[Bioc-devel] Vignette questions: location of html BiocStyle vignette, fake + real vignettes not necessary?

[Andrzej Oleś]

Hi Leonardo,

thank you for considering using BiocStyle for your vignettes! Please read
below.

On Thu, Oct 29, 2015 at 10:17 PM, Dan Tenenbaum <dtenenba at fredhutch.org>
wrote:

>
>
> ----- Original Message -----
> > From: "lcollado" <lcollado at jhu.edu>
> > To: "bioc-devel" <bioc-devel at r-project.org>
> > Sent: Thursday, October 29, 2015 2:09:57 PM
> > Subject: [Bioc-devel] Vignette questions: location of html BiocStyle
> vignette, fake + real vignettes not necessary?
>
> > Hi,
> > I'm planning on switching some vignettes to BiocStyle (html version).
> > Is there any particular reason why the html vignette for BiocStyle is
> > outside of the /vignettes directory? I see it at
> > /inst/rmarkdown/templates/html_document/skeleton/skeleton.Rmd with
> > /vignettes/HtmlStyle.Rmd linking to it. I didn't know that you could
> > write such links, thanks! But does it help with speeding up the
> > building or testing steps if the vignette is outside of /vignettes?
> > BiocCheck has its vignette at /vignettes/BiocCheck.Rmd so I know that
> > such a setup is not necessary.
>
> I will let the BiocStyle maintainers answer this one.
>

There is no specific reason for this other than convenience: at some point
I had the idea of providing a template for BiocStyle::html_document and
figured out that the package vignette itself could serve that purpose. So
to avoid redundancy I put the vignette in the templates directory which is
discovered by RStudio/rmarkdown, and link to it from the  vignettes
directory in order to build it as the package vignette as well. The other
way round wouldn't probably work as the vignette source gets copied to
/inst/doc upon installation. I don't see any obvious reason how this
approach could affect performance in any way, either positively or
negatively.

Cheers,
Andrzej

	[[alternative HTML version deleted]]