[SnapPea-planning] Finishing SnapPea 3.0

[Bill Thurston]

I'm  enthusiastically in favor of this project, but I'm of course very  
biassed and for that reason not necessarily in the best position for  
seeking NSF funding.  It shouldn't be hard to make a strong case. If  
NSF is constitutionally unsuited (or even if its suited) Nathan's  
other funding suggestions sound  good:
> ... One might have better luck with one of the smaller foundations  
> like AIM or Clay or Jim Simon's new Math/Physics thing.  Bus Jaco  
> used to be pretty tight with the head of AIM, for instance.
especially in light of Dylan's comment (I didn't know)
> That's a good idea.  John Morgan is leaving Columbia to be the
> founding director of the Simons institute; I could approach him.
since Morgan would almost certainly appreciate the significance of the  
project.
> 	1. The most obvious task is to continue adding old features into  
> the current OSX application until it can do everything the old one  
> could, but I'd like to ask around first.  Dylan, for example, would  
> like it to work on linux.  Are there other requests?
Yes the first priority should be getting the OSX version up to speed.  
Still it may be worth thinking ahead toward new capabilities,  
especially because it might make it easier to sell to NSF and others.   
Here are some quick thoughts, things that have nagged me for a while:
	a. It would seem natural to have a "good drillings" view that, at  
least for one cusp, would trace out the boundary
of the region in Dehn surgery  where tetrahedron in (one r  
triangulation | at least one from a set of triangulations) are  
positively oriented. Of course there are lots of additional  
enhancements from this e.g. displaying identification
information for the most significant Dehn fillngs.
	b. The dirichlet domain command should be more robust,  either by  
higher accuracy arithmetic or a smarter algorithm that
isn't killed by (I assume) floating point inaccuracy.
	c. An integrated hyperbolic viewer along the lines of maniview (from  
geomview). Or, as a start, just putting the larger
SO(3,1) group into the Dirichlet domain viewer.
	d.  Identifying Dehn-drilling axes within the dirichlet domain view  
and/or the cusp view.

These are things that don't seem too far out of the way from the  
current snappea. There's other longer-range stuff that would be cool,  
but probably take more thought and more work, such as a "wrapping up"  
search to look for supergroups of finite index (inverse to the finite  
covers command) and a "detriangulate" command that goes from a 3- 
manifold whose fundamental group has parabolic generators
to an embedding in S^3 as a link complement ... or even, drills stuff  
out and then constructs a link ...or ... or ... lots of possibilities.

and ... an iphone version.

	Bill
>
>
> 	2. There seem to be two obvious routes for funding: find someone  
> who wants to contribute unallocated grant funds, and/or apply for a  
> grant for the purpose.
>



> Nathaniel,
>
> Great to hear that you are interested in working on SnapPea.   Let  
> me give some quick answers now, with perhaps more once I've thought  
> about things a bit.
>
>> Two things need to happen first, both involving the current users  
>> of SnapPea:
>>
>> 	1. The most obvious task is to continue adding old features into  
>> the current OSX application until it can do everything the old one  
>> could, but I'd like to ask around first.  Dylan, for example, would  
>> like it to work on linux.  Are there other requests?
>
> In my opinion, a Linux version is probably the greatest need after  
> one that works on OS X.  Longer term, a new version that runs on  
> Windows might be a good idea.   I don't know whether the current  
> Windows port
>
> http://home.att.net/~Manoharan/SnapPea/snappea.html
>
> is/will be maintained (the source code is not posted, unfortunately)  
> and it's missing some of the graphical features of the original Mac  
> OS version (IIRC no pictures of Dirchlet domains).  Note: the  
> Windows port does run find under Wine on Linux and OS X/Intel and I  
> have used it as a stop-gap on those platforms.
>
>> Who should I consult regarding ideas about the requirements for the  
>> next version of SnapPea
>
> One additional person you should definitely consult is Marc Culler (culler at math.uic.edu 
> ), to whom I've taken the liberty of cc'ing this message.   Marc is  
> currently working on a new, better, Python interface to the SnapPea  
> kernel, using Pyrex/Cython instead of writing the wrappers by hand.   
> He has some of the graphical features working, including Dirichlet  
> domains, a very nice link editor, and some preliminary horoball  
> stuff.    See some pics here:
>
> http://dunfield.info/temp/screen-part.png
>
> He used the cross-platform Tk library for this, so it could form a  
> basis for a new GUI, but that is not (I think) his motivation at the  
> moment.
>
> More broadly, Saul Schleimer is heavy user of SnapPea who might have  
> some ideas on this.
>
>
>> , or grant funds for the purposes?
>
> The most reasonable person to contact at NSF is Joanna Kania- 
> Bartoszynsk.  I'm not sure how much the NSF likes to fund this sort  
> of thing, since, while it is clearly needed, it may seem to them  
> like maintenance, which they try to avoid, I think.  One might have  
> better luck with one of the smaller foundations like AIM or Clay or  
> Jim Simon's new Math/Physics thing.  Bus Jaco used to be pretty  
> tight with the head of AIM, for instance.
>
>> Does anyone want to participate in the process of deciding what  
>> should be done?
>
> Yes.
>
> 	Best,
>
> 	Nathan