saves like empty sources / families / persons, etc., that's the real Requirement? (Because I'd support the latter whole heartedly and turn off requests for confirmation of the former!)

[AdrianBruce]

I can't resist adding this tale against myself about the dangers of needing to confirm a Save... I've just closed down FH and - to my surprise - got a dialog box to do a back-up. This obviously implies I'd updated my file but my initial reaction was that I had done no such thing - I was about to cancel the back-up when I remembered that I had indeed updated the file, a couple of hours ago.

Now, imagine that was a "Confirm Save Y/N?" dialog - I'd have said "No" because I'd lost concentration and I'd have lost the data. (Well, I wouldn't actually because the AutoSave kicked in ages ago but AutoSave negates the point of asking to confirm a Save).

[jbtapscott]

I agree with your final paragraph Adrian - it is the saving of, effectively, "empty" records (of any type) that I don't think is correct (and goes totally against all my "database" training!) and for which a warning should be raised (in my opinion!) (Edit - this relates to Adrians first post of 2 above!)

[tatewise]

I sympathise with your suggestion, Adrian.
It is very rare that a user would want to leave a new record completely empty.
So at some point, it should be feasible to detect and warn the user about such empty records of any type.
If not soon after they are created, then at least when FH is closed.

[Gowermick]

I’ve been giving this a bit more thought, and have come to the conclusion that the problem isn’t necessarily the fault of the users creating empty citations, but the fault of the program in saving them!

As Adrian pointed out, most programs recognise when change have been made by a user, so prompts user to save the changes or lose them.

is it beyond the wit of a programmer to also spot that a newly created record is empty, so rather than save it, delete it instead. This isn’t restricted to citations but all record types.

[AdrianBruce]

... is it beyond the wit of a programmer to also spot that a newly created record is empty, so rather than save it, delete it instead. This isn’t restricted to citations but all record types.

I think that this is much closer to the real requirement that we want, and is something I would support. Couple of points, though:

• Citations aren't records in any GEDCOM sense of being top-level objects but you could quite easily drop the test down a level for Citations - not sure if anything else being empty should apply the same logic;
• Maybe the deletion shouldn't be automatic but should ask the user? I might want to create an empty source-record for use later on. Maybe??

[tatewise]

Citations can never be completely empty as they only exist linked to a Source record.
So, if the Source record is empty and is deleted, then the Citation goes with it, whether the other Citation fields are empty or not. That is how it has always worked.

[Gowermick]

Adrian,
I was referring to records in the FH sense rather than GEDCOM. i.e. Anything in FH that has a record ID assigned.
As for deletions, I was thinking along the lines of a regular garbage collection type cleanup routine, run at session closedown.
This would allow creation of a blank record during a session if user so wants, on the understanding it will be short lived and not last beyond the end of the session.

[tatewise]

That is the point Adrian and I are making. Citations are NOT records and do NOT have a Record ID.
Individual records and Source records have Record ID but Citations are just components of a parent record.

[Gowermick]

Miketate,
It may be my poor wording re source v citation. Whatever the correct term, blank records are being created. Whether it be a source or citation, this what needs deleting when blank.

[tatewise]

Sorry Mike, but you have been using FH for over 5 years!
So you should by now know the difference between the FH formal definition of a Citation and a Source.
A Citation is attached to Facts, etc, in the same way that Dates, Places and Notes are attached to Facts.
A Source is a separate record linked to a Citation in the same way a Media record is linked to a Media tab or a Citation.

All the types of record appear in the cunningly named Records Window each with their own tab.

There are some data items that could be deemed 'blank' that I would not want auto-deleted nor prompted for deletion every time FH closed, e.g. Facts with no subsidiary data.

[Gowermick]

I’d like to see an example of a data record that is blank that you would like to keep!
If it is blank, why would you want to keep it?

[tatewise]

I deliberately didn't say data record. I said data item like a Fact, which is subsidiary to a record.
As I keep trying to say 'record' has a very specific technical meaning in FH & GEDCOM terms that may be different from its common use elsewhere.

[Gowermick]

Nevertheless, I be interested in knowing why you would keep a blank fact or item!
Also if it is not a record, why would my suggested garbage collection affect it? It would only apply to blank Records!

[E Wilcock]

I think part of the trouble for me at least is the ambiguity over the word "citation".
A source record as a whole can be attached to a fact and then it becomes a citation?

So yes, I did accept the of option of a prepared citation as I wrongly thought it meant an existing source.
Thank you for the explanation. It is going to be useful to me.

[ColeValleyGirl]

I think part of the trouble for me at least is the ambiguity over the word "citation".
A source record as a whole can be attached to a fact and then it becomes a citation?

Strictly speaking, it is the link between the source and the fact that is the citation

I drew this to explain the same point to a friend. Ignore the technical symbols, just focus on the fact that a citation is a bridge between a source and a fact, and it can just be the bridge or it can have some information associated with it.

fh records_001.png (159.35 KiB) Viewed 3578 times

[PeterR]

Very useful diagram, but I think the connector between Citation and Source Record is reversed.

[ColeValleyGirl]

Very useful diagram, but I think the connector between Citation and Source Record is reversed.

Or should it be a one-one relationship. (My friend and I argued about this for ages).

[NickWalker]

Great diagram! When I tried to explain facts/citations/sources the other day I thought about drawing a (simpler) Entity-relationship diagram but didn't have the time. This is really helpful I think (at least for people used to database concepts).

[Edit: I added a comment agreeing with Peter about the source/citation link, but Helen edited the diagram at the same time making my point irrelevant ]

[ColeValleyGirl]

Nick, Peter -- I've fixed it (in my original post).

One day it might find its way into the KB, but then I'd have to do some other diagrams as well..

[tatewise]

One possible correction is that a Citation can be associated with much more than just Facts.
They can be associated directly with Individual, Family, and Place records to name a few, but may complicate the concept?

[ColeValleyGirl]

True, Mike, but as you say, it complicates matters -- especially as I want to avoid crossing lines -- and this was done in answer to a particular query.

If I do any more diagrams, I shall have to work out how to handle the full rats nest of connections.