Greg’s Translations of Nakajima Miyuki Song Lyrics

This is a test.

• This is a link

  WTF, why does it default to bold?

  This sucks.

  Why does my bulleted list not get a new bullet when I hit enter.

  Sucks as bad as Wordpress!

  
  

  • Fu** this.

OMFG.  Not again.  If you are using SSO with Dropbox on Linux, you need to trick the installer by logging onto Dropbox on the web prior to running dropbox.  Only then will it link the account.

https://www.dropboxforum.com/t5/Dropbox-installs-integrations/Entering-Link-Code-After-Install-Ubuntu-18-04/td-p/321876/page/2

clang-format

https://clang.llvm.org/docs/ClangFormat.html

autocomplete & more in emacs

https://github.com/Sarcasm/irony-mode

Check out all these different links to Stanford’s CS229 course. They seem to be different versions of the same course.

http://cs229.stanford.edu/syllabus.html

https://online.stanford.edu/courses/cs229-machine-learning

https://www.coursera.org/learn/machine-learning

http://openclassroom.stanford.edu/MainFolder/CoursePage.php?course=MachineLearning

https://see.stanford.edu/Course/CS229

MIT stuff

http://introtodeeplearning.com/

https://ocw.mit.edu/courses/electrical-engineering-and-computer-science/6-867-machine-learning-fall-2006/syllabus/

https://ocw.mit.edu/courses/sloan-school-of-management/15-097-prediction-machine-learning-and-statistics-spring-2012/lecture-notes/

MIT course references

https://docs.google.com/spreadsheets/d/1jtdtJHXZPbVSIT2xxF18OkQmwjfZrraAKai9e90mfeU/edit#gid=0

Discourse discussion

https://discourse.slicer.org/t/python-wrapping-of-plastimatch/6722/2

ITK module template

https://github.com/InsightSoftwareConsortium/ITKModuleTemplate

https://blog.kitware.com/python-packages-for-itk-modules/

https://itkpythonpackage.readthedocs.io/en/latest/Build_ITK_Module_Python_packages.html

Relevant section of the ITK software guide

https://itk.org/ITKSoftwareGuide/html/Book1/ITKSoftwareGuide-Book1ch3.html#x39-440003.7

https://itk.org/ITKSoftwareGuide/html/Book1/ITKSoftwareGuide-Book1ch9.html#x55-1520009.5

Consider also

https://github.com/pybind/pybind11

https://scikit-build.readthedocs.io/en/latest/

The hardware on sherbert is a Titan XP, compute capability 6.1

https://developer.nvidia.com/cuda-gpus

Currently we have package: nvidia-driver (390.116-1)

Backports has a package nvidia-driver (418.56-2~bpo9+1).  This could be installed to support CUDA 10.1.

https://packages.debian.org/stretch-backports/nvidia-driver

Debian stable/backport has only package nvidia-cuda-dev (9.1.85-8~bpo9+1)

This is not acceptable, because cuDNN does not support this version

Instead, install NVIDIA CUDA 9.0.176 for compatibility with cuDNN

https://docs.nvidia.com/deploy/cuda-compatibility/

cuDNN 7.6.0

https://docs.nvidia.com/deeplearning/sdk/cudnn-support-matrix/index.html

We therefore want Tensorflow 1.12.x

https://www.tensorflow.org/install/source

Which can be installed from package

pip install tensorflow-gpu==1.12.2

To test, run the following

import tensorflow as tf
print (tf.test.is_built_with_cuda())
print (tf.test.is_gpu_available(False,6.1))

There will be an exception for non-cuda available GPUs, so limit the available device as this:

CUDA_VISIBLE_DEVICES=0 python tftest.py

My good friend CP suggests:

1. OBS Studio for screen recording
2. Video Lan Movie Creator
3. Handbrake for conversions

Sad to say, this is what you need to link both a personal and business dropbox account on linux.

https://askubuntu.com/questions/475419/how-to-link-and-use-two-or-more-dropbox-accounts-simultaneously

https://realflash.wordpress.com/2016/02/11/multiple-dropboxes-stops-working-only-one-at-a-time/

Just a couple of links to remind me how to do this

https://docs.fedoraproject.org/en-US/packaging-guidelines/

https://fedoraproject.org/wiki/How_to_create_a_GNU_Hello_RPM_package

The issues which prevented me from using Open Live Writer have been resolved.  Glory be!

Anyway, building tensorflow using Debian’s CUDA package is a freakin’ nightmare.  Tensorflow uses an exotic build system called Bazel.  Bazel cannot successfully use Debian’s CUDA package because … (dramatic pause) … it creates a shell script that copies the entire contents of /usr/include into its own private directory, file by file, on a single line within the shell script, thereby exceeding the maximum line length allowed by the shell.

This is so mind-blowingly stupid that I can’t bring myself to file a bug report.  Makes me want to use Caffe2.  But I digress.

So, anyway, here’s how tensorflow can be built using Debian-managed NVIDIA driver, although not Debian-managed CUDA.

• Use Debian nvidia driver: as of this writing, version 390.87-2~bpo9+1
• Install NVIDIA CUDA, 9.1.85.3.  Install into /usr/local
• Install NVIDIA libcudnn and libcudnn-dev 7.4.1.5, these get installed into /usr/lib/x86_64-linux-gnu and /usr/include/x86_64-linux-gnu.
• Unzip NVIDIA nccl 2.3.7-1+cuda9.-0 into /usr/local/nccl_2.3.7
• Install bazel 0.19.2 from downloaded dpkg
• Checkout tensorflow into $HOME/tf
• Configure bazel options:
• /usr/bin/python
• ?? /usr/lib/python2.7/dist-packages  <- What does this mean?  Is it where my existing packages from Debian are?
• XLA JIT: Y
• OpenCL SYCL: N
• ROCm: N
• CUDA: y
• CUDA SDK: 9.1
• /usr/local/cuda
• /usr/local/cuda-9.1
• (Empty for cuDNN 7)
• cuDNN 7: /usr
• TensorRT: N
• NCCL: 2.3.7
• /usr/local/nccl_2.3.7
• CUDA compute: 6.1
• clang: N
• (Empty for /usr/bin/gcc) <- Debian default is 6.3.0
• MPI: N
• (Empty for -march=native -Wno-sign-compare)
• WORKSPACE: N
• “Bazel build”
• Wait … build succeeds
• “Build the package”
• Pip Install
• Add cuda and nccl to /etc/ld.so.conf.d
• Test https://www.tensorflow.org/tutorials/
• Mostly works, with following caveats:
• model = tf.keras.models.Sequential yields
• WARNING:tensorflow:From /PHShome/gcs6/.local/lib/python2.7/site-packages/tensorflow/python/ops/init_ops.py:1253: calling __init__ (from tensorflow.python.ops.init_ops) with dtype is deprecated and will be removed in a future version.
  
• model.compile(optimizer='adam' yeilds
• WARNING:tensorflow:From /PHShome/gcs6/.local/lib/python2.7/site-packages/tensorflow/python/ops/resource_variable_ops.py:439: colocate_with (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.
• model.fit yields
• 2018-12-28 14:03:34.416116: W tensorflow/compiler/xla/service/platform_util.cc:253] unable to create StreamExecutor for CUDA:0: failed initializing StreamExecutor for CUDA device ordinal 0: Internal: failed call to cuDevicePrimaryCtxRetain: CUDA_ERROR_INVALID_DEVICE: invalid device ordinal
• Random web search suggests the following method to choose the correct GPU
• import os
• os.environ["CUDA_DEVICE_ORDER"]="PCI_BUS_ID"
• os.environ["CUDA_VISIBLE_DEVICES"]="1"
• import tensorflow as tf
• sess = tf.Session()
• Success!

1. sudo systemctl stop wpa_supplicant
2. sudo systemctl restart systemd-networkd
3. sudo systemctl restart systemd-resolved

1. Firmware was not installed automatically.  This might work correctly with a full re-install but I am too lazy (tm).
2. Running lsusb confirms that usb device detected.  Running networkctl does not show wifi link.
3. Fully open source firmware exists, and that is what I want (firmware-ath9k-htc ADD LINK).  Download pkg to USB and install.  Easy-peasy.  Reboot.  Link found when running networkctl.  However, link not available yet.
4. Try to get iwconfig for more information.  This requires two packages: libiw30 and wireless-tools.  Manual download and install.  Now iwconfig works fine.
5. The above was not made easy due to the fact that I could not load packages from the USB stick that I used to install the base OS.  Multiple attempts were unsuccessful due to errors (unsigned packages, obsolete packages, unexpected syntax rules in sources.list, etc.).  I wish I had time to debug and send feedback, but life is short.
6. When I run NetworkManager, it does nothing.  Why?  Is the package not installed?
7. Apparently netctl is not installed. Sure.
8. To be continued.

1. wpa_supplicant is running, and seems to accept my passphrase.  But save_config does not seem to work, because it was not started using a config file that has update_config=1. 
2. I think systemd is using both NetworkManager and systemd-networkd. Not sure yet how to figure what it thinks it is doing.
3. Adding /etc/wpa_supplicant/wpa_supplicant.conf and restarting systemd puts my wlan link into "OPERATIONAL: no-carrier" / "SETUP: configuring" status. 
4. Killing wpa_supplicant and restarting allows configuration to be saved in /etc/wpa_supplicant/wpa_supplicant.conf, albeit with passphrase in plain text. This puts link in "OPERATIONAL: carrier" / "SETUP: configuring" state.  However, doing something like systetmctl restart wpa_supplicant.service does not work well, because two wpa_supplicant processes are run, and the working one is not being used.
5. After rebooting my computer, I found that the default debian install launches wpa_supplicant.service, and NetworkManager.service.  Let's try to get that way working first.
6.  I was able to create the wpa_supplicant-interface.conf file, and start the wpa_supplicant@interface service.  AFAIK, the authentication was successful.  However, networkctl still shows "OPERATIONAL n/a", "SETUP unmanaged".  This is because this command is not used with NetworkManager.  It is only used for systemd-networkd.
7. I can't seem to figure out how to get NetworkManager to use the device.  It might be because dhcp has not yet been run.  The default debian install contains isc-dhcp-client, which has a client named dhclient.  There appears to be no systemd wrapping for this.  Running the client on the command line does not work; DHCPDISCOVER repeats without succeeding.
8. Let's try systemd-networkd.  I disabled wpa_supplicant and NetworkManager, and enabled wpa_supplicant@interface and systemd-networkd.  Rebooted to be sure.  For some reason, NetworkManager didn't disable.  After stopping it, and randomly stopping and restarting systemd units, I have achieved "OPERATION routable" / "SETUP configured" state!
9. I can ping my router!
10. I can ping the outside world (www.oracle.com == 104.96.175.193)!!
11. The default debian install does not contain nsloookup, but it does have a utility called host.  This utility suggest that I am not resolving hostnames.  Starting systemd-resolved is insufficient for the host program to resolve names. However, using "systemd-resolve --status" does show that my link properly got the right DNS server (good 'ole cdns01.comcast.net).
12. The final ingredient is to perform "ln -s /run/systemd/resolve/stub-resolv.conf /etc/resolv.conf" as recommended.  Now, everything works.
13. However, it didn't stick on the reboot.  I have two wpa_supplicant running.  
14. For reasons I don't completely understand, I have a working system with networking.  A few cleanup ops like NetworkManager is still running, but generally good results

1. Rebooted, and still the same problem.  
2. Comment out mount  of previous SATA drives.  Problem resolved.
3. Random?  We'll see.

1. This time the error is different, "Host www.oracle.com not found: 2(SERVFAIL)"
2. Looking at journalctl, they suggest that I should delete /run/wpa_supplicant/interface
3. Let it be done. Reboot.  All seems OK.

RAJA

https://github.com/LLNL/RAJA

https://software.llnl.gov/RAJA/index.html

BSD License, Library requires C++11, MSVC support unclear

Kokkos

https://github.com/kokkos/kokkos

BSD License, Library requires C++11, no MSVC support

OpenMP, Pthreads, CUDA 7.5 rec

SYCL

https://www.khronos.org/sycl

Builds on OpenCL

OpenACC

https://www.openacc.org/

OpenMP 4.0

http://www.openmp.org/

OCCA

https://github.com/libocca/occa

Charm++

http://charm.cs.illinois.edu/research/charm

http://charmplusplus.org/

PGAS/GASPI

http://www.gaspi.de/

GASNet

http://gasnet.lbl.gov/

GASNetEX

https://sites.google.com/a/lbl.gov/gasnet-ex-collaboration/general

Obsolete?

Legion/Regent

http://legion.stanford.edu/

http://regent-lang.org/

Halide

http://halide-lang.org/

Thrust

https://github.com/thrust/thrust

https://developer.nvidia.com/thrust

HCC

http://gpuopen.com/compute-product/hcc-heterogeneous-compute-compiler/

https://github.com/RadeonOpenCompute/hcc

Are the above the same?

HPX

https://github.com/STEllAR-GROUP/hpx

HSA/HSAIL

http://www.hsafoundation.com

UPC

http://upc.lbl.gov/

C++ AMP

Written by Microsoft

Subsumed by hcc?

Agency

https://github.com/agency-library/agency

Written by NVIDIA research

References and other stuff

https://www.khronos.org/assets/uploads/developers/library/2016-supercomputing/SC16-compareSYCL-Michael-Wong_Nov16.pdf

https://devblogs.nvidia.com/parallelforall/building-cuda-applications-cmake/

https://github.com/LLNL/blt

https://charm.cs.illinois.edu/charmWorkshop/slides/CharmWorkshop2017_key_thakur.pdf

Excellent video series from ROScribbles

https://www.youtube.com/channel/UC-qqFCR2qMiFO22stTcB0CA

https://ourworldindata.org/war-and-peace-after-1945/

Maybe some day.  Check comments for additional ideas.

https://entertainment.slashdot.org/story/16/09/18/0645233/scifi-tv-series-space-patrol-orion-celebrates-its-50th-anniversary#comments

It has been acting up again.  This time, I think (maybe?) I could resolve by adding the kernel parameter “enable_mtrr_cleanup” in GRUB.

https://wiki.debian.org/NvidiaGraphicsDrivers/Troubleshooting

Could be very useful.  Generic phase space can be used instead of linac head model.

https://www-nds.iaea.org/phsp/phsp.htmlx

• EGSnrc / BEAMnrc / DOSXYZnrc
  • NRC
  • http://www.nrc-cnrc.gc.ca/eng/solutions/advisory/egsnrc_index.html
  • Latest version 2105
  • Unclear.  Web page says GNU GPL, but user guide says commercial use requires license
• Geant4 / TOPAS
  • CERN / TOPASMC
  • https://geant4.web.cern.ch/geant4/
  • http://www.topasmc.org/
  • Latest version 2015
  • Geant4 has BSD-like license with patent exclusion
  • Topas is licensed commercially
• FLUKA
  • INFN / CERN
  • http://www.fluka.org
  • Latest version 2015
  • Non-commercial / non-military license; modifications not allowed
• PENELOPE
  • University of Barcelona
  • http://www.oecd-nea.org/tools/abstract/detail/nea-1525
  • Latest version 2014
  • Distribution limited to NEA data bank countries
• MCNP / MCNPx
  • LANL
  • https://mcnp.lanl.gov/
  • https://mcnpx.lanl.gov/
  • Latest version 2014
  • Distribution limited by export control regulations
  • MCNP 6 mergest MCNPX features back into trunk
• DPM
  • Michigan / Catalonia
  • https://inte.upc.edu/downloads
  • Latest version 2001
  • No license file ?
  • Related to PENELOPE ?
• EGS4 / MCDOSE
  • SLAC / KEK / NRC
  • Latest version 1997
  • Superceded by EGSnrc
• PEREGRINE
  • Program availability unknown
• VMC++
  • Program availability unknown
• XVMC
  • Program availability unknown

I had some dead time at work between meetings, and I asked myself are there any productivity tools for C++ refactoring that I should look into?  The answer is maybe, but only one seems promising.

https://github.com/tuhdo/semantic-refactor

It is based on symantic-mode, included with emacs

https://www.gnu.org/software/emacs/manual/html_node/semantic/

Can’t wait to try it out!

Sometimes I get spam from ResearchGate.  Unfortunately RG also looks very nice.  But the spam is unacceptable.

The below blog mentions some alternatives.

http://blogs.exeter.ac.uk/openresearchexeter/2013/11/06/74/

• academia.edu
  • Have not yet investigated.
• mendelay.com
  • Probably good, but I’m more of a zotero guy.
• scholar.google.com
  • It is ok.  Does not do publishing.
• figshare.com
• Interesting.  You can publish citable pre-prints / tech reports here.

Need to remember to present this at the next journal club.

http://vision.ucsd.edu/sites/default/files/gestalt.pdf

I need to create some image galleries for my garden.  Wish me luck!  But how to choose a hosting site?

Especially, I heard there is some issue with sites claiming ownership of uploaded photos.

• Flickr
• I already have an account linked with Yahoo
• 1 TB storage
• I still own copyright
• Pinterest
• Tumblr
• Picasa
• Smugmug
• ImageShack / yfrog
• Photobucket / TinyPic
• Imagur
• min.us
• 500px
• Instagram
• Shutterfly
• Skydrive

The following might be useful

• PMD
  • http://sourceforge.net/projects/pmd/
  • BSD-style license
• ConQAT
  • https://www.cqse.eu/en/products/conqat/overview/
  • Apache license
• CCFinderX
  • http://www.ccfinder.net
  • MIT license
• Duplo
  • http://duplo.sourceforge.net/
  • GPL
• similarity-tester
  • http://dickgrune.com/Programs/similarity_tester/
  • (unknown license, in Debian)
• simhash
  • http://wiki.cs.pdx.edu/forge/simhash.html
  • BSD license
• same
  • http://sourceforge.net/projects/same/
  • GPL

Not open source, but free to use

• SourceMeter
  • https://www.sourcemeter.com/
  • Free to use
• Simian
  • http://www.redhillconsulting.com.au/products/simian/
  • Free for open source projects
• SolidSSD
  • http://www.solidsourceit.com/products/SolidSDD-code-duplication-cloning-analysis.html
  • Free for open source projects
• Atomiq
  • http://www.getatomiq.com/
  • Free for anyone
  • Windows only

These seem to be less useful

• Checkstyle
  • Java only
• FindBugs
  • Java only
• CppDepend
  • Proprietary
• SolidSSD
  • Free for open source

I was running Slicer, and suddenly X froze.  Needed to do a hard boot.  (Even /sbin/shutdown was not effective.)

Following the reboot, X was extremely sluggish.  Could not log in through console.  Xorg.0.log was clean, but dmesg had messages such as:

[   65.004180] NVRM: RmInitAdapter failed! (0x26:0x38:1219)
[   65.004198] NVRM: rm_init_adapter failed for device bearing minor number 1
[   65.004241] NVRM: nvidia_frontend_open: minor 1, module->open() failed, error –5

and

Aug 20 11:21:02 sherbert kernel: [  144.948750] NVRM: GPU at PCI:0000:06:00: GPU-4ae70410-b244-63da-efe7-b5356a89babe
Aug 20 11:21:02 sherbert kernel: [  144.948765] NVRM: Xid (PCI:0000:06:00): 13, Graphics Exception: Shader Program Header 11 Error
Aug 20 11:21:02 sherbert kernel: [  144.948772] NVRM: Xid (PCI:0000:06:00): 13, Graphics Exception: Shader Program Header 18 Error
Aug 20 11:21:02 sherbert kernel: [  144.948778] NVRM: Xid (PCI:0000:06:00): 13, Graphics Exception: ESR 0x405840=0xa2040800
Aug 20 11:21:02 sherbert kernel: [  144.948799] NVRM: Xid (PCI:0000:06:00): 13, Graphics Exception: ChID 0003, Class 00009197, Offset 00001f44, Data 00000000
Aug 20 11:29:04 sherbert kernel: [  626.673341] NVRM: Xid (PCI:0000:06:00): 13, Graphics Exception: Shader Program Header 11 Error
Aug 20 11:29:04 sherbert kernel: [  626.673347] NVRM: Xid (PCI:0000:06:00): 13, Graphics Exception: Shader Program Header 18 Error
Aug 20 11:29:04 sherbert kernel: [  626.673351] NVRM: Xid (PCI:0000:06:00): 13, Graphics Exception: ESR 0x405840=0xa2040800
Aug 20 11:29:04 sherbert kernel: [  626.673365] NVRM: Xid (PCI:0000:06:00): 13, Graphics Exception: ChID 0003, Class 00009197, Offset 00001614, Data 00000000

Decided to upgrade my distro and reboot.  Still no love.

Eventually a forum suggested this setting in xorg.conf.

Section "Device"
    Identifier     "Videocard0"
    Driver         "nvidia"
    Option "BackingStore" "True"
EndSection

It seems to have worked ??

Instructions on how to install debian in a qemu image:

Trial one:

https://gist.github.com/philipz/04a9a165f8ce561f7ddd

Mostly it worked, but I couldn’t copy the initd and linux from qemu to host.  The /boot partition was empty.

Trial two:

This seems to be more sophisticated.  The UEFI allows to boot from the disk image rather than from a separate initd and linux.

https://wiki.ubuntu.com/ARM64/QEMU

See also this:

http://blog.eciton.net/uefi/qemu-aarch64-jessie.html

This didn’t work too well.  It just kind of panics.  Seems to be that Qemu 2.2 is required for UEFI.

Trial three:

Basically a re-try of trial one, but exploring alternate approaches of getting the kernel off the VM. This page explains how to mount a host directory onto the guest

http://www.bennee.com/~alex/blog/2014/05/09/running-linux-in-qemus-aarch64-system-emulation-mode/

The following is a very impressive effort to document.  It explains how to transfer files by network instead of file system, and it explains how to retrieve a working initrd.

https://gmplib.org/~tege/qemu.html

Note however, that the above method did not work for mips, due to network adapter not found.  The below method did seem to allow to continue.

http://www.aurel32.net/info/debian_mips_qemu.php

See also this.

http://www.hellion.org.uk/blog/posts/debian-installer-arm64-dailies/

For reference, the up-to-date daily installer are here:

http://d-i.debian.org/daily-images/arm64/

Tony was nice enough to explain a basic idea how to do git branching for me today. 

1. Create new branch from master
1. git checkout –b regression-tests
2. Hack
3. Commit changes
1. git commit –am “Partial implementation of new regression test”
4. Create a second branch for a different problem
1. git checkout –b parser-bug
5. Hack
6. Commit changes
1. git commit –am “Fix parser bug”
7. Merge into master
1. git checkout master
2. git merge parser-bug
8. Delete branch that is no longer needed
1. git branch –d parser-bug
9. Push to remote if desired
1. git push origin master
10. Make first branch current again
11. Continue to hack

http://www.coopercenter.org/demographics/Racial-Dot-Map

This looks interesting.  However, requires one to use github.

https://travis-ci.org

Might need to invest in a more complex logging solution for MABS, to improve multithreading.

Here is what I learned:

http://www.cs.wustl.edu/~schmidt/ACE.html

http://www.codeproject.com/Articles/288827/g-log-An-efficient-asynchronous-logger-using-Cplus

https://code.google.com/p/google-glog/

For my infinite free time

http://entertainment.slashdot.org/story/15/03/31/230221/were-in-a-golden-age-of-star-trek-webseries-right-now

It is becoming increasingly apparent to me that a dataflow programming model will be useful for plastimatch.  Useful characteristics of such a dataflow system would be

• Well-specified programming language
• Library with embeddable interpreter
• Framework for calling C++ functions from interpreter
• Graphical editor of dataflows

A superficial survey suggests that the pickings are pretty slim indeed.  But here are a few options.

• NoFlo
• Has graphical editor
• Has a language, or uses json
• Interpreter written in coffeescript (javascript), therefore would require embedding a javascript interpreter
• CppFBP
• Contains a crude graphical editor
• Has language
• Interpreter?
• Requires Lua and Boost
• https://github.com/jpaulm/cppfbp
• Chameleon
• Has graphical editor
• Has language
• Not obvious interpreter can be embedded
• Written in C++
• http://www.shinoe.org/cameleon/
• Yadinku
• Has graphical editor
• Has language
• Compiler for C++, but no interpreter
• http://statecharts.org/
• CAL / RVC-CAL
• Language only
• orcc
• Has graphical editor
• Seemingly no interpreter
• Requires eclipse
• opendf
• Has compiler, but seemingly no interprete
• PureData
• Has graphical editor
• Has language
• Has audio-video cruft, not sure this can be separated
• Ptolemy II
• Has C compiler, but seemingly no interpreter
• Google Cloud Dataflow
• Apache Pig
• Lucid
• Language only
• plucid
• Interpreter, written in C
• GPL license
• https://code.google.com/p/plucid/
• SISAL
• Compiler, but no interpreter
• https://sourceforge.net/projects/sisal/
• http://ptolemy.eecs.berkeley.edu/ptolemyII/
• YAWL
• Has graphical editor
• More focused on business processes than code
• Seemingly tasks can map to Java, but not C++
• CPN Tools
• Has graphical editor
• Seemingly no interpreter, and no compile
• vvvv
• Has graphical editor
• Research-only license

References

• http://stackoverflow.com/questions/14613045/graphical-dataflow-composition-framework
• http://stackoverflow.com/questions/461796/dataflow-programming-languages/6774990
• http://stackoverflow.com/questions/71979/dataflow-programming-patterns-and-frameworks/21680911

This one cost me about 2 hours of my life.  It felt like 6 hours. 

Various symptoms such as:

$ id gcs6
id: gcs6: no such user

Or

$ ldapsearch –ZZ
ldap_start_tls: Connect error (-11)
   additional info: (unknown error code)

Turns out that the problem was a bug in TLS, which  is exposed when using self-signed certificates.  Doh!  Thanks to user67717 all other users who answered on this and various other forums.

http://serverfault.com/questions/398684/ubuntu-12-04-ldap-ssl-self-signed-cert-not-accepted

Some more information

http://stackoverflow.com/questions/7894633/disabling-msse

And this is interesting as well

http://lists.ros.org/lurker/message/20100323.100115.90abcbeb.pl.html

Maybe this is what I need?

http://doc.lubyk.org/dub.html

Cf.

http://purplepwny.com/blog/binding_lua_to_c_think_twice_before_eating_that_glue.html

and link from

http://realmensch.org/blog/fun-lua-bindings

Or else just use AngelScript or something else.

http://www.resilientcoders.org/

Some pleasure reading when I get the time.

http://linux.slashdot.org/comments.pl?sid=5620493&cid=47812373

Even better than reality TV.  Read the wordpress post by the ex-bf.

http://games.slashdot.org/story/14/09/04/1451217/combating-recent-ugly-incidents-of-misogyny-in-gamer-culture

And this too.

http://www.cracked.com/blog/7-ways-gamergate-debate-has-made-world-worse/

I wish there was a “knock-dead” answer to this.  Zotero only supports webdav for their sync system.  But the free webdav choices are not so great, even though there are many potential choices

Hm, I found this page which is more complete?

https://www.zotero.org/support/kb/webdav_services

This is also interesting

http://cloudstoragebuzz.com/dropbox/10-free-cloud-storage-services-more-secure-than-dropbox/

http://squte.com/

http://pipedot.org/

http://soylentnews.org/

QTM:  Doesn’t seem to support Blogger

Drivel:  The current version in debian 3.0.3 doesn’t seem to work.  I get an error message: “The server did not return a valid Atom response.”  The newer version 3.0.5 has an installer problem, which is that you cannot install into a directory specified by --prefix.

GScribble: To be evaluated

pyblogger: To be evaluated

** To set up a new box.

1) Install the homeshick software

git clone git://github.com/andsens/homeshick.git $HOME/.homesick/repos/homeshick

2) Download my dotfiles

$HOME/.homesick/repos/homeshick/bin/homeshick clone username/dotfiles

3) Done!

** To update an existing box

1) Run the refresh command

homeshick refresh

** To commit changes

1) Add new files

homeshick track dotfiles new_file

2) Commit changes

homeshick cd dotfiles
git commit -m "Added awesome tmux configuration"
git push origin master

Only the good ones

http://headneckbrainspine.com/

http://www.wesnorman.com/homepage.htm

http://www.anatomyatlases.org/

http://www.mhhe.com/biosci/ap/seeleyap/student/olc2/atlas.html

http://www.med.wayne.edu/diagRadiology/Anatomy_Modules/Page1.html

http://www.med.harvard.edu/AANLIB/home.html

http://www.lumen.luc.edu/lumen/meded/grossanatomy/x_sec/mainx_sec.htm

https://highered.mcgraw-hill.com/sites/0072495855/student_view0/cross_sectional_mini-atlas.html

http://www.dartmouth.edu/~rswenson/Atlas/

Not cross-sectional, but very valuable

https://commons.wikimedia.org/wiki/Special:Contributions/Anatomist90

Runner-up

http://www9.biostr.washington.edu/da.html

http://www.chestx-ray.com/Anatomy/Anatomy.html

http://library.med.utah.edu/WebPath/HISTHTML/ANATOMY/ANATOMY.html

Tracfone

StraightTalk

http://itcr.nci.nih.gov

The big ones are:

http://zillow.com/

http://trulia.com/

http://redfin.com

There are some interesting thoughts from real estate agents on these sites. 

http://www.inman.com/2014/03/27/what-to-do-when-seller-complains-but-zillow-says-my-house-is-worth-more/

N.b. the above article mentions three other AVM that you can use to compare with zestimate.

http://www.businessweek.com/articles/2013-03-07/why-redfin-zillow-and-trulia-havent-killed-off-real-estate-brokers

What a bunch of maroons.  On XFCE, the default terminal width for Debian X Terminal Emulator is 79 characters.  If you are using xfce4-terminal, you can do this:

http://www.linuxquestions.org/questions/slackware-14/adding-geometry-parameter-to-launcher-not-working-xfce-4-6-2-on-slackware-13-37-a-886411/

Often I wonder about what is the best way to code my C++.  For example, read the first paragraph of the following (excellent) article:

http://cpp-next.com/archive/2009/08/want-speed-pass-by-value/

In that article, Dave Abrahams suggests that C++0x compliance is sufficient for proper return value optimization.  Note also in the comments about the requirement to create a variable in the function definition to guarantee RVO. However the fact remains that many projects are still stuck on VS 2008.  I have not seen any detailed comparison of which optimizations are supported by each compiler version.

The generalities are clear, that you need a proper <omp.h>, vcomp.lib and vcompd.lib from the SDK, and that the SDK version must match your compiler version. 

VS 2005: http://blog.codekills.net/2007/09/20/openmp-and-visual-c++-the-free-way-%28sorta%29/

VS 2008: http://honya8bit.seesaa.net/article/163363596.html

VS 2010: http://www.ssidelnikov.ru/en/2011/05/openmp-visual-studio-2010-express/

VS 2012: OpenMP now supported by express version

• There are 3 ethernet jacks.  Only one seems to be effective.  Need to follow up with MW on that
• Add new local user & change root password
• Add local user to sudo-ers
  adduser foo sudo
• Install xfce
  sudo apt-get install xfce4
  https://wiki.debian.org/Xfce
• Partial uninstall of gnome.  Uninstall of everything seems rather harsh.  Revisit this sometime.
  sudo apt-get purge task-gnome-desktop gnome network-manager-gnome gnome-core gnome-keyring
• Fix colors on default terminal
• Set hostname
  https://wiki.debian.org/HowTo/ChangeHostname
• Enable ssh.  For some reason it didn't work out of the box.
  sudo dpkg-reconfigure openssh-server
  https://bugs.launchpad.net/ubuntu/+source/openssh/+bug/1005440
• Enable LDAP authentication

  ln -s /home /PHShome
  ln -s /bin/bash /bin/PHSshell

• Don't forget to add /bin/PHSshell into /etc/shells as well, this is required for LDAP user to have sudo access.
  

  sudo apt-get install nslcd libnss-ldapd libpam-ldapd
  

  The configuration tool will ask for:

  - ldap URI (ldap://ldap.partners.org)
  - search base (cn=users,dc=partners,dc=org)
  - services to configure (passwd,shadow)

• Add extra lines to nslcd.conf (and tls certificate) as described here:
  
  http://research.partners.org/wiki/index.php/Active_Directory_on_Unix
  
  Add the following line to /etc/pam.d/common-session
  

  session     required      pam_mkhomedir.so skel=/etc/skel umask=0022

  as described here:
  https://wiki.debian.org/LDAP/PAM
• Install synergy
  
  sudo apt-get install quicksynergy

Well, it seems that, in addition to extra hoops for installing onto RAID, the nouveau kernel module crashes when trying to boot.  Since (unlike Ubuntu or Mint), the debian installer works with RAID, I downgraded to debian stable ("squeeze").

1. Install debian without the desktop.  Reboot.
2. Read up on nvidia drivers.  Decide to try the "debian way".  (http://wiki.debian.org/NvidiaGraphicsDrivers)
3. Nvidia proprietary drivers are updated in backports repository.  Add it to /etc/apt/sources.list
   deb http://backports.debian.org/debian-backports squeeze-backports main
4. Install desktop after the fact (http://wiki.debian.org/Gnome)
   apt-get install aptitude tasksel
   tasksel install gnome-desktop --new-install
5. Remove nouveau (http://forums.debian.net/viewtopic.php?f=6&t=69888 | http://forums.debian.net/viewtopic.php?f=30&t=54330)
   echo "blacklist nouveau" >> /etc/modprobe.d/blacklist.conf
   apt-get purge --remove xserver-xorg-video-nouveau
6. Use aptitude, and remove every package with the word "nouveau" in it.  However, nouveau.ko is in the linux kernel package, so I can't actually remove it.
7. Continue, re-install nvidia drivers
   apt-get install --reinstall nvidia-glx
8. This is a key step (see http://linux.koolsolutions.com/2009/10/11/howto-blacklisting-kernel-module-from-auto-loading-in-debian/)
   depmod -a
   update-initramfs -u -k `uname -r` -v
9. Reboot.  Success!

1. Backup home directory.  Backup etc.  Unplug scratch drives
2. Boot live cd
3. Install mdadm and lvm2
4. Re-assemble the raid partitions
   mdadm --assemble --scan
5. Run linux mint installer.  The /dev/mdXX partitions are recognised and usable for install.  Re-format existing partitions (except /home). Choose /dev/md1 (/boot partition) as the partition for the boot loader.
6. I got an error in the install.  Maybe benign?  It says:
   An error occurred while migrating data.  More details may be found in /var/log/syslog.  The installation will continue, but some or all of the documents and settings you requested may not have been transferred to the installed system.
   The error is that it couldn't mount /dev/sdb5 (which belongs to /dev/md5), because it doesn't know the file system "linux raid member".  Click OK to continue.
7. Choose "continue testing" at the end of the install.
8. Now we have to set up the newly installed system to recognise the raid.  This is exactly as described in part 5 of this document:
   http://www.howtoforge.com/install-ubuntu-with-software-raid-10
   With the modificaiton of additional required partitions (/usr, /var, etc.), and installing lvm2 as well as mdadm.
9. Reboot.
10. Well, the reboot didn't work, because I neglected to update grub.  Here is the error that I got:
    GRUB Loading stage1.5
    GRUB loading, please wait...
    Error 15
11. Boot off of the live CD again, re-install mdadm & lvm2, re-scan raid, re-mount volumes, enter chroot.
    apt-get install mdadm lvm2
    mdadm --assemble --scan
    mkdir /myraid
    mount /dev/md5 /myraid
    mount /dev/md1 /myraid/boot
    mount /dev/md6 /myraid/usr
    mount /dev/md7 /myraid/var
    mount /dev/md8 /myraid/tmp
    mount --bind /dev /myraid/dev
    mount -t devpts devpts /myraid/dev/pts
    mount -t proc proc /myraid/proc
    mount -t sysfs sysfs /myraid/sys
    chroot /myraid
12. Follow instructions on this page:
    http://iiordanov.blogspot.com/2011/07/how-to-install-linux-ubuntu-debian-etc.html
    Start at step (10). It goes smoothly.  However, I notice that my swap is not RAID.  I wonder if this is on purpose?  So I left as two swap partitions in the fstab, to be resolved later.
13. Do step (11), note that I did not have to modify /etc/mdadm/mdadm.conf, since it already had the correct entries.
14. Do step (12).  This failed.  Note 1: the file /boot/grub/grub.cfg did not exist.  Note 2: grub-install was done to the raw disk, not the raid disk, such as:
    grub-install --recheck --modules="raid" /dev/sda
15. Reboot.
16. Now I got a grub menu, which listed a few options including "/dev/md1", "/dev/sda1", and "/dev/sdb1".  Choosting the first option causes a black screen of death.  Choosing "/dev/sda1" yields similar results.  It turns out that something is happening, but nothing is displayed.  Hitting e.g. arrow keys or ESC causes the messages to display.  The boot seems to hang on the following message:
    ** Starting CUPS printer/spooler server **    [ OK ]
17. Boot off of the live CD again, re-install mdadm & lvm2, re-scan raid, re-mount volumes, enter chroot.
18. Nothing interesting to see.  Here is boot.log:
    fsck from util-linux 2.19.1^M
    fsck from util-linux 2.19.1^M
    fsck from util-linux 2.19.1^M
    fsck from util-linux 2.19.1^M
    fsck from util-linux 2.19.1^M
    fsck from util-linux 2.19.1^M
    /dev/md5: clean, 9326/305824 files, 123307/1222912 blocks^M
    /dev/md8: clean, 12/305824 files, 54365/1222912 blocks^M
    /dev/md7: clean, 8815/305824 files, 130828/1222912 blocks^M
    /dev/md6: clean, 163285/1221600 files, 772988/4885744 blocks^M
    Reiserfs super block in block 16 on 0x90a of format 3.6 with standard journal^M
    Blocks (total/free): 110268128/33138842 by 4096 bytes^M
    Filesystem is clean^M
    /dev/md1: clean, 246/124928 files, 56798/497856 blocks^M
     * Starting mDNS/DNS-SD daemon^[[154G[ OK ]^M
     * Starting network connection manager^[[154G[ OK ]^M
     * Starting CUPS printing spooler/server^[[154G[ OK ]^M
     * Starting emergency keypress handling^[[154G[ OK ]^M
     * Stopping emergency keypress handling^[[154G[ OK ]^M
     * Starting System V runlevel compatibility^[[154G[ OK ]^M
     * Starting save system clock to hardware clock^[[154G[ OK ]^M
     * Starting save sound card(s') mixer state(s)^[[154G[ OK ]^M
     * Starting eCryptfs^[[154G[ OK ]^M
    Here is the end of kern.log:
    Apr  2 20:28:54 gelato kernel: [   18.535691] e1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: Rx/Tx
    Apr  2 20:28:54 gelato kernel: [   18.536853] ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
    Apr  2 20:29:04 gelato kernel: [   28.840999] eth0: no IPv6 routers present
    Apr  2 20:31:14 gelato kernel: Kernel logging (proc) stopped.
    Here is the end of syslog:
    Apr  2 20:29:04 gelato avahi: contact your administrator and convince him to use a different DNS domain,
    Apr  2 20:29:04 gelato avahi: since .local should be used exclusively for Zeroconf technology.
    Apr  2 20:29:04 gelato avahi: For more information, see http://avahi.org/wiki/AvahiAndUnicastDotLocal
    Apr  2 20:29:04 gelato kernel: [   28.840999] eth0: no IPv6 routers present
    Apr  2 20:29:12 gelato ntpdate[1303]: adjust time server 91.189.94.4 offset 0.213605 sec
    Apr  2 20:31:14 gelato kernel: Kernel logging (proc) stopped.
    Apr  2 20:31:14 gelato rsyslogd: [origin software="rsyslogd" swVersion="5.8.1" x-pid="1172" x-info="http://www.rsyslog.com"] exiting on signal 15.

When you try to run slicer remotely on the SPL computer, you get something like this:

$ ./Slicer-build/Slicer
Number of registered modules: 94
ERROR: In /home/gcs6/build/slicer-rt/SlicerRt-build/VTK/Rendering/vtkXOpenGLRenderWindow.cxx, line 404
vtkXOpenGLRenderWindow (0x333a200): Could not find a decent visual

ERROR: In /home/gcs6/build/slicer-rt/SlicerRt-build/VTK/Rendering/vtkXOpenGLRenderWindow.cxx, line 404
vtkXOpenGLRenderWindow (0x333a200): Could not find a decent visual

X Error: BadValue (integer parameter out of range for operation) 2
  Extension:    135 (Uknown extension)
  Minor opcode: 3 (Unknown request)
  Resource id:  0x21
ERROR: In /home/gcs6/build/slicer-rt/SlicerRt-build/VTK/Rendering/vtkXOpenGLRenderWindow.cxx, line 629
vtkXOpenGLRenderWindow (0x333a200): Cannot create GLX context.  Aborting.

error: [/workspace/gcs6/build/slicer-rt/SlicerRt-build/Slicer-build/bin/./SlicerQT-real] exit abnormally - Report the problem.

It seems you can workaround by doing this:

$ setenv LIBGL_ALWAYS_INDIRECT y