create/destroy name change proposal

[Mark Miller]

I hadn't considered auditing the entire API for consistency in naming
but I agree it could be improved. The various abbreviations for 'child'
you mention seem of the same ilk as my proposal.

So, I guess I may have wondered onto a slippery slope! We do have bigger
fish to fry then naming conventions. On the other hand, its a shame if
we don't make the tiny effort required to simplify the API where we
easily can and I think consistency in naming one such step. On the other
hand, I was aware when I proposed the name changes related to
remove/destroy operations that a larger audit of the API for consistency
in naming would probably be useful.

I am not sure what to do. I could easily be convinced to back off on my
proposal seeing as how it apparently results in consist-ifying only a
portion of the inconsistencies in the names. I could also be convinced
that we really should agree to a full audit of the API for consistent
names (as well as relative to other ITAPS APIs) sometime before the end
of this FY.

At this point, I don't think it hurts to at least get all the name
changes on the table so we can at least see 'em. So, I would ask you,
Jim, to send me a list of old --> new function name changes on the
'child' issue as you see 'em so that I can maybe combine them with my
proposal and draft a larger proposal on this topic.

Mark

On Thu, 2010-05-06 at 15:39, James Porter wrote:
> On Wed, 2010-05-05 at 10:16 -0400, Carl Ollivier-Gooch wrote:
> > On 10-04-29 09:30 PM, Mark Miller wrote:
> > > Proposal:
> > > ---------
> > > Change the name(s) of a few iMesh functions to make the API a little
> > > more consistently named. For convenience, I list below the names of many
> > > fuctions that add/remove or create/destroy things.
> [snip]
> 
> While we're on the subject of renaming, I think we should look at the
> consistency in abbreviations. For instance, "child"/"children" is
> spelled three different ways in iMesh: "iMesh_isChildOf",
> "iMesh_getNumChld"/"iMesh_addPrntChld", and "iMesh_getChldn". Likewise
> with "parent(s)". In most of these cases, we're only saving a couple
> characters (at most 4 for "iMesh_getNumChld"), at the expense of
> memorizing what abbreviation is used where.
> 
> I think we could just remove these abbreviations entirely, especially
> since the unabbreviated versions would by no means be the longest
> function names in iMesh.
> 
> Maybe it would also be worth looking at the iMesh_(get|set)Arr*Data
> functions. It looks like these are the only places in iMesh that still
> use "Arr" to refer to entity arrays (everywhere else uses "EntArr").
> 
> It would probably also be worth looking at naming consistency for iGeom,
> but that's a whole other can of worms.
> 
> - Jim
-- 
Mark C. Miller, Lawrence Livermore National Laboratory
================!!LLNL BUSINESS ONLY!!================
miller86 at llnl.gov      urgent: miller86 at pager.llnl.gov
T:8-6 (925)-423-5901     M/W/Th:7-12,2-7 (530)-753-851