Deal with unttaged reference

[fabiobatalha]

For Erudit the first main purpose of using Texture is to incorporate it in our Production Chain for Quality Control and content fixing.

Erudit now do not deal with fully marked up reference, it is one of our future improvement requirements for our production chain, and it is good that Texture is already quite advanced on it, but for now, we need that Texture take a look on how to deal with untagged references.

The untagged references have its purposes:

• Preserving the presentation style of each reference;
• Allow the check-up of metadata of references produced by automated processes (OTS, Crossref lookup, etc);
• Allow developers to produce output formats (conversions) for a variety of purposes, (Display, data processing, etc)
• Decrease the costs according to the stakeholders needs.
  • Identify all the metadata of a reference is an expensive task, stakeholders, can choose what metadata is relevant for their internal processes and needs. Ex: Erudit at this moment should use mainly the untagged reference, meanwhile we develop the expertise to produce fully marked up references with great quality and low costs.

[michael]

I will try to integrate a pub-type="other" with free form text. Will ping you when ready so you can test.

[fabiobatalha]

@michael

I don't think it is a good approach.

1. We still want to know the type of the reference;
2. pub-type=other should still being used by tagged references;

I was thinking we should have a specific field for this purpose. Maybe we can use the styled-content

It is allowed inside element-citation

we should have something like:

<element-citation>
    ...
    <styled-content specific-use="display|print|..." style-type="vancouver|abnt|iso">
    Reference.....
    </styled-content>
    ...
</element-citation>

[fabiobatalha]

We don't need to have at this moment a detailed attribute options, the specific-use="display" is enough for now.

[michael]

Good suggestion. I wonder if it makes sense to introduce this as an optional field generally, where you can override the display. E.g. when you're not happy with what the renderer produces or in your case, don't have structured info.

Not sure... i mean the ultimate goal should be to have structured info and a stable renderer. Maybe we should rather not show this field and use a configuration for Texture (like unstructuredReferences:true) that Erudit can use.

[fabiobatalha]

I agree with the configuration

On Wed, 7 Mar 2018 at 17:37 Michael Aufreiter notifications@github.com wrote:

Good suggestion. I wonder if it makes sense to introduce this as an optional field generally, where you can override the display. E.g. when you're not happy with what the renderer produces or in your case, don't have structured info.

Not sure... i mean the ultimate goal should be to have structured info and a stable renderer. Maybe we should rather not show this field and use a configuration for Texture (like unstructuredReferences:true) that Erudit can use.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/substance/texture/issues/412#issuecomment-371310498, or mute the thread https://github.com/notifications/unsubscribe-auth/AAexTuEESooOYmOaTvAihqe2VwXiFhmKks5tcGE0gaJpZM4Se7KO .

[fabiobatalha]

@michael

• Yes this should be an optional field.
• We may include a toggle display button for those who have the unstructuredReferences:true
• The toggle button switch between structured and unstructured display or the structured display is always available, and the toggle button display/hide the unstructured reference bellow the structured reference
• The toggle button could be above the reference list

I will try to mock the behaviour to give an example.

[fabiobatalha]

@michael My best shoot!

[michael]

Do we really want to mix it (like some refs are structured and others have a display string)? I think this would lead to bad UI/UX.

If we start the editor with the unstructuredReferences:true option, we could allow editing in place (such as with Footnotes, no modal dialog). However this means, you have to decide for a whole article, do I want to use structured references or manual plain references.

[fabiobatalha]

@michael

It is ok to go ahead with the second approach having in mind we still need to identify the type of the reference.

[michael]

Is this really pressing or could we put it on the agenda for the meeting? I know this sounds like a rather trivial feature, but it has implications on usability. Would be good to talk it through in detail and come up with something that satisfies both use cases well.

[fabiobatalha]

@michael

No pressure we can put it in our agenda.

[fabiobatalha]

@michael

We would have the ability of display at least the element pub-id together with the untagged reference. The purpose is to easily view the DOI number when it is available.

[fabiobatalha]

Hello @michael

We are documenting our Érudit Publishing Schema and writing the schematron rules to address our needs. It will be compliant with JATS4M, and include some extra validations.

One of these validation isn't officially defined in JATS4M.

Besides the UX issues, we must have one decision about the data structure to handle the untagged references.

Our suggestions it to go forward with the style-content element inside element-citation. It will be mandatory in the Érudit Publishing Schema, see: http://erudit-ps-documentation.readthedocs.io/en/latest/tagset/element-styled-content.html

Do you have any constrains about this approach?

[michael]

Hey @fabiobatalha!

I'm still not sure what the best way would be to handle untagged references. But I agree we should soon come up with a schema (tagging guideline) for those; the ui/ux can come later. Give me a bit of time and I will come back to you! (having a very busy week)

[fabiobatalha]

Hi @michael

No problem, take your time. We have some time before launching the first version of Érudit Publishing Schema.

[lpanebr]

This looks to me as the perfect use-case for mixed-citation as it may hold unstructured and untagged content as well as mix of both. Scielo currently uses both element-citation and mixed-citation. The former for bibliometrics and the later for display:

<ref id="B027">
  <label>27</label>
  <element-citation publication-type="journal">
    <person-group person-group-type="author"><name><surname>Yamamoto</surname><given-names>T</given-names></name><name><surname>Nohara</surname><given-names>K</given-names></name><name><surname>Uchida</surname><given-names>K</given-names></name><name><surname>Yamaguchi</surname><given-names>K</given-names></name></person-group>
    <article-title>Purification and characterization of secretory proteinase of <italic>Candida albicans.</italic></article-title><source>Microbiol Immunol</source>
    <year>1992</year><volume>36</volume><issue>6</issue>
    <fpage>637</fpage><lpage>641</lpage>
    <ext-link ext-link-type="uri" xlink:href="http://dx.doi.org/10.1111/j.1348-0421.1992.tb02064.x">http://dx.doi.org/10.1111/j.1348-0421.1992.tb02064.x</ext-link>
    <pub-id pub-id-type="pmid">1522813.</pub-id>
  </element-citation>
  <mixed-citation>Yamamoto T, Nohara K, Uchida K, Yamaguchi K. Purification and characterization of secretory proteinase of  Candida albicans.Microbiol Immunol. 1992;36(6):637-41. http://dx.doi.org/10.1111/j.1348-0421.1992.tb02064.x. PMid:1522813.</mixed-citation>
</ref>

An example of mixed content in the mixed-citation would be:

  <mixed-citation>Yamamoto T, Nohara K, Uchida K, Yamaguchi K. Purification and characterization of secretory proteinase of  Candida albicans. Microbiol Immunol. <year>1992</year>;36(6):637-41. <ext-link ext-link-type="uri" xlink:href="http://dx.doi.org/10.1111/j.1348-0421.1992.tb02064.x">http://dx.doi.org/10.1111/j.1348-0421.1992.tb02064.x</ext-link>. PMid:<pub-id pub-id-type="pmid">1522813.</pub-id>.</mixed-citation>

As for the UI/UX it is indeed a bit more challenging... It would help to think about the UI if @fabiobatalha could tell more about which fields MUST be tagged within Erudit's unttaged citations.