Re: [docbook] On family/given/first/last names

[Geraint North <geraint@transitive.com>]

>>> Geraint North wrote:
>>>> Out of interest, what do people use the firstname/surname/ 
>>>> lastname/givenname distinction for?  In Docbook, its main use is  
>>>> clearly to identify document authors visually on the page.
>>>
>>> Contributors? A group of authors?
>> Sure, but that's still basically the same thing - document authors.
>
> May be for your usage.
> It wasn't for mine? I wrote 95% of a document and received
> (much appreciated) contributions from two others.
> So no, contributors, not authors in my semantics?

Ah, I see where you're coming from.  No - I wasn't proposing that we  
remove the authorgroup and contributor semantics, just that they had  
identical requirements to <author> as far as the requirement to  
accurately describe a person's name is concerned.

>>>>  Given the variation in naming schemes internationally, I'd be
>>>> very unlikely to perform any sort of search restricted to  
>>>> givenname or familyname, so what else could it be used for?
>>>
>>> I like the suggestion on James page (one of the respondents) to  
>>> enable
>>> 'search term' so that I could search on what the person wants rather
>>> than what a non-local might use?
>> The thing is, if I'm searching for a document author, I don't want  
>> to miss a result because I'm unfamiliar with the particular local  
>> characteristics of the person's name - the only sensible thing to  
>> me would seem to be for the XML to contain every component of  
>> their name, and for me to search on all of them.
>
> But that's what the JC page elaborates?
> What you're searching on is i18N dependent?
> What the states might consider 'normal' to search on is
> way out different for Thailand or Burma.
> So yes, mark up each name component (and the 'linking'
> items! That was very subtle, a-b is different from a b)
> and then provide the metadata for that name, as used
> in that locale. <xname role='searchFirst'>.... etc
>
> This picks up the 'usage' idea?
> This man is known as Norm.
> His primary search 'name' is Walsh.
> His first name is 'Norman' etc.

I'd suggest that it isn't up the the document publisher to decide  
these things - it is up to the contributor/author to supply the  
writer with the appropriate tags.

So, in your example, Norm says to the publisher:

"Here is how I like to be known:
   <displayas>Norman Walsh</displayas>
   <namecomponent> Norman <namecomponent>
   <namecomponent> Norm   <namecomponent>
   <namecomponent> Walsh  <namecomponent>
"

If we had two people, "Walter Scott" and "Bob Walter", what use would  
a "primary" search term have?  If I searched for "Walter", I'd  
probably want to see them both.  Are you saying that the primary  
search term could be used to direct the ordering or relevance metric  
of the search results?

I guess the basis of my argument is that the only solution that would  
ensure that the i18n requirements were effectively dealt with would  
be to ask the person how _they_ want their name displayed and  
searched for.  This would mean that the journals wouldn't get their  
consistency, but as an author or reader, it would seem more important  
to me that a particular person was consistent _across_ different  
publications, rather than the list of authors be consistent _within_  
a single publication.

>  In order to do that, you need to know the
>>> function of each part of the name. Assume the journal wants the  
>>> names displayed as "Last, First M.". A guy named Luis Lopez  
>>> Penabad would obviously be displayed as "Penabad, Luis L."  
>>> although this is all wrong. He is Spanish, has one firstname  
>>> (Luis) and two lastnames (Lopez Penabad), one from each of his  
>>> parents. The correct display form is therefore "Lopez Penabad,  
>>> Luis". The firstname/surname/lastname/givenname elements somewhat  
>>> help to get this right, although you still face lots of problems  
>>> with non-Western names, and even with some standard US names.  
>>> E.g. if a person goes by his second "firstname", like L. Clinton  
>>> Webb, the current DocBook markup has a hard time to catch this
>
> Which points (me) to a need for a 'usage' class of metadata?
> searchTerm
> printOrder-1
> printOrder-2
> etc.
>
> which is likely usage oriented?

What would printOrder-1 and printOrder-2 represent?  I'd be generally  
against giving the publishing framework enough semantics for it to be  
able to rearrange my name as it saw fit, but them I'm biased, because  
"Geraint North" is many times more distinctive and memorable than  
"North, G". :-)  I'm sure L. Ron Hubbard and Bill S. Preston would  
also agree. :-)


>> If publishers do have strict guidelines in this regard, it would  
>> be interesting to know _why_ they have these guidelines - I  
>> suspect that they are artefacts from a time when search would be  
>> performed on the printed document itself, which (as Markus showed  
>> in his example) has clear disadvantages.
>
> After you in terms of getting them to change?
> King Canute is probably a good analogy!
> Stopping the tide might be easier though.

Yes - I agree that getting them to change is an unrealistic  
proposition, but by adding the 'wrong' tags to DocBook, I'd have  
thought that we'd be in danger of pushing them down the wrong path  
(for some subjective value of 'wrong').

For example, DocBook could take the approach of specifying the given  
name and family name of an author, and allowing the XSLT process to  
choose whether to render it as "<givenname> <familyname>" or  
"<familyname>, <givenname>".  I'd not be in favour of this approach,  
since it has all the i18n issues and corner cases that we're now  
aware of.  If DocBook only supported <displayas> and <searchterm>  
tags, then it would not be possible for the XSLT to trivially make  
such decisions.  This would be good for authors and readers, but  
might annoy the publishers.  I guess the question is whether such  
annoyances would cause them to change their ways, or to find an  
alternative to DocBook with authorship semantics that suit them better.

>> Obviously an impossible proposal, but if _I_ were King of the  
>> International Author Naming Committee, there seem to be only two  
>> things of interest:
>> - How each author would like their name displayed
>
> But display is the 'person' issue?
> What of the 'journal' issue
> and the other issues (I like to be known as... etc)
>
>> - An unordered set of all the components of their name (including  
>> variants), for search purposes.
>> e.g.
>> <author>
>>   <displayas>L. Ron Hubbard</displayas>
>>   <namecomponent>Lafayette<namecomponent>
>>   <namecomponent>Ron<namecomponent>
>>   <namecomponent>Ronald<namecomponent>
>>   <namecomponent>Hubbard<namecomponent>
>> </author>
>
> Which misses Markus and his darned journals,
> me and my 'I like to be called' stuff.
>
> It's the metadata about namecomponents that we
> are missing, to keep Markus.. sorry, his journal
> publishers and their quill pens, happy?

Yes - I guess my naive view is that I see no useful reason why the  
journals should refer to me by anything other than how I like to be  
called, and so the description of me in a journal should be  
completely under my control.  You are right that perhaps something  
extra is needed "how I like to be known", perhaps something like:

<author>
   <displayas role = "citation">Geraint North</displayas>
   <displayas role = "formal">Mr North</displayas>
   <displayas role = "informal">Geraint</displayas>
   <searchterm>Geraint</searchterm>
   <searchterm>Michael</searchterm>
   <searchterm>North</searchterm>
</author>

This would allow people to find me in searches easily, cite me  
properly in journals in the way that I'd like, and know how to  
address formal/informal communication with me.  I guess that the main  
difference between this approach and the existing DocBook one is that  
the focus shifts from "What are the components of your name? (so I  
can work out how to address you)" to the much more direct "How do you  
like to be addressed?"

As you say, very interesting stuff.   My experience thus far has  
really just been with Japan - when I get email from Japanese  
customers addressed to "Geraint-san", it is often not obvious from  
the individual's business card how I should address the reply.  I  
have no idea how I'd cope with a Thai customer!

Thanks,

Geraint North
Principal Engineer
Transitive
* The leader in cross-platform virtualization

This E-mail and any attachments hereto are strictly confidential and  
intended solely for the addressee. If you are not the intended  
addressee please notify the sender by return and delete the message.  
You must not disclose, forward or copy this E-mail or attachments to  
any third party without the prior consent of the sender.