[sword-devel] Getting stuff done (Re: External links)

[Matthew Talbert]

On Wed, Nov 26, 2008 at 4:15 AM, Peter von Kaehne <refdoc at gmx.net> wrote:
> Chris Little wrote:
>> then updating the module repeatedly until we
>> finally find a good markup standard, we can get things right the first
>> time
>
>
> That is a totally wrong approach on so many levels. It holds us back all
> the time, makes for module production times in years rather than weeks,
> drives good people away and holds the project back.
>
> All good frontends are capable of updating modules. It is easy.
>
> Automatic update would be easy to add to most frontends (and could be
> toggled for those behind dialups or under oppressive regimes)
>
> If someone brings us a good text to include, then we should be able to
> say - "here, it works" and then gradualy improve it to make it better,
> richer and more conforming to whatever lofty standards we want to achieve.
>
> By doing latter we create
>
> a) momentum
> b) gain additional helpers who remove bugs and update (instead of
> driving them away)
>
> If Wikipedia would rigidly enforce its increasingly arcane standards on
> how a good article should look like by simply deleting valid, but less
> then perfect articles it would be empty. Instead they put articles out,
> tag them as unclean, stub, poor style or whatever and allow others to
> work on them.
>
> We should look at their model and not that of paper publishers who have
> exactly one opportunity to prove themselves.
>
> Peter

I completely agree with this entire email. I would suggest that
front-ends do not actually install the modules, but notify the user
when updates are available.

I would just like to point out again that there is very little
transparency in how bugs are updated and fixed in modules. Neither
does it appear to happen quickly. For example, someone reported a bug
(several bugs) in the ISBE module to the sword-support list 4 days
ago, and no one has even replied. I do not know what to say, because I
do not know how bugs like those reported could/should get fixed.
Speaking from my experience, my satisfaction with software (whether
open source or commercial) depends in large part on the software
creators responding quickly to deficiencies in their software (to the
end user, there isn't much difference between the software and the
modules). In this case, the sword project is really lacking and
especially when it comes to module issues.  To the end user (that
would be me in some cases) it would appear that simple issues in
modules would require simple changes and an update. I've rarely seen
this happen.

Matthew