Published at LXer: Following AMD and FreeBSD Foundation collaborations and the Sovereign Tech Fund making a big investment into FreeBSD, the FreeBSD Foundation and Quantum Leap Research have...

Published at LXer: The Arch Linux-based and KDE Plasma-focused CachyOS distribution has a new ISO release for September 2024 adding various performance improvements and optimizations across the...

Published at LXer: We've already looked at the Radxa X4 kit featuring an Intel N100 SBC with a design similar to the Raspberry Pi 5 and accessories including a Radxa Power PD 30W power adapter, an...

Published at LXer: The Linux 6.12 merge window is wrapping up today with the release of Linux 6.12-rc1 in the coming hours. This is going to be a heck of an exciting kernel. Read More......

Published at LXer: Linus Torvalds announced today the general availability for public testing of the first Release Candidate (RC) development milestone of the upcoming Linux 6.12 kernel series. ...

Published at LXer: The HiFiBerry DAC8x and HiFiBerry Amp4 Pro are two significant upgrades for enhancing audio capabilities on the Raspberry Pi 5. These devices expand the audio options for users...

Published at LXer: In this article, you'll learn how to install the Deluge BitTorrent client on Ubuntu and other Linux distros, how to use it, and then how to remove it. Read More......

Published at LXer: VirtualBox 7.1.2 is the first such point release since the VirtualBox 7.1 series debuted earlier this month. Naturally, it builds on that major release with a flurry of big...

Published at LXer: The x86 platform driver changes that were merged last week for the Linux 6.12 kernel continue to be quite lively with changes for enhancing Linux laptop support along with other...

Published at LXer: Have an NVIDIA GPU? Use GE-Proton? You should probably make sure you're on GE-Proton 9-15 which was just released. This is a "hotfix" build to clear up some problems from the...

Published at LXer: As expected the Linux 6.12-rc1 kernel is out today in marking the end of the very exciting two-week Linux 6.12 merge window that saw many high profile features land. Read...

Published at LXer: Void Linux finally replaces Python 2 with Python 3 and upgrades its buildbot to enhance package delivery. Read More......

Published at LXer: The hardware monitoring "HWMON" subsystem updates for Linux 6.12 added some new drivers as well as adding new device support to some of the existing drivers. Read More......

Published at LXer: After a five-year hiatus, the RTorrent command-line BitTorrent client is back with v0.10, bringing performance upgrades and bug fixes. Read More......

Published at LXer: Linus Torvalds announced the release of Linux Kernel 6.12 RC1. Kernel 6.12 RC1 brings important new features like PREEMPT_RT and sched_ext. Read More......

Published at LXer: Environment variables are system-defined variables that store values used by the operating system and applications. For example, on a Windows machine, information like the CPU...

Published at LXer: In the latest monthly newsletter published today, Linux Mint project leader Clement Lefebvre shares a sneak peek at the new default Cinnamon theme coming to Linux Mint 22.1 later...

Published at LXer: Released remotely from Vienna, Linux kernel 6.11 is here, with improved monochrome TV support. Yes, in 2024. Emperor penguin Linus Torvalds was attending the Open Source Summit...

Published at LXer: While I have been very eager to test out the Core Ultra 200V Lunar Lake series on Linux in part due to the new Xe2 integrated graphics, after several days of pushing a new Lunar...

Published at LXer: The popular open-source email client, Mozilla Thunderbird, has launched a beta version of its Android app with a range of new features and improvements. Read More......

Published at LXer: Rspamd 3.10 spam filtering system brings enhanced MIME UTF8 support and negative score limits for improved email scanning. Read More......

Hello, I'm not sure what happened but I've always been able to tunnel to my xrdp GUI using a ssh putty session. Now when I try to connect after connecting via SSH it just shows a blank screen with an...

Published at LXer: Audacious 4.4.1 builds on the changes introduced in Audacious 4.4 (a release that brought GTK3 and Qt6 UI choices, the return of a dedicated lyrics plugin, and better...

Il y a des jours comme ça ou la chance vous sourit… Quart de finale (foot) Italie – Belgique, bien entendu je n’avais pas acheté de billets. Mais avec un collègue on s’est dit qu’on allait tenter notre bonne chance. Nous voilà donc là, arpentant les alentours...

Ein Coming-Out ist für die meisten homosexuellen Jugendlichen der letzte Ausweg, um ein Leben voller Lügen aufzugeben. Für viele fangen die größten Probleme jedoch erst danach an. Jakob hat sich geoutet, weil er sich selbst nicht mehr leugnen wollte. Sechs Jahre lang versteckte er seine Gefühle, aus Angst, ausgegrenzt zu werden. Seit sein Freund über ein Outing nachdenkt, begegnet er seiner eigenen Geschichte noch einmal.

Und wieder hat Deutschland sein neues "Littleton". Wie an dem aus Michael Moores Film "Bowling for Columbine" bekannten Ort, an dem zwölf Schüler ihr Leben verloren, stürmte Sebastian B. am 20. November 2006 seine ehemalige Schule in Emsdetten. Bewaffnet mit vier Gewehren und drei Rohrbomben schoss er um sich. Rund 30 Menschen wurden verletzt, drei davon schwer. Der 18-jährige Amokläufer tötete sich noch im Schulgebäude selbst. Ein Kommentar über die Hintergründe.

Jessy* strahlt über das ganze Gesicht, als sie ihren einjährigen Sohn in den Armen hält. Erst vor kurzem ist sie zu ihrem Freund nach London gezogen, damit die kleine Familie endlich zusammen ist. Deutschland weint die 23-Jährige nicht nach, denn mit dem Umzug in ein neues Land lässt sie eine schreckliche Vergangenheit hinter sich. Jessy wurde zehn Jahre lang von ihrem Stiefvater Ralf* missbraucht.

Die EU-Grenzschutzagentur Frontex hat gemeldet, dass sie die deutsche Bundesregierung bereits im Frühjahr 2015 über die hohen Flüchtlingszahlen informiert habe. Der Leiter von Frontex, Fabrice Leggeri bestätigte, dass er die Bundesregierung im März über eine Zahl von ca. 500.000 bis 1 Millionen Flüchtlinge informiert habe. Bereits zu diesem Zeitpunkt war ...

Etwa 90 Minuten vor dem geplanten Anstoß im Fußball-Freundschaftsländerspiel zwischen Deutschland und den Niederlanden hat Bundesinnenminister Thomas de Maizière in Absprache mit seinem niedersächsischen Amtskollegen Boris Pistorius die Absage des Spiels verfügt. Durch Lautsprecherdurchsagen wurden die Besucher des Spiels nach Hause geschickt. In einer Pressekonferenz erläuterte de Maizière die Gründe für ...

Russische Kriegsschiffe haben am 20. November Cruise Missiles auf Ziele in Syrien abgefeuert und dabei sieben islamistische Stellungen in den nördlich Provincen Raqqa, Idlib and Aleppo getroffen. Col Patrick Ryder kritisierte, bei den russischen Angriffen würden auch gemäßigte syrische Oppositionsgruppe in Mitleidenschaft gezogen. Die Nachrichtensender BBC und Channel 4 veröffentlichten ...

Am 25. November 2015 wurde der verbreitete Free Pascal - Compiler (FPC) in der aktualisierten Version 3.0 (Pestering Peacock) veröffentlicht. Die Vielzahl der Neuerungen hat die Entwickler zum ersten großen Versionssprung seit 10 Jahren bewogen. Version 2.0 war 2005 eingeführt worden, die letzte Hauptversion war 2.6 aus dem Jahre 2012. Version ...

I have just released GNU Gatekeeper version 5.3.

You can download it from https://www.gnugk.org/h323download.html

This release has a number of new features as well as some important bug
fixes.

Whats new ?

• LRQ loop detection to optimize calls flows between multiple neighbor gatekeepers This new feature has the potential to significantly reduce the load on all gatekeepers and prevent "LRQ storms".
• new routing policy to set call destinations by querying HTTP or REST servers, see [Routing::Http]
• much improved support for SNMP
• important bug fix for TLS encryption of signaling channels
• important bug fixes for H.460.18 NAT traversal (for H.245 tunneling and for multi-homed servers)
• performance optimization: this version can handle 5-10% more proxied  calls on the same hardware
• performance optimization: re-authenticate lightweight, additive registrations only when new aliases differ. This significantly reduces the load on password databases.

Enjoy!


Full change log:

- BUGFIX(ProxyChannel.cxx) don't send H.245 address to tunneling
  H.460.18 endpoint, breaks call when H.245 multiplexing
- performance optimization: 5% faster UDP handling
- changed default: [SNMP] Implementation=PTlib
- remove unfinished Windows-SNMP implementation, use PTLib-SNMP on Windows
- support SET and GET-NEXT in PTLib-SNMP
- support SNMP sysUpTime when running as standalone agent
- BUGFIX(configure.in) LARGE_FDSET defaults to off
- new SNMP OID 1.3.6.1.4.1.27938.11.1.9 to query total bandwidth allocated to ongoing calls
- BUGFIX(ProxyChannel.cxx) fix hangup when making many TLS calls quickly one after another
- BUGFIX(RasSrv.cxx) don't require H.460.22 parameters in ARQs
- BUGFIX(ProxyChannel.cxx) fix TLS without LARGE_FDSET
- BUGFIX(ProxyChannel.cxx) don't send H.460.22 priority field in SCI
- BUGFIX(gkauth.cxx) free memory from cached and expired passwords
- re-authenticate lightweight, additive registrations only when new aliases differ
- remove switch [Proxy]DisableRTPQueueing, always disabled now
- new routing policy: http with config section [Routing::Http]
- BUGFIX(ProxyChannel.cxx) fix H.460.18 on multi-homed servers (SCI comes from the correct IP now)
- new switch to disable SNMP traps [SNMP] EnableTraps=0
- BUGFIX(ProxyChannel.cxx) don't throw SNMP trap on H.245 connection errors
  (causes crash under load with Net-SNMP)
- BUGFIX(snmp.cxx) shutdown GnuGk when SNMP agent can't be started
- BUGFIX(snmp.cxx) protect NetSNMP library calls with mutex
- changed default: ForwardResponse now defaults to 1 in [RasSrv::LRQFeatures] and [Neighbor::...]
- new feature: loop detection for LRQs [RasSrv::LRQFeatures] LoopDetection=1
- BUGFIX(Neighbor.cxx) some settings in [RasSrv::LRQFeatures] were ignored if not set in [Neighbor::...]

I am happy to announce the release of GNU Gatekeeper 5.4.

You can download it from https://www.gnugk.org/h323download.html

New features:

• new accounting module to send accounting data to an MQTT server
• support for redis as database (eg. as backend for password storage)

Bug fixes:

• important fix for H.245 tunneling translation with H.460.18 endpoints
• fix for snmpwalk in PTLib-SNMP implementation
• fix sending alternate gatekeeper list to endpoints with assigned gatekeeper
• improved DRQ from child gatekeepers
• fix TLS with neighbor gatekeeper

Please also note that a bug has been found in PTLib that can cause a crash in any GnuGk version if you use the status port (manually of from an application). Please upgrade to PTLib 2.10.9.3!

I am happy to announce the release of GNU Gatekeeper 5.5.

This release has new features and bug fixes when you run clustered gatekeepers. It also improves the port detection feature and we have a complete and up to date Chinese documentation.

You can download it from https://www.gnugk.org/h323download.html

New features:

• new feature GnuGkAssignedGatekeeper to push endpoints back to their intended home gatekeepers in the cluster, even if the endpoints don't support assigned gatekeepers
• support new PBKDF2 password hashes for ssh logins to the status port
• new switches to fine tune port detection for H.239 channels (IgnoreSignaledPublicH239IPsFrom=x and IgnoreSignaledAllH239IPs=1)
• new Chinese manual

Bug fixes:

• select correct source IP for neighbor pings
• set altGKisPermanent=true when redirecting endpoints
• fix RRJ to include alternates when RedirectGK=Endpoints limit is reached
• fix reading of AllowSignaledIPs= switch
• don't complain about [Neighbor::xxx] SendAliases switch when using--strict

Enjoy!

Today GNU Gatekeeper version 5.6 has  been released.

Download: https://www.gnugk.org/h323download.html

It contains an important bug fix to H.460.19 multiplexing and H.460.26 (media over TCP) when using GnuGk's internal call forwarding (ForwadOnFacility).

I have also added an interop tweak to be able to call video services that don't understand H.323 URL aliases (eg. videobutler.nl). You can enable it with

Most H.323 vendors did not implement encrypting the signaling connection with TLS. They only encrypt the media (RTP). But you can use the two GNU Gatekeepers to encrypt you call signaling even when your endpoints don't support this natively.

Suppose you have 2 locations and want to connect them securely over the public internet.

GnuGk can encrypt call signalling between those locations using TLS and encrypt the media (RTP) using H.235.6 (AES encryption). 

 Configuration for GNU Gatekeeper 1 (prefix 01)

 

[Gatekeeper::Main]

[RoutedMode]
GKRouted=1
H245Routed=1
CallSignalPort=1720
AcceptUnregisteredCalls=1
; make sure H.245 gets tunneled for TLS
H245TunnelingTranslation=1
; add AES media encryption if the endpoint doesn't encrypt itself
EnableH235HalfCallMedia=1
; only allow encrypted calls
RequireH235HalfCallMedia=1
; change the media key after 2^31 operations
EnableH235HalfCallMediaKeyUpdates=1

[Proxy]
Enable=1

[ModeSelection]
0.0.0.0/0=PROXY
; only use routed mode for local calls
192.168.0.0/18=H245ROUTED

[TLS]
EnableTLS=1
PrivateKey=/path/to/server.pem
Certificates=/path/to/server.pem
CAFile=/path/to/rootcert.pem
Passphrase=MySecret
CheckCertificateIP=1

[Gatekeeper::Auth]
FileIPAuth=required;Setup

[FileIPAuth]
; allow all calls from local network
192.168.1.0/24=allow
; only allow TLS encrypted and authenticated calls from elsewhere
any=onlyTLS

[RasSrv::PermanentEndpoints]
; the GnuGk in the other location, serving prefix 02
1.2.3.4:1300=remote-gw;02

[EP::remote-gw]
; use TLS to call remote GnuGk
UseTLS=1 

Configuration for GNU Gatekeeper 2 (prefix 02)

[Gatekeeper::Main]

[RoutedMode]
GKRouted=1
H245Routed=1
CallSignalPort=1720
AcceptUnregisteredCalls=1
; make sure H.245 gets tunneled for TLS
H245TunnelingTranslation=1
; add AES media encryption if the endpoint doesn't encrypt itself
EnableH235HalfCallMedia=1
; only allow encrypted calls
RequireH235HalfCallMedia=1
; change the media key after 2^31 operations
EnableH235HalfCallMediaKeyUpdates=1

[Proxy]
Enable=1

[ModeSelection]
0.0.0.0/0=PROXY
; only use routed mode for local calls
192.168.0.0/18=H245ROUTED

[TLS]
EnableTLS=1
PrivateKey=/path/to/server.pem
Certificates=/path/to/server.pem
CAFile=/path/to/rootcert.pem
Passphrase=MySecret
CheckCertificateIP=1

[Gatekeeper::Auth]
FileIPAuth=required;Setup

[FileIPAuth]
; allow all calls from local network
192.168.1.0/24=allow
; only allow TLS encrypted and authenticated calls from elsewhere
any=onlyTLS

[RasSrv::PermanentEndpoints]
; the GnuGk in the other location, serving prefix 01
1.2.3.5:1300=remote-gw;01

[EP::remote-gw]
; use TLS to call remote GnuGk
UseTLS=1 

Other options

You could also configure the remote GNU Gatekeeper as a neighbor, but beware that the RAS traffic between neighbors will show meta data (whois is caling who) in clear text! 

See the GnuGk manual section on TLS for more details and examples how to generate the OpenSSL certificates. 

 

 H323Plus 1.27.2 and PTLib 2.10.9.4 have been released.

Changes in H323Plus:

- support for Alpine Linux for smaller container images
- crash fixed on invalid RTCP packets
- memory leaks fixed
- GetCrytoMasterKey() restored that got lost in 1.27.1
- better support for cross-compiling
- various updates for newer compilers
- some smaller bug fixes

https://www.h323plus.org/source/

 

Changes in PTLib:

- support for Alpine Linux
- better support for cross-compiling
- various smaller bug fixes

https://github.com/willamowius/ptlib/releases

GNU Gatekeeper version 5.7 has some important bug fixes, improves interoperability
with other vendors and also has a few new features.
 

Several severe crashes and a few memory leaks have been fixed.

Improved interoperability with:

• Lifesize endpoints
• Poly's Microsoft Teams gateway
• Polycom RealPresence Capture Server

New features:

• You get a warning in the GUI / on the status port if one of your endpoints has an incorrect time setting and this password authentication fails. This makes trouble shooting a lot easier.
• Invalid TPKT packets (eg. due to network errors) now don't necessarily take down an otherwise healthy call. Use the new  AbortOnInvalidTPKT=0 switch to enable.
• GnuGk will now also return unused memory back to the OS periodically to make it available again to other applications on the same server.
  
• You have a new %{Vendor} variable for SqlAuth RegQuerys and LuaAuth

Full change log:

- BUGFIX(ProxyChannel.cxx) fix crash on non-standard H.245 Indication from
  Polycom RealPresence Capture Server
- BUGFIX(ProxyChannel.cxx) fix possible crashes on non-standard generic information in OLCs
- print warning message on status port when passwords get rejected due to wrong time
- BUGFIX(httpacct.cxx) fix memory leak
- BUGFIX(ProxyChannel.cxx) fix possible crash
- BUGFIX(gk.cxx) avoid crash when terminating in the middle of program startup,
  set non-zero exit code so restarter notices error
- return unused memory back to OS periodically
- new switch: [RoutedMode] AbortOnInvalidTPKT=0 for more graceful handling of network errors
- BUGFIX(gk.cxx) fix for running on Alpine Linux (needs updated PTLib, too)
- don't start GnuGk if RTP multiplexing is configured, but we can't start the listener
- new switch: [RoutedMode] MatchH239SessionsByType=0 to fix presentations with
  LifeSize endpoints over Poly's Microsoft Teams gateway
- BUGFIX(ProxyChannel.cxx) make sure we don't set RTP address on multiplexed RTCP keepalive
- BUGFIX(RasSrv.cxx) look at all tokens for H.235.TSSM
- add %{Vendor} variable for SqlAuth RegQuery and LuaAuth

GNU Gatekeeper version 5.8 has been released with a number of bug fixes and a few new features.

To stay updated on new releases, please also follow us on Twitter!

Improved interoperability with:

• EdgeProtect
• Avaya
  

 New features:

• experimental support for Avaya's non-standard version of H.323 (./configure --enable-avaya) (thanks Konstantin Prokazov)
• consider RFC 6598 shared network space (100.64.0.0/10) and Zeroconf (169.254.0.0/16) as private IPs
  
• new switch [Proxy] AllowSignaledIPsFrom= to skip auto-detect for messages received directly from certain IPs when IgnoreSignaledIPs=1
• new switch [Proxy] AllowAnyRTPSourcePortForH239From= to handle incorrect RTCP addresses in H.239 OLC (EdgeProtect interop)
• new switch [RoutedMode] MatchH239SessionsByIDOnly= to never attempt to match a H.239 reverse channel by type for improved interoperability with EdgeProtect
• new switches to set the HTTP Content-Type header in HttpAcct, HttpPasswordAuth and Routing::Http
• new switch [Routing::Http] JSONResponse=1 to send more flexible routing data in the HTML reponses
• many new status port shortcuts (see manual section for details)

Bug fixes:

• fix H.460.18/.19 on multi-homed servers
• fix race condition when handling H.460.19 multiplex IDs
• fix media loop on half port-detected channel when media is very early
• fix Net-SNMP query for total bandwidth
• save RTCP address from OLC for port-detection
• always check AllowSignaledIPs= before applying IgnoreSignaledAllH239IPs or IgnoreSignaledPrivateH239IPs
• handle extensions and CSRC in RTP header with H.235 half-call media
• better endpointIDs on Windows when compiling without OpenSSL

GNU Gatekeeper version 5.9 is out with a number of bug fixes and a few new features.

Download: https://www.gnugk.org/h323download.html

New features:

• new switches [Proxy] CachePortDetection=1 and CachePortDetectionDuration= to cache port detection packets for faster media connects when IgnoreSignaledIPs= is active
  
• new switch: [EP::] ForceTerminalType=
• new place holder for port notifications: %t for port type
• experimental: better error recovery if multiplexed RTP sending fails

Please note that Radius support is disabled by default now. You can enable it with the --enable-radius switch when running configure.

Bug fixes:

• fix bug in port detection with AllowSignaledIPsFrom=
• when DNS name resolves to IP without alias, remove alias from ACF completely (Cisco interop)
• remove RTP session 0 from internal tables once H.245 master has assigned a session ID
• fix compilation of Avaya support
• initialized cmsg struct to zero before using
• fix regression introduced with MatchH239SessionsByIDOnly= switch

 

GNU Gatekeeper version 5.10 has been released.

Download: https://www.gnugk.org/h323download.html

This is a bug fix release.

Bugs fixed:

• fix a crash when handling the MasterSlaveDetermination message
  
• fix the documentation of [RasSrv::LRQFeatures] NeighborTimeout and consistently treat the value as 10th of a second in the program
  

GNU Gatekeeper version 5.11 has been released.

Download: https://www.gnugk.org/h323download.html

This is a bug fix release with a few new features added.

An important bug in the handling of the ExternalIP switch has been fixed.

We also added a few features that make it easier to use GnuGk with Graphana and InfluxDB monitoring.

Changes and additions:

• remove non-working command line switch -e / --externalip, use config file to set ExternalIP
• new accounting variables %{registrations}, %{calls}, %{total-calls}, %{successful-calls}, %{allocated-bandwidth}
• new switch [HttpAcct] Authorization= to send authorization headers to support InfluxDB
• replace and in HttpAcct body with carriage return and line feed characters
• new switch: [RasSrv::LRQFeatures] PreserveDestination=1 (helpful when calling Pexip servers)

 GNU Gatekeeper version 5.12 has been released.

Download: https://www.gnugk.org/h323downldad.html

This is a bug fix release with a few new features added.

Another important bug in the handling of the ExternalIP switch has been fixed as well as Y2K38 issues.

This release also adds features:

• support for Oracle databases
• easier cloud deployment with IP detection with STUN
• better load scaling by mixing proxied with direct mode endpoints in a single gatekeeper
• Windows 64bit executables with VS2022

Full list of changes:

• enable more runtime hardening flags from OpenSSF recommendation 11/2023
• fix bug with H.245 address when using ExternalIP= switch without H.460.18/.19
• auto-detect public IP with ExternalIP=STUN and STUNServer=stun.example.com
• compiler support for VS2022
• new database driver for Oracle and new timestamp format 'Oracle'
• new switch [EP::xxx] ForceDirectMode=1 to handle all calls from this endpoint in direct mode
• BUGFIX(RasSrv.cxx, gkauth.cxx) make sure time_t is handled unsigned to avoid Y2K38 issue
• BUGFIX(ProxyChannel.cxx) check for too small packets when acting as encryption proxy

 

I have just bundeled up the changes and bug fixes of the past 2 years and released PTLib 2.10.9.6.

Most notable in this release is working IPv6 on *BSD, macOSX and Solaris as well as support for newer compilers and many small platform fixes.

Since PTLib is the foundation for the GNU Gatekeeper and many H323Plus projects, all these improvements get propagated into those projects as well.

Changes:
- IPv6 support fixed for *BSD, macOSX and Solaris
- support for newer compiler, eg. gcc 13 and VS2022
- support for C++-17
- support for Win64 builds
- support AIX as platform
- small OpenBSD fixes
- other small fixes

Download from https://www.h323plus.org/source/

GNU Gatekeeper version 5.13 has been released.

Download: https://www.gnugk.org/h323download.html

This is a bug fix release with a no new features added. It fixes an issue with the RTP port detection and 2 other minor issues.

I like the convenience of a charger for my phone in my car or by my desk at the office. The constant plugging and unplugging a micro-usb cord is a bit harsh though, a least from a first world problem perspective. I ran across a post on the XDA-Developers forum that described modding a Wireless Charger […]

The post A Magnetic Mount for a Wireless Fast Charging Dock first appeared on robotthoughts.

I am fascinated with the success of the Amazon Echo. A company founded on selling books has worked very hard to become a hardware powerhouse and I think they achieved that goal with the Amazon Echo. I bought an Echo to play with home automation but when Amazon posted instructions on how to build your […]

The post Building An Amazon Echo on the Raspberry Pi Model B Revision 2 first appeared on robotthoughts.

I’ve recently been looking at building instruments using a 3D printer. I’ve been playing the guitar for over 40 years but most of the builds for 3D printed guitars seemed a little daunting as a place to start. I found a couple of 3D printed ukuleles that looked like quicker one day builds. I finally […]

The post Build a 3D Printed Ukulele first appeared on robotthoughts.

In a couple of weeks time, it will be the first anniversary of a 35-year old intellectual killing himself on the steps of a church on the Harvard campus.

I discovered Mitchell Heisman’s “Suicide Note” via a concise article on responses to the story.

I’ve been reading “Suicide Note” since I found the article, on and off. Mitchell might have benefitted from an editor, but there is no doubt the work is philosophically an opus par excellence.

Nihilism is not my thing - I do not agree with his core philosophy that life is entirely without meaning - but the way he gets there, and some of the ideas he presents are wonderful. There are things to take away from it all that will likely resonate with me for the rest of my life - as works by all good philosophers have.

To this day, Wikipedia have repressed information about him based on a subjective rules that don’t recognise that the guy’s work is actually worth reading. I expect in due course academics will start to cite him, and that situation will change.

Out there is a growing movement to recognise him. There have already been calls from some - perhaps over-excited - individuals for him to be award a Nobel Prize in literature posthumously. I wouldn’t go that far, but I would encourage those who can deal with it to consider his work. 

This is not an exciting story, despite the title. But it’s true. And it happens to dozens of people every day, and is the reason why getting to work in London can sometimes take so long.

First, let me explain that this is not a story of me causing a fire alarm to go off, for anti-terrorist police to close a station for half an hour, or some dramatic incident that has left TfL seeking an ASBO against me.

This is a story that starts with a strap of a backpack. This strap, in fact:

This morning I caught a tube from Baron’s Court on the District Line heading East. Normally I change at South Kensington for a Circle Line to Moorgate, or hop off at Mansion House and walk up to the office through the City. This morning I had decided to stay on the District line until Blackfriars, and change there for a Circle line. It’s a man’s perogative, etc.

The tube this morning was very busy. During the Olympics it has on the whole been very quiet, but this morning it was the normal 8:15-8:45am peak time crush. I was stood right next to the door at the very front of the train, crushed in by about 20 other souls attempting to share the exact same square foot I was stood on.

At Victoria, as is often the way for the District Line, a lot of hustling and bustling went on as people fought their way out to the platform, and others tried to struggle onto the train. After around a minute, the doors closed.

Except for the one next to me. Looking down, it was jammed on my bag strap.

Swearing, I attempted to free it. It was jammed solid because the hydraulic pressure of the door was pushing against it, but not with sufficient force for the door to close. The guy next to me tried to help. The guy on the platform waiting for the next train also tried to help. Neither of us could free it. Moving it simply led to the door moving along a bit, keeping the strap jammed.

Then the sound of hydraulics releasing was heard, all the doors on the train went to open, and the driver climbed out of the cab. The release of pressure had allowed me to unjam the strap, and recover it into the train. The driver confirmed we were all fine, climbed back into the cab, closed the doors, and off we went.

I apologised to those around me for delaying their journey, even though the total delay was perhaps 60-90 seconds.

Then realised everybody else on the train was delayed, too.

Then a thought about queuing theory and a little knowledge about how loaded that line is with train traffic at that time of the morning hit me: I had delayed tens of thousands of people.

Let me explain how I worked this out.

The District Line is composed of rather large gauge trains. I estimate that conservatively, each train is capable of shifting 2,000 people during peak times. There were certainly at least 2,000 people on my train this morning. Yes, they are only 6 carriages each, but each is certainly capable of holding nearly 350 people, and frequently does. I’m prepared to revise my numbers down if shown evidence.

In addition, the District Line platforms are not just used by the District Line. They’re also used by the Circle line between Gloucester Road and Tower Hill.

A glance at any “passenger information display” on a platform along this part of the network during rush hour will tell you the mean time between trains is 1 minute. There are close to 60 trains an hour going along that piece of track during rush hour.

Because my train was delayed for over a minute, this must have caused the train behind it to be given a red signal. This in turn would have caused the train behind that to be given a red signal, and so on. This buffer effect would be dampened beyond Gloucester Road going West, because the Circle and District lines diverge, giving more time for the red signals to switch to green, meaning scheduled trains would not have to stop in an unscheduled manner.

However, there would have been at least - I think - 5 trains affected by this delay in addition to my own. So we’re now up to 12,000 people in total delayed by my bag strap jamming a door.

It gets worse.

I changed at Blackfriars to a Circle line train. I got off the train I had delayed, waited 60 seconds on the platform and got on the Circle line train immediately following it, obviously now delayed. Cautiously making sure my bag was far from any doors, I boarded aware this train was now at least 2 minutes late against schedule.

Satisfied at the figure I had come up with of around 12,000 delayed passengers, I had assumed I had done no more damage, until we got to Aldgate.

The tube system has a tendency to expect passengers always want to be moving all of the time. Any delay of more than a minute or two at a station is always explained via an announcement. As we sat at Aldgate, the driver announced we were being “regulated” by a red signal. Looking out of the window, I could see an East-bound Metropolitan line train crossing our tracks to head across to East London.

That’s when it hit me. We were “out of position”. The train was a couple of minutes late, and so the guys running the switching had decided to give priority to the Metropolitan Line train, and we were held for approximately 4-5 minutes.

Whilst this part of the Circle line between Aldgate and Tower Hill was not as busy as the District/Circle line Tower Hill back West, a 4 minute delay was enough to ensure that the train behind us was going to be red signalled waiting for us to clear the platform.

That would be enough for the train behind that to be stopped.

And that would be enough for the train behind that to be stopped, which would probably be on the shared part of the network. That would be enough to cascade across the whole part of that line back to Gloucester Road, causing delays to perhaps 12 trains in total.

By now the numbers per carriage were down a little as we were close to the end of peak, but there was probably at least 1,000 people per train out there. Rounding up for the few more probably still around the Victoria area, and we’re up to 25,000 people.

There’s obviously some fudging here - people boarding trains at the “correct time” for them, did not realise the train they were getting was in fact the one after the one they had expected, and they did not suffer any delay. But I also suspect that this effect wasn’t dampened until after the peak ended at around 9:30am, and there were people who boarded their trains at 8:30am or before still out there (it can take 60 minutes easily to get from the “end” of a line into central London), whose journey had taken at least a few minutes longer than normal.

I doubt many noticed. I doubt anybody cares.

But it did make me think about how queueing theory applies to real world problems, and how when TfL moan about people keeping coats, bags and belongings clear of the doors, or jamming the doors to squeeze on rather than wait 6 more minutes for the next train, that they might have a point.

If you cause a train to be delayed, you are not simply inconveniencing the dozen or so people glaring at you in your vicinity. Or the people on the rest of the train who would glare at you if they could. But in fact, you have a cascade effect down the rest of the network. Tens of thousands of people delayed, because you didn’t want to wait 5 minutes. Or because you didn’t keep an eye on your belongings near the door.

I’ll certainly be more careful in future.

The next time I’m sat waiting for a signal to clear or am told that we are “being regulated”, I’ll wonder about whose bag or foot was to blame, and how the numbers of people flowing through London make butterflies flapping their wings on the network capable of huge cascading effects on transport infrastructure.