* GEDCOM Criticisms

[DonF]

If I can just revert back up to AdrianBruce's commentary on 'Witness' in Is V5 to V6 upgrade worthwhile? (12468) - firstly, I wouldn't blame the name on TMG's use of it for marriage witnesses (especially as in TMG you can have a witness to any event).
Second, I think the origin is in GEDCOM, which used to have WITN (Witness) tag, up to GEDCOM 5.4, when it was removed and replaced with the ASSO (Associates) tag, to be used with a RELA (relationship) sub-tag.
However with true GEDCOM stupidity, ASSO was to be used to relate people to each other, not to relate people to events. The GEDCOM rules then said that if you wanted to relate associates to an event, you should use _ASSO as a sub-tag of the event. Note that any tag with a name starting with the underscore character is meant to be a GEDCOM extension, but here we have the GEDCOM authors (who frown on extensions) saying you should use it for 'normal' use. Unbelievable.
So GEDCOM does allow witnesses, but given the rules for this are as chaotic as the rest of that woeful so-called standard, it is little wonder that most programs ignore ASSO.

Don

[AdrianBruce]

... GEDCOM, which used to have WITN (Witness) tag, up to GEDCOM 5.4, when it was removed and replaced with the ASSO (Associates) tag, to be used with a RELA (relationship) sub-tag.
However with true GEDCOM stupidity, ASSO was to be used to relate people to each other, not to relate people to events. The GEDCOM rules then said that if you wanted to relate associates to an event, you should use _ASSO as a sub-tag of the event. ...

Thanks for that Don. I'd not realised GEDCOM once had a WITN tag. I concur that ASSO, while wider in its use, also loses meaning. And advocating an extension tag (_ASSO) seems to radically miss the concept of what a standard should be.

Just to be consistent in my own comments over the years - I can't agree that GEDCOM is a woeful standard, though. Or at least, no more woeful than any other standard in the known universe. Most GEDCOM issues come from programmers who consider themselves too important to read standards (NB - I'm an ex-programmer!) Most standards have contained lunacies in them somewhere, sometime. Yes, GEDCOM has been frozen and that's absurd. (There's a hobby horse looking at me in expectation - I think I'll let him be this time! )

[DonF]

Yes, Adrian, its an extremely large hobby horse, and probably not worth the effort of trying to ride!

But to say GEDCOM is a standard worries me (it fails any logical test of such a definition) and I also fear it is not other programmers who are at fault - just one small example: the GEDCOM specification uses the phrase '8-bit ASCII'. There is no such thing as 8-bit ASCII. ASCII is a 7-bit character set. This is an error that has been in GEDCOM since forever, which they were told about repeatedly, which has a simple fix, but which remains as an indictment of the author's ability to write anything accurately.
I could go on...... but I'll leave the hobby horse alone in the hope it rides off into the sunset. Quickly.

Don

[AdrianBruce]

... But to say GEDCOM is a standard worries me (it fails any logical test of such a definition) ...

Well, yes. I would have to agree - though it's all we have right now (that hobby horse is round the corner again...) Oh, and yes, you are right about 8-bit ASCII. It's just that I find it difficult to accept criticism of the "standard" when most of the errors that people use as evidence of GEDCOM's incompetence are programming errors, not errors in the "design" of the "standard".

[DonF]

Well I'm not sure it's 'most of the errors'. The following gives an analysis of only some of GEDCOM's stupidities (before highlighting programming issues):
http://www.tamurajones.net/GEDCOMCharac ... ings.xhtml

and while this mainly supports your programming stupidities argument, the bit on CONC usage is pure GEDCOM author idiocy:
http://www.gaenovium.com/presentations/ ... 0Right.pdf

The GEDCOM specs show other cases where rules are stated, then Examples are given that contradict the rules - EVEN (event) being the most-quoted classic example.
But we've probably bored everyone else with this discussion by now.....

Don

[AdrianBruce]

....
But we've probably bored everyone else with this discussion by now..... Don

Surely not! I'd only suggest that Tamura is not "most people". He's a very thorough and very logical guy and I still bet most standards couldn't stand up to his sort of scrutiny!

'nuff said!

[tatewise]

What those two links to an analysis of Gedcoms say to me is:

• Primarily, most genealogy programs are poor at implementing the specification, rather than the specification is poor.
• The CHAR specification criticism is valid, but trivial in comparison to its abuse.
• The CONC specification criticism is valid, but has relatively minor consequences.

As the two links also say, there are fairly easy ways to compensate not only for deficiences in the specification, but more importantly deficiencies in its implementaion.

[jimlad68]

I don't think this discussion can ever go anywhere. The stark fact is that regardless of how good or bad it is, at present Gedcom is the only universal portable data method (I put that badly, but you know what I mean), we have to use it as best we can, and considering how old it is, it is amazing it still functions as well as it does a nd can form the basis of FH.

For the future there is http://fhiso.org/ which FH Simon Orde is a member/ contributor, I wish it well but I am not convinced that is going anywhere fast without software developer support, and then LDS seem to eventually seem to be taking things into their own hands again with new API (or similar) for their new online trees.