MS-RFC-27: Label Priority

[Ned Harding]

quicksort is not stable so within the same priority class you could get
different orderings depending on the set of labels.

Daniel's multiple cache support would seem to be stable, which is great.
If for any reason people wanted to use a sort instead, lets make it a
stable sort?

Overall I am in favour of this RFC and it will be a good addition.
Would it be possible to have some pre-render test as items are getting
added to the cache that we already know of a collision at this or a
higher level, therefore we don't cache it at all?  Something like this
could save a lot of memory for dense layers.

Ned.

-----Original Message-----
From: UMN MapServer Developers List [mailto:MAPSERVER-DEV at LISTS.UMN.EDU]
On Behalf Of Frank Warmerdam
Sent: Tuesday, May 22, 2007 1:47 PM
To: MAPSERVER-DEV at LISTS.UMN.EDU
Subject: Re: [UMN_MAPSERVER-DEV] MS-RFC-27: Label Priority

Daniel Morissette wrote:
> I have created RFC-27 that defines a simple mechanism to handle label 
> prioritization in MapServer:
>   http://mapserver.gis.umn.edu/development/rfc/ms-rfc-27
> 
> Let's take a day or two for comments and then I'll propose that we 
> vote to adopt this RFC.

Daniel,

Can you comment on why you use a multiple cache approach instead of
doing a quicksort on a single cache based on the priority value?  A
quicksort is order n-log-n which should be a very modest performance
impact.

Also, is the priority something that can come from an attribute?  I
think i'm asking if there is a corresponding PRIORITYITEM though I'm not
sure.

Lastly, should the RFC touch on the corresponding mapscript changes?

Best regards,
-- 
---------------------------------------+--------------------------------
---------------------------------------+------
I set the clouds in motion - turn up   | Frank Warmerdam,
warmerdam at pobox.com
light and sound - activate the windows | http://pobox.com/~warmerdam
and watch the world go round - Rush    | President OSGeo,
http://osgeo.org