savi-developers — SaVi developers - Tcl/Tk/C expertise.

[savi] Getting Mini-SaVi to run: SaVi with network traffic

[<llo...@us...>]

SaVi is now also a constellation network simulator!

Putting SaVi together with mini-net to do network simulation of an intersatellite mesh was described in:

Zhu Tang et al., Mini-Savi: Realistic Satellite Network Simulation Platform Based on Open-Source Tools, Proceedings of the Fourth International Conference on Frontiers of Computers and Communication Engineering (FCCE), Xiamen, China, January 2023.

https://www.researchgate.net/publication/373789005_Mini-Savi_Realistic_Satellite_Network_Simulation_Platform_Based_on_Open-Source_Tools 
http://dx.doi.org/10.1109/FCCE58525.2023.00013

the code used (a customised SaVi 1.5.1 talking to network simulation) was put on github a year later, in March 2024, and, a year later again in April 2025, I've finally just noticed that that public code exists.

So far, I've been able to build and run Mini-Savi. There are a couple of modified Iridium scripts that hand over topology from SaVi to a python mini-net that runs frrr (isisd and zebra routing code).

My comprehensive, detailed, and easy-to-follow installation instructions are at:

https://github.com/zhutang/mini-savi/issues/1 

I hope this proves to be interesting and useful.

Noted on SaVi's papers page, along with the still-private intersatellite link modifications:
http://savi.sf.net/papers/

Recent SaVi betas have added some map lines and Tcl 9.0 compatibility
http://savi.sf.net/beta/

best

Lloyd Wood 
SaVi satellite constellation visualization 
http://savi.sf.net/ 

[savi] On SaVi with Tcl/Tk on Mac OS X - trials and tribulations

[<llo...@us...>]

In the course of making SaVi http://savi.sf.net work better on a Mac,
I've been looking further into the state of Tcl/Tk on the Mac,
which is... messy. So, I'm making some notes on this.
Corrections to my errors and understanding would be appreciated.

https://www.python.org/download/mac/tcltk/

has some relevant useful information about Tcl/Tk variants.

If I can summarise this, hopefully correctly:

The Aqua Tcl/Tk (Carbon or Cocoa) comes with your Mac. That is
often broken: in Mac OS X 10.5 the menubar didn't work (which is
why I implemented SaVi's -redrawn-menus command line flag to use
buttons for each menu item instead. Implementing hierarchical menus
in SaVi 1.6.0 broke this behaviour. If you need -redrawn-menus,
use SaVi 1.5.1 or earlier until it's fixed. But it's a very specific
fix for a very specific problem in one Aqua Tk variant on an obsolete
operating system, so you likely won't need it.)

Apparently Mac OS X 12 onwards gives a warning that the supplied 
version of Aqua Tk is deprecated, and if you try to use it, Tk windows
spawn without any useful contents, they're just blank. I was
surprised to discover this behaviour on a new M4 Mac, with supplied
Tcl 8.5.9 on Mac OS X Sequoia 15.1.1, which was released in (checks
notes) 2010:

https://wiki.tcl-lang.org/page/Tcl+chronology

The Tcl/Tk supplied with the Mac does not appear well supported,
and seems best avoided. So you have to install a separate
version of Tcl/Tk instead if you want to use SaVi natively on
a Mac, since SaVi is a Tcl/Tk application.

You can either:

- Install ActiveState Tcl/Tk, which used to have a free version
  but now, as far as I can see, costs money after a trial.
  http://www.activestate.com

- compile and build Tcl/Tk from source, as you would on any other
  platform. This assumes that you're happy with the unix shell
  command line. Download the Tcl and Tk packages from http://tcl.tk 

-- now you have options. The easiest is to cd into each package and:
     cd unix ; ./configure ; make ; sudo make install
   in both the unpacked tcl and tk directories.
   This builds the X11 variant, which loses out on various
   Mac-specific niceties such as Mac menu bar support which the
   supplied Aqua version has. (okay, I'm not sure what the other
   niceties are.) You have to use a menu bar in each window, like
   an absolute savage! Or like a Windows user! But I repeat myself.

   But this builds, it works, and is very much the path of least
   resistance.

-- The harder option is to read the instructions in the macosx
   directory, and try to follow them to build a flavour of Tcl/Tk
   that understands the Mac. The README there is very old (10.3?
   Darwin?) and I'm not sure that I can understand and rely on it.
   But that's primarily because I'm not a Mac developer, so the
   fault is probably mine. Could the unix/ directory autodetect
   it was on a Mac and offer the option of building an Aqua variant?
   Don't know.

My conclusion: Tcl/Tk supplied with the Mac is very much a
second-class citizen with abandonment issues, not to be trusted,
so bring your own reliable replacement, and build and use SaVi with
that.

My impression is that Apple doesn't care much about Tcl/Tk.

What I find particularly annoying is that this happens every
time I buy a Mac. In 2007 I bought a blackbook, it had Mac OS X
10.5 and that Tk menu problem had to be worked around:

https://web.archive.org/web/20170629150444/http://www.advogato.org/person/lloydwood/diary/63.html

Now, with a new M4 machine running Mac OS X 10.15.blah, it's
finding that the installed Aqua Tk still doesn't usefully
just work, over fifteen years later.

That is my understanding of the Tcl/Tk/Apple Macintosh state of affairs.
Corrections and enlightenment are welcome.

thanks

Lloyd Wood
llo...@us...



 On Friday 17 January 2025 at 00:28:56 GMT+11, lloydwood--- via savi-developers <sav...@li...> wrote:


http://savi.sf.net/beta

I've just put support for Tcl/Tk 9.0 and better support for Mac OS X
into SaVi development. This adds to the new simulation scripts, Mars
simulation, and coverage map options that are already in development. 

(The Tk supplied with current Mac OS X is broken, so you do need to
 install your own version of Tcl/Tk before building SaVi.
 Menus are not currently in the menubar, but are at the top of
 windows just like other platforms; that needs a little more thought.
 But at least they're not blue!)

And a belated happy new year to you all.

Lloyd Wood
llo...@us...

[savi] New SaVi development drop - Tcl/Tk 9.0 and Mac OS X compatibility

[<llo...@us...>]

http://savi.sf.net/beta

I've just put support for Tcl/Tk 9.0 and better support for Mac OS X
into SaVi development. This adds to the new simulation scripts, Mars
simulation, and coverage map options that are already in development. 

(The Tk supplied with current Mac OS X is broken, so you do need to
 install your own version of Tcl/Tk before building SaVi.
 Menus are not currently in the menubar, but are at the top of
 windows just like other platforms; that needs a little more thought.
 But at least they're not blue!)

And a belated happy new year to you all.

Lloyd Wood
llo...@us...

[savi] Mars, baby!

[<llo...@us...>]

https://savi.sourceforge.io/beta/

The latest development drop of SaVi allows you to select Mars in the Simulation constants. If you turn off the Earth map and set the blue decay to red, you can almost imagine you're there... drawing appropriate landmaps is a future problem.

The Clarke three-satellite stationary script works at the right height, through drift (which is different to Earth's) is not simulated that accurately.

A nice article on how drifting happens at Mars stationary:
https://www.planetary.org/articles/stationkeeping-in-mars-orbit

This is very much an experimental feature.

Suggestions on other bodies worth adding welcomed.

Lloyd Wood
llo...@us...

[savi] new SaVi beta drop

[<llo...@us...>]

https://savi.sourceforge.io/beta/


The latest beta drop of SaVi includes support for drawing the equator and tropics of Cancer and Capricorn in the Coverage View.


This turned out to be rather more effort than I'd expected for just drawing a couple of lines...


Lloyd Wood
llo...@us...

[savi] SaVi 1.6.0 released

[<llo...@us...>]

SaVi is software for simulation and visualization of satellite orbits, especially
satellite constellations such as O3b, Starlink and Project Kuiper.

The SaVi 1.6.0 release (30 April 2023) integrates with v1.2 of the satellite plot
scripts for better presentation of SaVi coverage maps, adds a reworked hierarchical
menu structure with submenus, adds simulations of Audacy, Karousel, Project Kuiper,
Laser Light/HALO, RADARSAT, Starlink, Telesat, Theia and Viasat LEO constellations,
adds simulations of Singh's geosynchronous 4-satellite constellations as described
in Nature Communications, increases maximum number of coverage colors to 29 for
megaconstellations with a direct popup menu, adds menu options for coverage map size,
adds -small-map and -all-maps options, associates fisheye, sunlight and geostationary
satellite behaviour with simulation constants, and suppports wider downloading of
TLEs using lynx
Enjoy!
Lloyd Woodlhttp://savi.sf.net/

[savi] New SaVi development drop

[<llo...@us...>]

http://savi.sf.net/beta

today's development drop adds a command-line -all-maps flag that returns the unpopular projections that no-one uses to the coverage menu, and finally ties the fisheye's geostationary arc to simulation constants, so if you mess with e.g. the time of the sidereal day, the arc moves. (Tcl maths is quite painful to get right. I can see why everyone moved to Python!)

Little things...

Lloyd Wood
llo...@us...

[savi] SaVi and Geomview showing intersatellite links

[<llo...@us...>]

I stumbled across a video suggesting a variant of SaVi has been modified to simulate intersatellite links -- functionality which doesn't yet exist in mainstream SaVi releases.


https://www.youtube.com/watch?v=M49yyJ0o5YU


This seems related to some work presented to the Internet Engineering Task Force at IETF 112 in November 2021.


https://datatracker.ietf.org/meeting/112/materials/slides-112-rtgwg-3-satelite-semantic-addressing-00


Related internet drafts:

https://datatracker.ietf.org/doc/draft-lhan-problems-requirements-satellite-net/

https://datatracker.ietf.org/doc/draft-lhan-satellite-semantic-addressing/

https://datatracker.ietf.org/doc/draft-lhan-satellite-instructive-routing/



This intersatellite link functionality would be a great addition to SaVi's capabilities, and could be combined with the modifications I've been slowly making (rendering, resizing, submenus...) to improve SaVi for a new release.

http://savi.sf.net/beta/


But, other than the existence of this video as proof of existence, I know nothing about the development of this SaVi variant. Information welcomed...


Lloyd Wood

llo...@us...


SaVi satellite constellation visualization http://savi.sf.net/

[savi] SaVi 1,.5.1a released

[<llo...@ya...>]

Happy New Year

SaVi 1.5.1a is just a quick repackage to add a gcc switch
so that SaVi compiles and links as before on gcc 10 and 11.

If you already have SaVi working, you don't need this.

New features continue to be added slowly in beta at
http://savi.sf.net/beta

Lloyd Wood
http://savi.sf.net/

Re: [savi] [savi-users] New SaVi beta snapshot

[<llo...@ya...>]

Well, 'some other developments' was an understatement, and it's not just
ViaSat's system that has changed. There's a good summary of filings:


https://www.linkedin.com/pulse/new-fcc-processing-round-prompts-license-requests-frhr-von-der-ropp/? 



New FCC processing round prompts license requests for 81,195 NGSO satellites


So a number of scripts are no longer up to date with where filings are at.
Telesat's is likely the most easily tweakable.


***


Given that some FCC deadlines are due shortly, there may
be some other developments with other competitors soon.


Lloyd Wood 
http://savi.sf.net/

[savi] New SaVi beta snapshot

[<llo...@ya...>]

https://savi.sourceforge.io/beta/


A new beta of SaVi is available. This one adds a simulation
of ViaSat's modified MEO filing, where they went from MEO
to LEO and from 20 satellites to 288 satellites - a rosette,
so not the same as the Boeing 288-satellite Teledesic design.


There is some commentary on this on a ViaSat earning call
today:
https://www.fool.com/earnings/call-transcripts/2020/05/26/viasat-inc-vsat-q4-2020-earnings-call-transcript.aspx


Given that some FCC deadlines are due shortly, there may
be some other developments with other competitors soon.



Lloyd Wood 
http://savi.sf.net/

[savi] New SaVi beta snapshot

[<llo...@ya...>]

https://savi.sourceforge.io/beta/


A new beta of SaVi is available. This one tidied up the submenus. 


Surprisingly, I'm not getting as much done on SaVi during lockdown 
as I had hoped. Such are the vagaries of life.



Lloyd Wood 
http://savi.sf.net/

[savi] Latest SaVi beta

[<llo...@ya...>]

https://savi.sourceforge.io/beta/


savi-dev-2020-02-02


The latest SaVi beta includes a first attempt at simulating
the 'efficient' four-satellite Singh constellations published
in this month's Nature Communications. Rather hard to tell
if I have the simulations correct based on the figures that
were published in that paper... they're down the Constellations
menu in the 'Other' section.


I'm thinking about the eventual release of SaVi 1.6.0.
Still to do for that is supporting satellites of multiple
colours, which is needed for telling the difference between
overlapping orbital shells of megaconstellations. And
simulating all the shells of Starlink, which would be
subject to change at any point. And doing some tidyup
to maps and to the satellite plot scripts.


Any other suggestions would be welcome.



Lloyd Wood
http://savi.sf.net/

[savi] tcl-lint-vars utility

[<llo...@ya...>]

I wrote a small utility to help clean up Tcl files by
identifying unneeded global variable declarations; there
were quite a few to remove from SaVi, and I couldn't
find anything else that would ease the process.


As is the nature of code cleanup utilities, it has 'lint' in its name.

https://en.wikipedia.org/wiki/Lint_(software)


README and code are at:


https://sourceforge.net/p/savi/tcl-lint-vars/HEAD/tree/


and there's a tarball under:

http://savi.sf.net/beta


It's probably only useful to me -- but, hey, feel free to
prove me wrong on that. Comments welcome.



Lloyd Wood

http://savi.sf.net/

[savi] new SaVi beta tarball release

[<llo...@ya...>]

http://savi.sf.net/beta

A new SaVi development beta release is out.

This allows resizing the coverage map at whim,without that annoying 'what size do you want?'dialog.


Feedback is welcome.


Lloyd Wood 
http://savi.sf.net/

  

[savi] new SaVI beta tarball release

[<llo...@ya...>]

http://savi.sf.net/beta

A new SaVi development beta release is out.

This includes a simulation of the RADARSAT constellation,and adds a -verbose flag to restore the OOGL comments that areno longer sent to Geomview as unnecessary - useful fordebugging geomview interaction, with 
./savi -fake-geomview > oogl.log

(I'm debugging Geomview because adding colour choices toindividual satellites turns out to require some reworkingof how Geomview defines each satellite...)


Lloyd Wood 
http://savi.sf.net/

[savi] Latest SaVi beta - Audacy, Laser Light, Theia

[<llo...@ya...>]

https://savi.sourceforge.io/beta/hh


this latest 22 Sep SaVi tarball adds simulation scripts
for the Audacy, Laser Light (HALO) and Theia proposals.


Further suggestions for things to simulate,
likely based on descriptions in FCC filings,
appreciated.


Lloyd Wood

http://savi.sf.net/

[savi] New SaVi beta - updates Starlink

[<llo...@ya...>]

https://savi.sourceforge.io/beta/


I updated the Starlink simuulation to attempt to match their latest filing.

I don't know what they're doing, but then, it seems, neither do they.



Feedback is welcome.


Lloyd Wood

http://savi.sf.net/

[savi] New SaVi beta - adds Telesat simulation

[<llo...@ya...>]

https://savi.sourceforge.io/beta/
A new SaVi beta drop is available. Do give it a try.Feedback is welcome.

This adds a Telesat simulation. Their geometry doesnot seem well optimised, but is as close to thepatent application description as I think I can get it.
https://spacenews.com/spacex-says-more-starlink-orbits-will-speed-service-reduce-launch-needs/
SpaceX have recently filed with the FCC to changetheir constellation geometry - again - but I haven'tyet found that paperwork to modify the Starlink
simulation. "Why bother, it will only change again"is my thought...
best

Lloyd Woodhttp://savi.sf.net/

[savi] New SaVi beta drop

[<llo...@ya...>]

https://savi.sourceforge.io/beta/

A new beta drop of SaVi is available as we slowly stumble towards the 1.6.0 release.

This includes full simulation of Project Kuiper - but if you want to see partial deployment to match the FCC filing plans you will need to edit the first and second 'shell' scripts.

There's also an attempt at the Karousel constellation, which is not quite right.


Lloyd Wood
http://savi.sf.net/

[savi] new SaVi beta

[<llo...@ya...>]

http://savi.sf.net/beta
The new SaVi beta adds more control over colour shades,and increased colour shades for megaconstellations, aswell as the hopefully-improved popup menu system groupingconstellation scripts.
Feedback is welcome.
thanks

Lloyd Woodhttp://savi.sf.net/

[savi] new SaVi development beta

[<llo...@ya...>]

https://savi.sourceforge.io/beta/

finally, submenus, so we can group constellation scripts
by categories. Suggestions on improved organisation welcome.

(I'm missing a Telesat simulation, and a couple more Kuiper
and Starlink shell simulations)

Lloyd Wood
http://savi.sf.net/

[savi] New SaVi development drop

[<llo...@ya...>]

http://savi.sourceforge.net/beta/

The latest SaVi development drop (17 July 2019) improves the
number of web TLE options that can be loaded from Celestrak,
as well as including "first-shell" Starlink and Project Kuiper
simulations.

Feedback is, as ever, welcome.

Lloyd Wood
http://savi.sf.net/

[savi] SaVi Amazon Project Kuiper script

[<llo...@ya...>]

 https://sourceforge.net/p/savi/savi-dev/HEAD/tree/data/amazon-project-kuiper-first.tcl

I've added a first attempt at a simulation of the initial planned shell
of Amazon's Project Kuiper system, similar to my previous SpaceX Starlink
addition. This is based on very recent FCC filings; the script
explains all.

Comments welcome.

regards

Lloyd Wood
http://savi.sf.net/

Re: [savi] savi-dev not working under geomview

[<llo...@ya...>]

This development crash turned out to be due to use of
the -ansi debug flag for gcc; I've removed that, and
all is as it should be when building with recent gcc.

Recent release packages are unaffected by this.

I've uploaded a new development tarball (also adding
the Starlink inner shell simulation and output
for the ns sat plot scripts) to
http://savi.sf/net/dev

(my, svn tarballs are way bigger than the cvs I'm
used to - huge databases...)

regards

Lloyd Wood
http://savi.sf.net/



 On Saturday, 15 June 2019, 16:22:36 GMT+10, lloyd.wood  wrote: 


This note is just to say that I've noticed that
when compiling savi, it's not working with geomview,
and promptly crashes when launched under geomview,
on ubuntu and on cygwin.

savi-dev works standalone; the savi package still
works with the geomview package, so if you have
savi and geomview installed as packages,
geomview -run savi
works just fine, while
geomview -run ./savi-dev/savi
does not work as it should.

This looks like a change in the linux environment,
rather than within savi or in geomview.

Reports of how this behaves on other systems welcome;
diagnoses of the problem, doubly so.