[geos-devel] Unclear objects lifetime and ownership issues inMonotone Chain components

[Obe, Regina]

Martin,

> This seems like it might be building GEOS to be a bit too specific to 
> its use in PostGIS, however.  I think there's lots of people who want to 
> use it in other environments.

Slightly off-topic - but have you been looking at how those evil C# people have been interpreting JTS?
http://sourceforge.net/project/showfiles.php?group_id=144924&package_id=159419&release_id=443988

Although I think they have a much easier time porting JTS code - since well C# doesn't have to worry about all that pointer arithmetic minutia and its syntax is much closer to Java (e.g. they have ArrayList too) so is a closer one to one code compare with JTS.  In fact - its almost identical code except for some casing choices to conform to C# standard here and there and use of IEnumerator vs iterator.  Scary close.

I, feeling somewhat attached to the evil empire - can't help but point out these irrelevant similarities :)

I'm still working on the proof of existence theorem - although I am leaning toward the direction of a meta language
that no one actually programs in but that enforces a certain level of rigidness such that (ala Microsoft's IL where a VB.NET compilation to IL is almost as efficient as C# to IL or C--) 

Valid Transform(JTS, TSM) -> implies - Transform(TSM, GEOS) is efficient
Valid Transform(GEOS,TSM) -> implies - Transform(TSM,JTS) is efficient
Valid Transform(NTS, TSM) -> implies - Transform(TSM,JTS) , Transform(TSM,GEOS) is efficient

By the simple fact it chokes you know there is no way in gods hell of making it efficient without significant rewrite or tinkering with the transform function.  What exactly that magical function looks like (if it exists) I'm still stirring in my head (I suspect it involves cataloging the various patterns of programming in each).  

The problem I have with things like GCJ conversions (talking from the side of my mouth since I don't quite understand them) is 
they just guarantee at best the code will run but not that it will run efficiently in the new environment.  I also tend to think the task is much more simple if you limit the scope of possibilities whereas a bytecode full blown thingy has to consider everything.

Thanks,
Regina














-----------------------------------------
The substance of this message, including any attachments, may be
confidential, legally privileged and/or exempt from disclosure
pursuant to Massachusetts law. It is intended
solely for the addressee. If you received this in error, please
contact the sender and delete the material from any computer.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.osgeo.org/pipermail/geos-devel/attachments/20080926/d781f892/attachment.html