[tei-council] FR nuncles: new element tei:objectType

[Sebastian Rahtz]

On 7 Jan 2011, at 14:41, Lou Burnard wrote:

> What is the problem we're trying to solve here?

two problems

  a) use of technology which assumes *nix, and which no-one under 50 understands
  b) lack of a rigorous test procedure
> 
> Where is the evidence that the current build process is seriously broken?

do a quick poll, and see how many people can build P5 from scratch
on their desktop

> 
> I hypothesize that one problem with it is seen to be that too few people 
> have direct experience/understanding of it. So while introducing an 
> entirely different system would level the playing field (by ensuring 
> that nobody had knowledge of it) it would hardly advance the status quo 
> very much.

but if the new system was already more familiar to more people, this
would not apply. and if it was automated, no-one _needs_ to understand it
how to run it. 

> 
> I also hypothesize that there is a perceived need to enhance the current 
> environment to do continuous validation better. But, as I said before, 
> this is pointless until we have better test materials. And if we had 
> better test materials then we'd see immediate benefits in our current 
> build process, with no need to go to a completely new environment.


I don't buy these arguments. 

a) our current testing is already pretty effective,
(certainly compared to P4!). It regularly catches typos or brainlosses perpetrated
by you or me or syd or an other.

b) to run the current testing  is manual, tedious, local, and poorly documented. Simply
automating that so that any commit triggered a change and report on a web site which
anyone could see would be hugely beneficial.
--
Sebastian Rahtz      
 Information and Support Group Manager
Oxford University Computing Services
13 Banbury Road, Oxford OX2 6NN. Phone +44 1865 283431

Sólo le pido a Dios
que el futuro no me sea indiferente