Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • Aug 01 2018 17:48
    @jnh5y banned @matrixbot
Martin Davis
@dr-jts
@jnh5y any reason for keeping the 1.17.1 docs around?
I'm always a bit leering of taking on the administrative debt of cleaning up old artifacts
and also it can be a pain having doc URLS with ver numbers - makes it hard to have a stable link
(for instance, the link in the JTS README doesn't point to the new docs)
James Hughes
@jnh5y
ah! good catch on the link
I like the idea of having previous versions of documentation around
that said, I am not committed to it or to my particular approach
Martin Davis
@dr-jts
@jnh5y how about putting the latest ver under javadoc and keeping older ones in versioned dirs?
James Hughes
@jnh5y
I like that idea; I'll make that change now
Martin Davis
@dr-jts
great
James Hughes
@jnh5y
done and done! Thanks for the suggestion
James Hughes
@jnh5y
nice blog post; it did take me a little bit to get that the yellow areas were the output of the algorithm (and hence it was taking me longer than necessary to get that the simplification example was as shockingly bad as it is:))
Martin Davis
@dr-jts
hmm, good point. The TestBuilder does now have the ability to provide a legend - I should use that!
image.png
James Hughes
@jnh5y
Beautiful.
Only thing you could do from there is rename "A" to something indicated that it is excluded
Martin Davis
@dr-jts
image.png
James Hughes
@jnh5y
+1
If you keep up this ratio of words to line of code go forward, I think you'll be writing an applied computational geometry / topology textbook
Martin Davis
@dr-jts
or at least The JTS Book!
James Hughes
@jnh5y
I'll take two! Signed please!;)
Vicky Vergara
@cvvergara

The fix consists of about 4 lines of actual code.

I heard once a story about a man that charged $1001 to change a screw,. Why $1001 to only change a screw?
$1 the cost of the screw, and $1000 for knowing which screw to change.

@dr-jts impressive

Martin Davis
@dr-jts
@cvvergara Thanks. I know that story too... and it often applies in software development!
alongsang
@alongsang
1.png
May I ask some questions? Does jts support this function,can't find it
Martin Davis
@dr-jts
Yes, you just have. :)
Nope, JTS doesn't support Reshape. That seems like more of an interactive function, I think?
You can probably simulate it by simply closing the "shaping" line and then using union or difference
alongsang
@alongsang
got it.Thanks for answering
Martin Davis
@dr-jts
Hey devs... I was reminded about this long-standing issue with the JTS API semantics: locationtech/jts#159
What are people's thoughts about changing this?
James Hughes
@jnh5y
I think this is reasonable. @jiayuasu do you know if Sedona would be impacted by a change to Geometry.equals? I think that keeping hashcode the same ought to be sufficient
Martin Davis
@dr-jts
I haven't done it before because it's one of those changes which is technically simple but has potentially wide impact. Although really the topological equals() method is not very useful, so should not be used too often.
James Hughes
@jnh5y
I wonder if we should put it on a list of minimum changes that we'd want in a JTS 2.0.0?
Martin Davis
@dr-jts
Well yes, definitely. But since JTS 2.0 has no timeline, would be good to get it out in JTS 1.x
James Hughes
@jnh5y
I'm interested in a JTS 2.0 so that we start thinking about using the new Java record classes for Point/Coordinate;)
Jia Yu
@jiayuasu
@jnh5y Thanks for asking. I think Sedona won’t be affected by the change in equal. In Sedona 1.0, we no longer use that function.
James Hughes
@jnh5y
@jiayuasu glad to hear it! Also, thank you again for upgrading Sedona (soon to 1.0.0) to the latest JTS!
Jia Yu
@jiayuasu
Thank you all for the great work of JTS 1.18! :-)
Martin Davis
@dr-jts
This is an interesting one: locationtech/jts#663
More work required on the WKBWriter and WKBReader!
Julius R Friedman
@juliusfriedman
Beg of work :p
as previously stated by @jiayuasu , & @dr-jts , Thank you for your library and efforts! May your library stand the test of time for all such time to come!
James Gill
@jagill
^^^ Geometry.equals(Geometry) should be the Java object equality. It's too confusing otherwise.
Just a heads up: I've found some TopologyExceptions in the OverlayNG. I'm working on getting minimal examples.
James Gill
@jagill
Also, I'm running into the requirement that "inputs must be homogeneous". I could prune any lower-dim entries of a GeometryCollection, but that involves a correctness change for my users. I could separate by dimensions and use the distributive properties of intersection/union, but maybe that would better be done as a helper function in JTS?
Martin Davis
@dr-jts
Thanks for the vote about changing Geometry.equals(Geometry). Seems like that should happen.
As for issue of overlay handling GeometryCollections, I am working on that now. As you say, it involves using the distributive/associative properties of the various overlay operations.
James Gill
@jagill
Great! Looking forward to it :).