[sword-devel] Wycliffe modules and beta testing
chrislit at crosswire.org
Sun Jan 18 14:23:21 MST 2009
Peter von Kaehne wrote:
> My understanding of a discussion with Troy was that we are not supposed
> to do beta testing on the Wycliffe modules as they are kind a separate
> category - supplied and maintained by the organisation - and we only
> host them
We need to test them to the extent that we need to know what failings
exist in either our conversion process or in frontends. For the most
part, we aren't testing the content quality--I pretty much know where
the SF->OSIS problems are and have corrected them myself.
It may have been our hope that Wycliffe maintain their own content, but
that has not been the case, partly because we haven't set up a process
for them and partly because (I believe) it's not their desire to manage
content on our servers (and I haven't heard anything to the contrary;
but correct me if I'm wrong).
> If this understanding is correct could we either move them in bulk into
> the main repository or- in my view vastly preferable - create a vendor
> repository for Wycliffe and move them there?
If WBT wants to maintain its own repository, then they can do that. I
don't think it serves them to move offsite and essentially lose
exposure, and I don't think it would serve us to maintain additional
> The updates to the module manager will make this much easier (by
> automatically binding in other repositories) and I think we are doing
> ourselves no favour by keeping them either in beta or in main. As I
> remember the testing which was in full flow came to a grinding halt when
> people started looking at this list.
The Wycliffe modules should essentially mirror one another in terms of
errors and issues. If we test one or two of them adequately, then the
others should show the same issues and need not be tested.
We set aside the Wycliffe modules from testing (both literally and
figuratively) at your request and that occurred significantly prior to
the cessation of testing. So to assign any connection between their
presence and no volunteers doing any testing would be an error of fact.
More information about the sword-devel