RFC: Change LWCURVE to LWCIRCSTRING (was Re: [postgis-devel] STYLE)

[Mark Cave-Ayland]

Paul Ramsey wrote:
> On Fri, Dec 19, 2008 at 3:52 AM, Mark Cave-Ayland
> <mark.cave-ayland at siriusit.co.uk> wrote:

>> BTW I noticed your r3450 commit, and this got me thinking. One of things
>> that really trips me up at the moment is LWCURVE - why oh why was it not
>> called LWCIRCSTRING??! This tends to trip me up a lot, since SQL-MM defines
>> a curve as a LINESTRING or a CIRCULARSTRING and so makes the logic a lot
>> harder to follow when trying to map SQL types to C types.
> 
> Hmmm, working this week, I was surprised to see my attempt to use
> LINESTRINGTYPE fail :)  Perhaps it is in that tradition.  Not that I
> would mind seeing the internal names harmonized with the standard
> names, for cognitive assonance.
> 
> P.

Yeah - for me LINETYPE vs. LINESTRINGTYPE isn't really an issue since 
it's clear when you come to read the code that when you see LINETYPE, 
the geometry involved is a a line.

The catch with LWCURVE is that the SQL-MM spec defines a curve as a 
superclass of LINESTRING and CIRCULARSTRING; and so when looking at the 
code, there is immediate confusion since the name implies that the type 
is a collection when in fact it is not. Changing this to LWCIRCSTRING 
would both tell you immediately that its associated text is 
CIRCULARSTRING, and that it is not a collection. Note that the 
LWMULTICURVE (MULTICURVETYPE) already correctly uses this convention. 
since it can contain curve elements which can be either a CIRCULARSTRING 
or a LINESTRING.

I see that it was Mark L that originally committed the code. Mark, do 
you have any comment on this?


ATB,

Mark.

-- 
Mark Cave-Ayland
Sirius Corporation - The Open Source Experts
http://www.siriusit.co.uk
T: +44 870 608 0063