(j3.2006) (SC22WG5.3605) [ukfortran] Ballot processing and SC22 meeting

Malcolm Cohen malcolm
Fri Oct 3 01:46:15 EDT 2008



John Reid wrote:
> I do not expect WG5 to discuss the papers referenced in the first two 
> sections of N1741, Editorial repairs and Technical repairs, since this 
> level of detail is normally delegated to J3.
I'd like to disagree with this assertion.  In the past WG5 has taken its 
lead from J3 at this level, but the draft CD is WG5's document and it is 
WG5's responsibility to take ownership of any changes.  For example, 
during F2003 development we voted up minor corrections as a block; I 
believe these were absent any so-called "technical corrections" though 
some typos might have been fixed that had some trivial technical effect.

In particular, the great majority of "technical fixes" are emphatically 
*NOT* typo fixes, but actual technical CHANGES in some cases from second 
thoughts about how some things were supposed to work and in other cases 
actual changes in previously-agreed technical content.

Even for the so-called editorial changes, despite the best efforts and 
intentions there could be defective edits or edits with technical 
effect; in order to encourage people to have a look at what they are 
agreeing to I'd like to see that paper block-voted.  If indeed we got it 
right (or people don't care) it will take 5 minutes at most.  If we got 
it wrong it will result in the loss of a lot more than 5 minutes!

The "technical changes" paper needs *SERIOUS* attention by WG5 in my 
opinion.  These are not "oh look a typo".  Some of them are "oops we 
forgot about this and that interaction" ... and the chosen solution 
needs to be agreed!  Others of them are "we changed our mind on this 
one" - in some cases deleting some part of a feature ... and such change 
of content certainly needs to be agreed.  There is also the cumulative 
effect of multiple minor technical changes on a feature; individually 
these might well be close to insignificant but as a whole they can 
certainly change the whole character (and thus desirability) of a 
feature.  It's possible that we did such a good job on this paper that 
WG5 will find nothing wrong and just vote the whole thing up as a block 
(though I'd be a little surprised if we got all of these right, to be 
honest); and if there is any technical change there that does not have 
WG5 agreement then yes it will take longer but it needs to take longer 
and achieve agreement if not consensus.

I will note that I have not produced the two documents under discussion 
here, viz 08-019r1 and 08-020r1.  Sorry about that; I have started, but 
I am under some pressure on the day job as well so these are likely to 
take another couple of weeks.  If they are as trivial and 
uncontroversial as John implies that won't be a problem; to the extent 
that they are not it underscores the need to WG5 to agree to them.

...
> I think we may need to make changes in the structure of the early part 
> of our draft standard.
Sure.  We have already made substantial progress towards making our 
standard fit the guidelines, so I am not too worried about this one.  
The structural change you suggest is minor compared to the major one of 
actually adhering to ISO guidelines re terms we define!  (And we are 
still not quite finished doing that, though it is much better than before.)

Cheers,
-- 
.....................Malcolm Cohen.





More information about the J3 mailing list