(j3.2006) Spreadsheet from meeting 167

[Malcolm Cohen]

I take slight exception to the word "veto" to indicate subgroup discussion that 
ended up with the proposal not being supported.  J3 normally uses the procedure 
of having a subgroup act as a filter for proposals (whether to do with new 
feature suggestions or not).  "was vetoed" -> "failed to get support" (or just 
"failed").

>In 2004, we put some focused effort into evaluating proposals for the
>2008 work plan.  These were all discussed in plenary sessions (see item
>1.4 in the minutes, 04-272r1).

Right, and given the large number of proposals, one might think that plenary did 
not discuss them in great detail...  OTOH I was not there, so what do I know...

...but I do know that "undecided" was not an option, so "like" really means 
"don't object"...

>I reviewed
>the spreadsheet 04-265r1.xls, and noticed that the following proposals
>had overwhelmingly positive sympathy,

Yes, according to the minutes they had positive sympathy for
  "Should JOR spend time on the proposal?"

That is some distance from whether they should actually go into the standard.

> but were not added to the 2008 work plan,
>mostly because it was thought that other work was more desirable.

In some cases, but hardly all.

>Is there interest to pursue any of the ones that have not yet been
>reconsidered?

I don't think that resurrecting old ideas we already considered (sometimes more 
than once) and decided not to do is a particularly good approach to identifying 
how to improve the language.  The spreadsheet you are looking at is just a 
snapshot in time (and not even the most uptodate version of it - it was added to 
at later meetings) of what the committee thought was worth investigating - in 
every case, further investigation was actually done and either the feature was 
dropped, we did something else instead, or it got to WG5 and failed to win 
support there.  In a few cases this was due to resource limitations rather than 
a subsequent lack of support, but even there the case for reconsideration is 
rather weak.

Languages can get too big you know.  Every new feature should start with minus 
100 points...

Also, many of the "not yet been reconsidered" items are full-blown new features, 
not in any sense "wart removal".  No matter how small they might appear at first 
sight (none of the even partly-desirable ones appear so to me!), these really 
need a strikingly good case to be made for them if there is to be any 
reconsideration, where "strikingly good" means "so overwhelming we are willing 
to delay the standard and delay the compilers implementing other features".

Cheers,
-- 
................................Malcolm Cohen, Nihon NAG, Tokyo.