Hi everyone,
this is not strictly GMT, but I am not sure where to turn…
Does anyone know who manages the TTT software now that Paul is no longer with us? I bought. a license way back when… and the computer it ran on is on it’s last legs, and I am having trouble getting it to run properly on my newer computers…
Thought one of the GMT Team might know
Thanks!
-jose
Hi,
I have wondered about that too but I never saw the TTT source code. Did you get when you get that license?
I have no idea what TTT software is, but one general solution for an old software that cannot be setup on a new computer is to run it in a virtual machine.
E.g. an existing old working system’s hard drive can be copied/cloned to a virtual machine disk file and then configured as a virtual machine’s hard drive. This way ensures there’s always a working copy if an old PC is falling apart.
Or an old OS can simply be installed in a virtual machine and then the software installed the usual way.
Computes the so called Tsunami Travel Times. It’s a C code written by Paul but the source code was not public.
Yes sure, I googled it. Did not read beyond the title page so no working idea what’s actually needed to install and run it.
I also have a TTT code created by me in Mirone (a C-MEX file) and decided to put Claude to work. After a few iterations we came out with a new module in GMT.jl (in #master for now).
using GMT
G = gmtread("@earth_relief_01m", R=:PT); # Grid size (729, 1508)
@time ttt = wave_travel_time(G, [-11.0, 35.9]);
0.770920 seconds (25 allocations: 12.748 MiB) # After warm up
viz(ttt, contour=true, colorbar=true, coast=true, proj=:guess) # Looks like contour defaults are too thick
Hi Joaquim,
Thanks for looking into it. Can we discuss via private email…? I’m interested in the tool you made. But I have no idea how to use julia.. As you know, I am generally barely hanging on when it comes to this stuff. i have some well established workflows and I don’t deal well when there are perturbations in my flow! ha ha… Maybe you can help me understand why the TTT isn’t working on the other Windows machines, it seems like a simple path issue, one of the '' in the path name is in wrong direction when it’s looking for the bathymetry data upon which to do the TTT calc… It doesn’t do that on the original windows laptop wher it is all working (the one that is close to the end of its service life) I’m sure it’s something stupid… I just don’t know how this stuff works… the ‘ttt_64’ executable itself runs, it just fails on the step when it is looking for the bathy grid to run on, and I don’t know where to change that. I’ll do a deeper dive on the man page and see if I can figure it out… Thanks!
Hi Joaquim,
well i messed around on the windows machine and changed the path in the environment variable that TTT uses to find the bathymetry and guess what, it worked! So I am back up and running. so um, yeah, disregard!
But that said, I will ask around in the tsunami world to see if there were any handoff plans for TTT… I know the tsunami warning centres use it and they bought a long term license or something so that it could be distributed to foreign govenments for tsunami preparedness etc. I have no idea who holds the source code, but it woudl be good to find out.
Sure. And note that I never saw TTT so don’t know how it works and is installed.
Hi Jose et al.,
This is a timely post because there are some potential issues with Paul’s software that I’m currently trying to pinpoint/diagnose. The problem relates to discrepancies between TTTv3 and TTTv4 (the most recent release) output…yikes. I was planning on posting something soon. Anyway, we (NOAA/NWS/Pacific Tsunami Warning Center) routinely use the Geoware TTT software to generate travel times for messages, maps, etc. The source code can be downloaded from https://ftp.soest.hawaii.edu/ptwc/TTT/. It compiles w/o difficulty on Linux systems (and Macs too I think but it’s been a while since I tried). Hopefully this helps you and Joaquim can take a look at the source code.
Jonathan
Hi Jonathan,
I guess that it would be a bit difficult to inspect the code since the README says
Executables may not be shared with other entities. The TTT
source code (obfuscated for protection)
I am surprised to the extents that Paul gave to hide de source code. I would still be curious to look at the obfuscated code … but that bash script to extract the source from somewhere is in itself a good obfuscation to me. Would you be able to get it (me Windows)?
Another thing that you guys that have a TTT runing and know how to use it is to compare with the results of the wave_travel_time program in GMT.jl that I posted above. I know that it has issues “around corners” and shadowed zones but for the rest, the tests I did at the time I wrote it showed an agreement of the order of minutes with a non-linear shallow water code.
Hi Joaquim,
That’s interesting. I tarred up the tttv3 folder from one of our systems. Just remove the .zip extension. It looks like the source code for the different tools is accessible in the top level. The usage is slightly different than tttv4 but I suspect the computational guts are more or less the same. Please note that I excluded the share folder with the large bathymetry grid files. "
I will try to compare/contrast with the wave_travel_time program you previously posted.
Thank you!
Jonathan
ttt3.tgz.zip (147.1 KB)
Thanks,
Meanwhile I was able to find the 4.0 version of the code in a dir of your linked site. But, like the README says, code is obfuscated.
What is the difference in behavior that you find between v3.o and 4.0?
Hi Joaquim,
Thank you for the response. What do you mean by obfuscated? Were you able to look at the guts of the TTTv3 code? Please see the attached text file that is my initial attempt at documenting the problem. Also attached is the tide station location file (tide.tmp) that I’m using to extract the travel times. Remove the .txt suffix to correspond with the commands. I haven’t had a chance to compare against your new code yet but will do so asap. Let me know what you think if/when you have some time. I may also post something to the tsunami bulletin board in case others have seen the same behavior…not sure how many folks are using the software.
Cheers,
Jonathan
tide.tmp.txt (7.5 KB)
ttt_testing_jw.txt (2.7 KB)
Hi, j(ohn?)
Well, have a look at the code and you’ll understand better than explanations (that you can find here Obfuscation (software) - Wikipedia).
So, no, I didn’t loot at the guts, impossible … but Claude was able to do an, apparently, good job in translating it to Julia. So, I may have a code that I can integrate in GMT.jl (big add on is that it won’t be restricted to GMT native binary grids anymore). However, that discrepancy between versions (my conversion was from v4) add an extra layer of work/complexity. Historical/real cases should help disambiguating which TTT is more realistic.