Noding Basic SegmentStrings vs Noded SegmentStrings

[micycle1]

Is there any difference/advantage in passing a list of NodedSegmentString vs BasicSegmentString to a Noder (noder.computeNodes())?

[dr-jts]

Yes, there is. I have just noticed that the Javadoc on BasicSegmentString is not correct, so thanks for asking this question (I will fix it pronto).

The difference is:

• BasicSegmentString allows accessing segments, but does not support adding nodes
• NodedSegmentString allows accessing segments, and also allows adding nodes to the segments, and then splitting the segstring at the nodes to produce a set of sub-strings.

Noders are used for both detecting intersections and for computing noded arrangements. For detection it is fine to use BasicSegmentString. This avoids the (small) overhead of creating a SegmentNodeList for each string. But if full noding is required then NodedSegmentString must be used.

(Really NodedSegmentString could be a subclass of BasicSegmentString - not sure why this is not so. And perhaps the SegmentNodeList should be created lazily, which would reduce the overhead to the point where only NodedSegmentString would be needed).

[dr-jts]

The Javadoc for BasicSegmentString and NodedSegmentString has been improved.