Subject: [boost] Another GGL review
From: Jose (jmalv04_at_[hidden])
Date: 2009-11-16 13:57:51
YES, the library should be accepted (only show stopper is that it has
to include the non-cartesian coordinates so that the library can be
useful as is without resorting to additional extensions)
- What is your evaluation of the design?
The design is good and has taken input in multiple iterations from the
community. Also, it's a very strong point that the library provides
open benchmarks with multiple libraries so that the performance
standing of the library can be evaluated.
The week point is that the library does not satisfy the needs of some
application domains (e.g. VLSI domain) where floating point types are
It's unfortunate that this review has been muddled by the recent
boost.polygon review. This is a topic beyond this review and specific
to the boost review process.
I think in this case it should be determined by at least two experts
whether it is possible to have a generic library that can satisfy the
multiple application domains, and if yes, then take GGL as the base,
update the design and incorporate boost polygon algorithms as
- What is your evaluation of the implementation?
The library should incorporate the key extensions that make it useful
right away, i.e. non-cartesian coordinates. Not including these will
hamper the usefulness and adoption of the library. An example where
this happened was the GIL library, it had a wonderful generic design
but a key numeric extension was not included in the Boost library. As
a result, the adoption has been very limited and the key extension not
updated or bug-fixed, with some algorithms being ported and some
libraries being built on top, but not wide use.
I think the problem with the PROJ dependency could be handled similary
to the MPI library, that provides bindings to the C library.
- What is your evaluation of the documentation?
The docs have to show how it can be used in multiple domains, GIS,
CAD, VLSI, .. Same with the examples.
- What is your evaluation of the potential usefulness of the library?
Very high. A broad library that could be a great match for Boost.
- Did you try to use the library? With what compiler?
Yes. g++ 4.4.1 on x86_64
Did you have any problems?
YES, because the non-cartesian extension was not part of the library.
- How much effort did you put into your evaluation? A glance? A quick
reading? In-depth study?
10+ hours. Studied both GTL/GGL papers/presentations presented in BoostCon09.
- Are you knowledgeable about the problem domain?
Yes, the GIS application domain but not core geometry and other
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk