Dear all,welcome to the Wordpress for Scientists Google Group. The goal of this group is to promote Wordpress as a tool for writing and distributing scholarly content. Wordpress is used by many science bloggers, and I'm sure we can improve Wordpress as a writing tool for them. Wordpress is also used for lab notebooks and scholarly papers, but several limitations (e.g. reference management) have so far limited a more widespread use for this purpose. This group can hopefully become a central discussion place for these ideas.IdeasWhat plugins, themes or other Wordpress tools should we develop? What are the most interesting projects and who is interested in helping with them?Best Practices
- What are recommended plugins for science blogging, writing scholarly papers, etc.?
- Should plugin developers use a common tag (e.g. "science") when submitting their plugins to the Wordpress plugins directory?
- Can we agree on standards for scholarly content in Wordpress, e.g. import/export formats?
DevelopmentThe group should support developers interested in creating science-specific plugins. This could mean practical advice and/or help in testing. I'm particularly interested in making the different tools work better together. It would also be nice to coordinate the work so that we for example come up with one great paragraph-level commenting tool instead of five that work most of the time. And it would be helpful to know who is already working on a particular project or idea.I'm personally most interested in developing solid reference management tools for Wordpress, including integration with available reference managers. I'm fairly new to Wordpress, but have released four Wordpress plugins in the past few weeks (Contact Info Options, BibTex Importer, Link to Link, ePub Export).Kind regards,Martin
Dear list,
I am quite excited to see Martin's post <http://blogs.plos.org/mfenner/2011/02/14/how-to-use-citation-typing-ontology-cito-in-your-blog-posts/> yesterday on using CiTo through his updated Link-to-Link plugin. I have just started trying this out, so far seems to be working quite well and seems to include the most common categories I will need. I am curious how to access this information or make it visible/accessible to my readers. Just skimmed through the Shotton paper <http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=2903725&tool=pmcentrez&rendertype=abstract> again, not clear if there's a central location from which to discover/explore the network of citations, or how to make the reason for citation visible as in Shotton's Figure 5?
It would be useful for the citation system to generate footnote-based citations rather thank links. papercite had a reasonable way of doing this with its shortcodes, but I've experienced performance issues with it on large .bib files that have led me to disable that plugin. Kcite seems to do a very nice job of this part, but in general citing by my bibtex keys is a bit faster than having to look up the doi's manually. For some reason kcite citation gives me a link to "Bibliography in JSON format" rather than the DOI link so a reader can access the original paper. Is there a way to toggle this?
Thanks to Martin and others for creating this list. A little (possibly uncessary) background: I use Wordpress for my research lab notebook in theoretical & computational ecology and evolution. After experimenting with a few different platforms I've found it a particularly flexible and effective tool. Overall I'm interested in further Wordpress development in several directions:
A) Basic needs
Citations -- Integrate to my workflow (Mendeley/bibtex and the above plugins?)
Figures -- Including links to version of code generating it (trying this by integration with github+flickr)
Equations -- LaTeX/mathjax seems to do this fine.
B) Integrate lab notebook with scientific databases & link back from pubs/databases to notebook
C) Integrate/represent social context of research.
-Carl
On Fri, Feb 4, 2011 at 1:42 PM, Martin Fenner <fenner...@mh-hannover.de> wrote:
Carl,
my goal is to make something like Shotton’s Figure 5 possible using Wordpress. I will work on a citation parser/bibliography formatting tool at the Peter Murray Rust hackfest March 12/13 in Cambridge. The kcite plugin obviously provides some of this functionality already. I’ve just updated my Link to Link plugin to version 1.1.1 with two changes:
- Use the kcite shortcode (when selected in the plugin options)
- Use the cito:disagreesWith format for CiTO relationships
Link to Link and kcite now work in combination to insert citations and create a bibliography. The kcite shortcode uses source=url when the link is not a DOI. And it includes the CiTO relationship when this is selected in the plugin options. Maybe these two modifications require an update of the kcite plugin ;).
I think that CiTO relationships work better with links than with footnotes. A paper can be cited several times, and the CiTO relationship might not always be the same. And the CiTO relationship is embedded directly in the text.
> It would be useful for the citation system to generate footnote-based citations rather thank links.
Link to Link handles everything that has a link, not just DOIs. It
doesn't handle references without a URL.
Kind regards,
Martin
Carl Boettiger <cboe...@gmail.com> writes:
> It would be useful for the citation system to generate footnote-based
> citations rather thank links. papercite had a reasonable way of doing this
> with its shortcodes, but I've experienced performance issues with it on
> large .bib files that have led me to disable that plugin. Kcite seems to do
> a very nice job of this part, but in general citing by my bibtex keys is a
> bit faster than having to look up the doi's manually.
Yes, adding in DOIs is a pain. We are starting to work on proper tool
chains for doing this. My own tool chain is a little unusual
(Emacs/Bibtex/asciidoc), but was a relatively small amount of work to
make usuable.
http://www.russet.org.uk/blog/2011/02/tooled-up-bibliographies/
Now I can do regexp look-up through my bib files, and auto-insert the
cite shortcode. Obviously, we want to look at something more widespread
than my own tool chain.
We are also refactoring kcite at the moment; I want to make it easier to
add new forms of citation besides the two we support at the moment, that
is DOI and pubmed ID. Most importantly, we want to support URLs, so that
you can formally cite other kblogs; the last thing that I want is to
encourage people to use DOIs (which our kblog articles have), as DOIs
are, to my mind, relatively pointless.
If there is a demand, we can add a public API to kcite, making this
extensible.
> For some reason kcite citation gives me a link to "Bibliography in
> JSON format" rather than the DOI link so a reader can access the
> original paper. Is there a way to toggle this?
Hmmm. Well, no, there isn't a way to toggle it, because it sounds like a
bug. Can you send a message to knowledgeb...@knowledgeblog.org
for me? Preferably with a URL or equivalent so that we can see. The
"Bibliography in JSON format" is supposed to offer you a download of all
the references.
> Thanks to Martin and others for creating this list. A little (possibly
> uncessary) background: I use Wordpress for my research lab notebook in
> theoretical & computational ecology and evolution. After experimenting with
> a few different platforms I've found it a particularly flexible and
> effective tool. Overall I'm interested in further Wordpress development in
> several directions:
>
> A) Basic needs
> Citations -- Integrate to my workflow (Mendeley/bibtex and the above
> plugins?)
Mendeley has a public API so should be easy to extend to. Bibtex also I
think should work, although, for me, this already works client side.
Where do you edit the text? In wordpress, or outside of it?
> Figures -- Including links to version of code generating it (trying this by
> integration with github+flickr)
We are really interested in this; at the moment, I think we are going to
be pushed to do anything more than a rough and read implementation
within the scope of our current grant.
> Equations -- LaTeX/mathjax seems to do this fine.
It does seem to do the job, doesn't it.
> B) Integrate lab notebook with scientific databases & link back from
> pubs/databases to notebook
We're also planning to look at this; it's is explicitly mentioned in our
JISC grant. Basically, I think this is easy, and can largely be done
with pingback/trackbacks. For those databases which don't support this (which is
almost all of them at the moment!), we use a pingback/trackback proxy.
> C) Integrate/represent social context of research.
Well, that one is a bit outside of something with a read software
solution.
Phil
> * Use the kcite shortcode (when selected in the plugin options)
> * Use the cito:disagreesWith format for CiTO relationships
>
> Link to Link and kcite now work in combination to insert citations and
> create a bibliography. The kcite shortcode uses source=url when the link is
> not a DOI. And it includes the CiTO relationship when this is selected in
> the plugin options. Maybe these two modifications require an update of the
> kcite plugin ;).
Neither of which are rocket-science. The URL is in the works. CiTO I've
wanted to use for ages. I think it's easy enough to add.
>> It would be useful for the citation system to generate footnote-based
> citations rather thank links.
>
> I think that CiTO relationships work better with links than with footnotes.
> A paper can be cited several times, and the CiTO relationship might not
> always be the same. And the CiTO relationship is embedded directly in the
> text.
This is true. I think it's a general issue for CiTO, though. There needs
to be a way to say "I am not talking about this internal anchor, but the
link at the end of it".
Phil
Kind regards,
Martin
> On Wed, Feb 16, 2011 at 2:27 PM, Carl Boettiger <cboe...@gmail.com> wrote:
>> Hmm, for some reason it just inserts
>> [link2link id=""]CiTo[/link2link]
>> rather than the kcite code when I try inserting a reference using the
> link to link plugin. I believe I have everything updated correctly. any
> ideas?
>
> I now get the link to link to insert the correct code, i.e.
>
> [cite source='doi' rel=cito:cites]10.1186/2041-1480-1-S1-S6[/cite]
>
> but the rel=cito:cites seems to break the kcite shortcode, so the shortcode
> is displayed raw rather than as a citation.
It will. Kcite does not have support for this form of citation yet. Even
if it did, it would probably be.
[cite source='doi' rel='cito:cites']10.1186/2041-1480-1-S1-S6[/cite]
I am working on this as we speak. The next version will not break with a
rel="cito" attribute, although it won't do anything with it. I'm happy
to add cito support to the version after that.
Phil
Can't remember. Do handles return metadata or is that just the DOI? Do
you want it for preprints or some such?
Otherwise, "functional" is just going to mean sticking the handle at the
end in the list. More than happy to do this. I think, in general, though
the idea of having a pluggable identifier scheme interface is looking
more and more attractive.
Phil
Mark Hahnel <mha...@imperial.ac.uk> writes:
--
Phillip Lord, Phone: +44 (0) 191 222 7827
Lecturer in Bioinformatics, Email: philli...@newcastle.ac.uk
School of Computing Science, http://homepages.cs.ncl.ac.uk/phillip.lord
Room 914 Claremont Tower, skype: russet_apples
Newcastle University, msn: m...@russet.org.uk
NE1 7RU twitter: phillord
As Simon says (erm...), we think that doing 1) is the correct approach,
but that it needs support from 2.
Most of our decisions with knowledgeblog have been around what the
authors want; by and large, this is "make my life easy, and don't make
me learn anything new". So, we support this.
Ironically, if that were all we cared about, we would, I think, largely
be done. Most of the tool chains already support HTML. Take for instance
the references on this article:
http://www.russet.org.uk/blog/2010/07/realism-and-science/
All looks good to me. All done with latex, bibtex and latextowordpress.
Who needs a plugin?
So, why kcite? Surely, it is solving a problem that is, basically,
solved. Well, in this case, I think, we are trying to support the reader
and consumer of the articles, not the author. As Simon points out, I
think we can do this so that the author doesn't notice; we just fiddle
with the tool chain to generate kcite shortcodes. But, what does the
reader get out of this. Well:
- We are using the online database, keyed on a public identifier,
whether it be URL, pubmed ID or DOI. Kcite makes it easier for the
authors to get this right, rather than wrong. Which is better for the
reader.
- We, the publisher, have the data in a computational form. It's not by
chance that this fits well with adding CITO form of semantics; it's
by design. We can alter the presentation in anyway we choose. Hard to
do if word/latex/google docs generates the bibliography directly.
- We, the publisher, can give the data in a computational form to the
reader. Or, rather, in most cases, the reader's client. Our hope is
that we can use something like citeproc.js, so that the reader can
choose the citation style that they want, on the fly, in the client.
It is with this plan in mind that kcite generates a bibliography in
JSON, in case any of you were wondering about that otherwise obscure
decision.
Finally, of course, if the consumer is not human, well, they probably
will be glad just to get the public identifiers without all the cruft
that humans need. I hope to add this as well (to knowledgeblog rather
than kcite).
The "do it in wordpress" approach is fine; indeed, there is also a
public identifier solution for this from cross ref.
http://www.crossref.org/CrossTech/2008/02/crossref_citation_plugin_for_w.html
But, for my money, giving wordpress access to the public IDs has too
many advantages to be ignored.
Phil
Simon Cockell <sjco...@gmail.com> writes:
--