Re: [boost] [polygon] tutorial done and library ready for release

Subject: Re: [boost] [polygon] tutorial done and library ready for release
From: Simonson, Lucanus J (lucanus.j.simonson_at_[hidden])
Date: 2010-05-24 12:31:45

Dean Michael Berris wrote:
> I just took a look at this having some interest in using the polygon
> library as a user. Please take these first impressions with a grain of
> salt because I didn't participate in the review. That said here's some
> of my thoughts (unfiltered):
> 
>   * You use quite a lot of jargon that's specific to the chip design
> field. I'm not sure if normal C++ developers are your target audience,
> but a tutorial on how to do VLSI layout might not be the best. If you
> still really want to use that domain, then you might as well explain a
> few things before you start diving into the rationale of why you'd
> want to check your chip layouts using a program.
> 
>   * I understand that there are already industry standards when it
> comes to chip layout and other things, however I find just the normal
> problems with polygons solved easily would be a huge and powerful
> tutorial. Things like what the minimum dimensions should be for a set
> of smaller polygons to be contained in a single rectangle
> (minimization problem, maybe for the garment industry) or the maximum
> number of triangles of a given dimension can be fit in a given polygon
> (maximization problem, maybe for the construction industry). Naive
> question: does it have to be specific to VLSI chip design or did it
> just come out in the review?

If by "it" you mean usage of the library, then no.  The library generally handles polygons with integer coordinates.

I'm not sure either of your suggestions turns out to be easy using the library.  You are suggesting difficult things in their own right that would make good extensions to a geometry library.

>   * Given that I don't see where this tutorial fits in the whole
> scheme of documentation, this tutorial seems to rely on a lot of
> missing context. Is that intentional by virtue of placement, or is
> this supposed to be a stand-alone tutorial?

My intention was that the tutorial be self contained and that the documentation should help the reader understand the usage of the library in the tutorial.

> I haven't really dug into the details of the tutorial and I'll trust
> the experts when it comes to the details of the solution. However, I
> think it would be powerful to show how the library can be used for
> "normal" everyday and considerably non-trivial problems that are made
> trivial by the library.
> 
> Thanks and I hope this helps!

It does help.  I was concerned that the tutorial would be unapproachable for people not familiar with the VLSI domain.  Suggestions on a simpler tutorial would be helpful.

Thanks,
Luke