[sword-devel] modules for debian

Ben Armstrong sword-devel@crosswire.org
Mon, 29 Jan 2001 18:16:36 -0400

On Mon, Jan 29, 2001 at 09:30:02PM +0000, Trevor Jenkins wrote:
> > ... "Core" suggests "stuff that libsword
> > needs to work properly". 
> The implication is greater than that. If only certain modules are included
> in the Debian distribution/mirrors that exactly what people will
> think. "These files are included in the distibution ergo they must be
> essential to the correct opertaion of the program."

That's simply not the way Debian works.  Things basically get packaged for
Debian on a demand basis.  The "Depends" vs. "Suggests" makes it clear
what is required and what is not.  In no way does a package's presence in
Debian indicate "essentialness" of that package.  Have you ever really
looked at Debian?  It's *huge*.  A Debian CD will contain all kinds of
packages and only subsets of data for many of them.  Nowhere does Debian
represent the data sets provided as "essential".  If people are getting
that impression, it is merely from ignorance (i.e., not reading the docs).
If Dan is really concerned about people not being led astray, he can
place it prominently in the description of each module.  "This is one of a
sampling of modules for Sword which has been packaged for Debian.  For the
full range of available modules see <insert-url-to-modules-here>."  Of
course, that is assuming that the user even bothers to look at the package
description ... but if they're not even doing that, I'm afraid there's no
help for them if they are getting the wrong impression.

> > It seems what we are looking for is a usable
> > enough sampling of the modules available for sword that the package could
> > be used with what Debian alone provides quite successfully by most
> > (English-speaking) people.  It gives the user an idea of what sword is
> > capable of without having to provide every single module within Debian.
> Whilst that's a laudable intent I do not believe that this is how people
> will view the inclusion of a selected few modules.

Regardless of how they view it, there is a practical problem Dan has to
resolve here that has nothing to do with peoples' impressions.  There is
no "data" section in Debian yet, so he cannot burden the many Debian
mirrors with the full range of distributable modules available for Sword. 
Therefore, it is necessary to select some subset of modules for starters. 
I believe it should be representative enough so that people can get a
taste of what is available.  This is exactly the same situation as Debian
has with themeable window managers.  There are many, many themes
available.  Debian doesn't package all of them.  Instead, a subset of them
are selected by the maintainer based on whatever criteria ("usability",
"good looking", "small") the maintainer deems to be appropriate.  If/when
the user decides to try out some other themes, they have all of themes.org
to select from.  It should be the same way with Sword's modules.  The
modules packaged for Debian should satisfy the new Sword user's curiosity. 
Additionally, it would be nice to include a *useful* subset of modules, as
some users will not have the means to easily obtain additional modules
(e.g. if the user is not the person responsible for installing the package
on the system, or the user doesn't have the bandwidth, etc.)  Basically,
it is up to Dan to decide which modules get included.  Of course, it is
also up to the Debian archive maintainer to put on the brakes if Dan
decides to add 500M of modules to Debian, so Dan has to choose wisely. :)

If I read his original request correctly, then, it was for advice in
coming up with this subset of modules, not to decide what is "essential" 
and what is "not essential".  The problem is simply one of arriving at a
"best compromise" given the space constraints in the main archive.

    nSLUG       http://www.nslug.ns.ca      synrg@sanctuary.nslug.ns.ca
    Debian      http://www.debian.org       synrg@debian.org
[ pgp key fingerprint = 7F DA 09 4B BA 2C 0D E0  1B B1 31 ED C6 A9 39 4F ]
[ gpg key fingerprint = 395C F3A4 35D3 D247 1387  2D9E 5A94 F3CA 0B27 13C8 ]