Thanks for the replies.  The errors occur even though the layer I am 
editing is in the same projection bounds as the first layer open, and thus 
should be the same as the cosmetic layer (which also has the same bounds as 
the first table opened?).  I think I will try jacques' suggestion and 
create a map layer that is used instead of the cosmetic layer to add object 
into.

It may be that there is no problem afterall however.  The difference may be 
in the display of adjoining polygons when they do not share all the same 
nodes.  (It was late - I was not looking for lateral solutions).  The 
problem manifests itself when you look closely at the adjoining line 
between two adjoining polygons, one of which is split with the mapbasic 
program.  There is a 4 cm gap.  However, if you add a node to the adjoining 
polygon (that was not split), the gap disappears.  I have noticed this 
phenomena before.  The question is "If a gap or overlap appears that 
MapInfo cannot see, does that gap or overlap really exist at all??

One problem in eliminating this problem is that overlaying the nodes of the 
unsplit polygon with the split polygon does not give another node in the 
unsplit polygon. (Obviously did not recognise it as overlapping).  I was 
forced to do some creative work arounds to cut the split line back with a 
20cm buffer around the original object, select out all the objects from the 
original layer  that intersect with this line object, update their objects 
with OverlayNodes(obj,PLIneObject), then go about splitting the other 
polygon as per original plan.  Works fine now....

Some interesting phenomena on the way...
I have two previously adjoining polygons, of which I split one using the 
program written to split polygons with a line.  When you zoom i close along 
a adjoining boundary, they now appear to be separated by 4 cm.  When I 
 buffer one of the split objects by less than 5 cm while using the 
Australian Map Grid Projection, the resultant polygon is no different than 
the first.  If I make a buffer 5cm around the first polygon, you can see 
that it now overlaps the original adjoining polygon. Maybe it all in the 
preception.

Does not solve the issue of inaccuracies with the cosmetic layer, but hey, 
MapInfo could go the ArcView course and not allow you to display layers 
with different projections together at all - that would be a great 
solution.

-----Original Message-----
From:   Steve Wallace [SMTP:[EMAIL PROTECTED]]
Sent:   Wednesday, November 24, 1999 2:40 AM
To:     [EMAIL PROTECTED]
Subject:        Re: MI: errors in coordinates

Did you see Eric's response? The cosmetic layer will have the coordsys and
bounds of the first table opened. Make sure you create the mapper with the
one you want to control the cosmetic layer, then add the rest of your 
layers.

Let us know how this works -- it may be that there are other problems Eric
didn't realize.

Steve

At 01:52 AM 11/24/1999 +1000, you wrote:
>As it so happens, I am reprogramming some functions in a MapBasic program
>at present and I am encountering problems with the coordinate shifts that
>have been discussed recently.
>
>In the program, the line is drawn into the cosmetic layer, the polygon in
>the editable layer.  Both objects are extracted to variables for
>processing, where the polygon object is split by the line into two
>polygons.  The process seems to work fine (BTW the extractnodes function 
is
>a great new feature).
>
>However, on closer inspection, there is a problem.  When the split polygon
>objects are added back to the original layer, they have shifted about 4cm
>from the original object.  This will mean that they no longer adjoin the
>neighbouring objects.  The split also occurs about 4cm from the original.
>  In light of the recent discussions, it all seems to make sense, but it
>does not help the situation.
>
>?? Any suggestions??  Its 2 AM here so I am going to bed, but maybe
>  tomorrow I will try overlaying nodes with neighbouring polygons etc. to
>see if that works... Or maybe buffer the polygons by 4cm and erase the
>result with the neighbouring polygons.
>
>Robert Crossley
>Trinity Software
>10 Trinity Street
>Parramatta Park
>CAIRNS   4870
>AUSTRALIA
>
>Phone: 61-7-40314877
>Fax:     61-7-40314810
>email: [EMAIL PROTECTED]
>web: www.trinitysoftware.com.au
>
>
>----------------------------------------------------------------------
>To unsubscribe from this list, send e-mail to [EMAIL PROTECTED] and put
>"unsubscribe MAPINFO-L" in the message body, or contact [EMAIL PROTECTED]
----------------------------------------------------------------------
To unsubscribe from this list, send e-mail to [EMAIL PROTECTED] and put
"unsubscribe MAPINFO-L" in the message body, or contact [EMAIL PROTECTED]

Reply via email to