Re: FW: [jats-list] Element for wrapping a group of xref elements

[Wendell Piez <wapiez@xxxxxxxxxxxxxxx>]

Hi Alf,

I have just a bit of color commentary to add to what Jeff said.

It's certainly possible to handle <xref rid="f1"/> given a
specification of how content is to be derived or generated for it in
display. For example, maybe the link target always has a 'label'
element <label>Fig. 1</label>, which can be retrieved; or maybe your
system auto-generates labels.

Implementing such a system is straightforward (indeed I believe the
JATS Preview XSLT will do either of these if things are set up right),
but does entail specifying constraints that will preclude anomalies.
For example, you might have rule that requires that anything targeted
by an xref have a (non-empty) 'label' element. Or that any target of
an xref be one of a set of element types for which you are
implementing automatic label generation. Etc. Schematron is good for
enforcing constraints of this sort over your data.

If PMC isn't doing this, it's likely because it feels that imposing
such constraints across its (very wide) data set isn't feasible, given
the range of tag usage they wish to support. At that scale, simpler is
better.

Similarly, the hard part of designing support for <xref rid="f1 f2
t3"/> (which the Preview stylesheets do not do) is in specifying how
it should look and work in all cases. Your suggestion that a web page
with scripting could implement a two-step link (click once to see a
list of links, a second time to follow one of them) is just such a
spec -- and given the spec, the code could be written.

Note that having done this, however, you'll be limiting the
interchangeability of your data with other systems expecting JATS
following other rules: it'll work in your stack, but not theirs. For
this purpose, you might wish to have another transformation to create
an "interchange version", converting your empty xref into something
else, such as a series of discrete xrefs with literal content, which
is more likely to be portable.

Cheers,
Wendell





On Wed, Jan 9, 2013 at 9:19 AM, Alf Eaton <eaton.alf@xxxxxxxxx> wrote:
> Thanks Jeff. I was thinking that in places where a single xref links
> to multiple target ids, a list of the targets could be displayed and
> the user can choose which one they want (this would work for
> references, tables or figures). The default - for when Javascript is
> disabled - could be to link just to the first of the targets.
>
> I haven't tried that yet to see how well it will work in practise, though.
>
> On 9 January 2013 12:52, Beck, Jeff (NIH/NLM/NCBI) [E]
> <beck@xxxxxxxxxxxxxxxx> wrote:
>> Hi Alf,
>>
>> PMC can handle multiple values in xref/@rid. We do have issues generating
>> the display if the xref is empty (but we have that for single values,
>> too), and of course the links are problematic.
>>
>> <xref ref-type="fig" rid="d349665qxx"/> or <xref ref-type="fig"
>> rid="d349665qxx d34888324qqq"/> are both tough to figure out what to do
>> with.
>>
>> <xref ref-type="fig" rid="d349665qxx d34888324qqq">Figs. 1 &amp; 2</xref>
>> is not a problem unless you want "1" to link one place and "2" to link to
>> another - which is generally what users expect.
>>
>> I'd prefer
>>
>> Figs. <xref ref-type="fig" rid="d349665qxx">1</xref> &amp; <xref
>> ref-type="fig" rid="d34888324qqq">2</xref>
>>
>> Or even
>>
>> <xref ref-type="fig" rid="d349665qxx">Figs. 1</xref><xref ref-type="fig"
>> rid="d34888324qqq">&amp; 2</xref>
>>
>>
>> Remember, the PMC Tagging Guidelines are just that - guidelines. PMC
>> requires content that is well-formed, is valid to the schema it claims to
>> follow, and represents the article(s) and publication model of the journal
>> accurately.
>>
>> Questions on PMC can be addressed to pmc@xxxxxxxxxxxxxxxxx
>>
>> Jeff
>>
>>
>>
>>
>>
>> On 1/8/13 5:40 PM, "Alf Eaton" <eaton.alf@xxxxxxxxx> wrote:
>>
>>>Jeff, would you be able to comment on whether this actually is a
>>>restriction of PubMed Central? It seems like it would be nice to mark
>>>up a collection of references as one element, but I guess it's
>>>forbidden by PubMed Central because the HTML display can't handle it?
>>>
>>>I could work around the restriction by adding another element around
>>>the whole group of references, perhaps, though I'm not sure what
>>>element that would be.
>>>
>>>Alf
>>>
>>>
>>>---------- Forwarded message ----------
>>>From: Alexander Schwarzman <aschwarzman@xxxxxxxxx>
>>>Date: 5 December 2012 18:57
>>>Subject: Re: [jats-list] Element for wrapping a group of xref elements
>>>To: jats-list@xxxxxxxxxxxxxxxxxxxxxx
>>>
>>>
>>>I think one has to consider whether the articles are going to be
>>>submitted to PubMedCentral.
>>>
>>>PMC Guidelines
>>>http://www.ncbi.nlm.nih.gov/pmc/pmcdoc/tagging-guidelines/article/tags.htm
>>>l#el-xref
>>>admonish:
>>>
>>>Do not tag multiple id values in @rid.
>>>
>>>--Sasha
>>>
>>>On Wed, Dec 5, 2012 at 8:36 AM, Bruce Rosenblum <bruce@xxxxxxxxx> wrote:
>>>> Alf,
>>>>
>>>> You can tag in all versions of NLM or in JATS as:
>>>>
>>>>         "induce co-stimulatory molecule elevation<xref rid="r14 r15
>>>>r18">14,
>>>> 15, 18</xref>"
>>>>
>>>> This works fine for references, and you can also do the same with
>>>>tables or
>>>> figures, as in:
>>>>
>>>>         <xref rid="t1 t2 t3 t4">Tables 1-4</xref>
>>>>
>>>> but beware that this may create a usability problem for the reader
>>>>because
>>>> there's no "point" on which you can click to go to table 2 or table 3.
>>>>This
>>>> is less of a problem with references since they are all in the same
>>>> proximity.
>>>>
>>>> In both cases, the key behind this is that the DTD declares rid to be
>>>> IDREFS, not simply IDREF.
>>>>
>>>> Bruce
>>>>
>>>> At 08:26 AM 12/5/2012, Alf Eaton wrote:
>>>>>
>>>>> Is there an element in the JATS DTD that is suitable for denoting a
>>>>> collection of xref elements in the body of an article?
>>>>>
>>>>> For example, in various citation styles:
>>>>>
>>>>> "induce co-stimulatory molecule
>>>>>
>>>>>
>>>>>elevation<xref-group><xref>14</xref>,<xref>15</xref>,<xref>18</xref></xr
>>>>>ef-group>."
>>>>>
>>>>> "induce co-stimulatory molecules elevation
>>>>> <xref-group>(<xref>Gagliardi 2000</xref>; <xref>Berin
>>>>> 2008</xref>)</xref-group>."
>>>>>
>>>>> "induce co-stimulatory molecule
>>>>> elevation<xref-group>[<xref>14</xref>-<xref>16</xref>]</xref-group>."
>>>>>
>>>>> If the citations were marked up in this way, then it would be possible
>>>>> for the renderer to choose which style
>>>>> (numeric/superscript/parenthetical) to use when rendering them.
>>>>>
>>>>> For example, <xref-group><xref rid="ref-14">14</xref><xref
>>>>> rid="ref-15">15</xref><xref rid="ref-16">16</xref></xref-group> could
>>>>> be turned into a comma-separated list of numbers, a hyphenated pair of
>>>>> numbers, or a semi-colon-separated list of author/year pairs, with the
>>>>> appropriate punctuation.
>>>>>
>>>>> Thanks,
>>>>> Alf
>>>>
>>>>
>>>> -------------------------------------------------------------------
>>>> This email message and any attachments are confidential. If you are not
>>>>the
>>>> intended recipient, please immediately reply to the sender or call
>>>> 617-932-1932 and delete the message from your email system. Thank you.
>>>> -------------------------------------------------------------------
>>>> Bruce D. Rosenblum
>>>> Inera Inc.
>>>> 19 Flett Road
>>>> Belmont, MA 02478
>>>> 617-932-1932 (office)
>>>> bruce@xxxxxxxxx
>



-- 
Wendell Piez | http://www.wendellpiez.com
XML | XSLT | electronic publishing
Eat Your Vegetables
_____oo_________o_o___ooooo____ooooooo_^