[MAPSERVER-USERS] Problems calculating MAXSCALES

[Steve Lime]

I've not tweaked anything since that ticket was closed. My take on this thread is that:

1) With the same extent 4.8.3 and 5.x output different scale computations. This is true. There
were inconsistencies in what an "extent" meant to MapServer in older versions. That was standardized
in 5.0.

2) ESRI desktop scale computations do not match MapServer computations. Again true probably because
the number of pixels per inch value used in each package is different. The underlying pixel/extent model
shouldn't matter I guess.

In this case I think Rob is seeing the effect of 1) and if so would need to adjust his layer MAXSCALE and
move on.

Perhaps 2166 does need to be reopened but I don't think that has to do with the problem here which seems
to be an artifact of changing versions.

Steve

>>> On 8/13/2008 at 3:45 PM, in message <48A347DC.4040800 at swoodbridge.com>, Stephen
Woodbridge <woodbri at swoodbridge.com> wrote:
> Steve,
> 
> I have not checked this recently, but I think there is a fundamental 
> principal that if I request a map at SCALE=20000 and some X,Y that the 
> map rendered includes all LAYERs that are in the 20000 SCALE +- 1.0 for 
> rounding. This was the bug that I reported in
>     http://trac.osgeo.org/mapserver/ticket/2166 
> and I hope that that has not regressed. Was this bug fixed after 5.0.2 
> that Rob is working with.
> 
> Rob,
> 
> Have you tried a later build of mapserver like 5.2 and see if the 
> problem exists there?
> 
> Best regards,
>    -Stephen Woodbridge
>     http://imaptools.com/ 
> 
> Steve Lime wrote:
>> It's likely the desktop product you're using uses a different pixel
>> model than MapServer's center-to-center model. The code in 4.8.3 was
>> incorrect so I'm not sure what we can do. A WMS extent simply isn't
>> equal to a MapServer extent. The MapServer WMS and WCS code account
>> for the difference between the MapServer and OGC models by
>> transforming between models (shrink or expand by .5 pixels).
>> 
>> My opinion is that you adjust the reference scales in your mapfile
>> based on the now correct output from 5.0.2.
>> 
>> Steve
>> 
>> 
>>>>> On 8/13/2008 at 9:07 AM, in message
>> <f3b73b7d0808130707r2df2aeeg826ed5f07dbc033a at mail.gmail.com>, "Tamas
>> Szekeres" <szekerest at gmail.com> wrote:
>>> 2008/8/13 Rob <gis at vanbooth.com>:
>>>> If I change the (width-1) to be width, then I get scale figures
>>>> that correspond to what I would expect, and what the ESRI GIS
>>>> Clients hitting my WMS service seems to think that the scale
>>>> should be.  From the dates in TRAC it certainly looks like this
>>>> was changed between 4.8.3 and 5.0.2.
>>>> 
>>> I'd prefer back out r6529 and go with a double based calculation 
>>> instead of it. However I'm not sure how this will affect the
>>> original problem, so you should reopen this bug by pointing out the
>>> problems with the current scale calculation you encountered.
>>> 
>>> Best regards,
>>> 
>>> Tamas _______________________________________________ 
>>> mapserver-users mailing list mapserver-users at lists.osgeo.org 
>>> http://lists.osgeo.org/mailman/listinfo/mapserver-users 
>> 
>> _______________________________________________ mapserver-users
>> mailing list mapserver-users at lists.osgeo.org 
>> http://lists.osgeo.org/mailman/listinfo/mapserver-users