[bt-devel] RFC: Remove modules installed by the native package manager

[Jonathan Marsden]

Peter von Kaehne wrote:

> My considered and often stated view is that this should not happen.

The fact remains that such modules do exist.  And there are people in 
the world who, unlike you,  feel that packaged SWORD modules do in fact 
have their place -- as evidenced in part by the very fact that such 
modules exist :)  Since many SWORD modules are explicitly public domain, 
there is no way to "force" packaged SWORD data to cease to exist, even 
if everyone on this list fully agreed with your viewpoint.

> These are texts, documents and not programme modules.

There are, as I am sure you know, many other packages for Linux 
distributions which are data packages, not program packages.  SWORD is 
not at all exceptional in that regard.  Further, SWORD modules currently 
need to be "installed", so they are rather unlike other texts and 
documents which generally speaking do not require such installation.

Given that packaged SWORD data modules *do* exist, having front ends be 
able to appropriately handle the existence of such packaged modules is a 
clear benefit to their (our) users.

> No one is packing a few sample PDFs with evince.

When there are as many different SWORD modules in the world as there are 
PDFs, and when SWORD modules do not need to be "installed", but are 
simply a single document file in a defined and formally standardized 
file format (maybe in a hypothetical SWORD 4.x, the underlying library 
storage representation of a module might be a single OSIS 2.9.9 XML 
file, with indexes being created and cached as needed??), *then* 
packaging individual document files might be universally seen as totally 
useless, and comparison with PDF would be a lot more valid.

For now, I suggest that we focus on the real world, in which such 
packaged SWORD modules really *do* exist, and improve the end user 
experience in that (possibly imperfect) world, rather than discuss a 
hypothetical situation in which no such packages "should" exist :)

In that context, Sveinung's work is IMO a valuable contribution.

Jonathan