Millennium Line - Broadway Extension: Planning & Pre-Construction

[cornholio]

Originally Posted by Reecemartin

I'm going to keep it straight and simple, IT IS NOT that simple. Cracking good encryption is virtually impossible, it is not the same as DRM.

Running the ICBM's or Skytrain on floppys is just as much an accident waiting to happen, the very materials they are made out of degrade with age.

As someone with a comp science degree who instead implements systems (wont say what) for multinational corporations now in EMEA I disagree with you. Things get complicated fast and very difficult to test. If it aint broken don't fix it is usually a good rule of thumb. A system that gets the job done and continues to run for 20 years with minimal problems is golden. You keep it going until you cant because of some other reasons like integration with something new, or enhancements it can no longer handle etc.

Also agree with Jlousa.

[afreak]

Originally Posted by Reecemartin

I'm going to keep it straight and simple, IT IS NOT that simple. Cracking good encryption is virtually impossible, it is not the same as DRM.

Running the ICBM's or Skytrain on floppys is just as much an accident waiting to happen, the very materials they are made out of degrade with age.

I am someone who works in an information security role that has a major part in an industrial control environment for a very large employer; I can safely say that you're wrong.

A stable industrial control system like SkyTrain does not need to be changed and if you're concerned about the use of floppies as storage material to operate the system, I am all ears on what you have to propose but I am willing to bet that it is inadequate and will be more likely to break. I am familiar with SkyTrain's requirements and they're not simple to explain but the way it is setup is sensible.

The system as it is is not perfect but it also works and you don't go and fix that. Banks still rely on software that is 40-50 years old on that note and you can still make purchases via Interac every day without many problems.

[swimmer_spe]

It is interesting. Because of my field, I have learned about how Nuclear Power stations in Canada are having a problem. The controllers they use for the systems by law must be proven for at least 10 years. The problem is, there are no companies that keep controllers that long anymore.

While using old systems may seem like a good idea, when they start to need parts, where do you get them from? I am ex-navy. We had a challenge getting parts as supplies don't make the 30 year old parts or those companies no longer exist.

So, this upgrade may be out of necessity.

[Migrant_Coconut]

Originally Posted by Reecemartin

I'm going to keep it straight and simple, IT IS NOT that simple. Cracking good encryption is virtually impossible, it is not the same as DRM.

Running the ICBM's or Skytrain on floppys is just as much an accident waiting to happen, the very materials they are made out of degrade with age.

Not sure if funny or sad - dude, people have been hacking DVD encryption since 1999. Blu-ray encryption fell just last year. And it's not like USBs and disks can't get worn out, corrupted or scratched just the same as floppies.

There's an old joke for this:

When NASA first started sending up astronauts, they quickly discovered that ballpoint pens would not work in zero gravity. To combat the problem, NASA scientists put their heads together, and spent a decade and $12 billion to develop a pen that writes in zero gravity, upside down, underwater, on almost any surface, and at temperatures ranging from below freezing to 300 degrees Celsius.

The Russians opted to use pencils instead.

[cganuelas1995]

Originally Posted by swimmer_spe

It is interesting. Because of my field, I have learned about how Nuclear Power stations in Canada are having a problem. The controllers they use for the systems by law must be proven for at least 10 years. The problem is, there are no companies that keep controllers that long anymore.

While using old systems may seem like a good idea, when they start to need parts, where do you get them from? I am ex-navy. We had a challenge getting parts as supplies don't make the 30 year old parts or those companies no longer exist.

So, this upgrade may be out of necessity.

Yes, upgrade it to meet modern standards but don't change the technology. Old cable with a new cable.

Also, think about something like Stuxnet happening with a wireless system for the SkyTrain.

ABSOLUTE LAST THING WE NEED. Full stop.

[nname]

Originally Posted by jlousa

We can't even get our radios to work reliably in a 400ft zone downtown anymore due to the saturation of radio waves thanks to everything being connected. And that is with having a dedicated frequency via industry Canada. Hardwired is the only way to do for reliability let allow security. And no encryption is not an unhackable solution. Put anything of value worth cracking and someone somewhere will circumvent it. Heck there are probably members on here that have mining rigs set up that could brute force the keys given enough time.

I remember when Taipei's Neihu Line first opened, there are so many issues due to lost wireless signals so the system was randomly dropping trains during revenue service. There were even talk of switching to wired system because wireless was so unreliable.

Originally Posted by cornholio

If it aint broken don't fix it is usually a good rule of thumb. A system that gets the job done and continues to run for 20 years with minimal problems is golden.

As a software developer, I agree with you wholeheartedly. Upgrade a system is just asking for trouble and may require years of planning. Some companies prefer using softwares that are decades old just because they works, and they are very stable by now. Stability is the key here, not the latest greatest feature.

[cganuelas1995]

Originally Posted by jlousa

We can't even get our radios to work reliably in a 400ft zone downtown anymore due to the saturation of radio waves thanks to everything being connected.

I gotta say that the cellphone signal in the Canada Line tunnel is awful and tends to drop out and not work until I get out of the tunnel on my commute from work. It says it's connected, but is incredibly slow from Marine Drive to Vancouver City Centre, but when I go from Vancouver City Centre to Richmond-Brighouse, it's fine.

[CanSpice]

Could it be that both sides are right?

On one hand you're relying on technology that isn't manufactured any more, or is expensive to manufacture because it's outdated, or you can't find people familiar with the technology. All of these drive up the costs of maintaining the current situation.

On the other hand the newer technology doesn't have the proven track record as the old technology, suffers from different types of failure modes (some of which could be caused by external factors), or may not have the longevity of older technology.

Either way there needs to be a cost-benefit analysis done. What happens when paying someone to maintain legacy hardware becomes cost prohibitive? What happens when trying to buy replacement hardware is either impossible because it doesn't exist, or it's way too expensive? What happens when the failure rate of new technology is too high?

I'm pretty familiar with using outdated technology and the struggles that it can bring. One of the instruments for one of the telescopes I used to work for ran its software on a CP/M machine. We had a spare floppy disk for it, and only one floppy drive. When we decided this wasn't very good in case a disk or the drive broke, we tried to find replacements, and we ended up having to go on eBay to find it. While probably cheaper than traditional routes, government institutions buying mission-critical hardware off eBay probably isn't how things should be done on a regular basis.

But upgrading the software to run on a modern operating system would have been cost prohibitive, so we didn't do that (most observatories aren't exactly swimming in cash).

[cganuelas1995]

Originally Posted by CanSpice

Could it be that both sides are right?

On one hand you're relying on technology that isn't manufactured any more, or is expensive to manufacture because it's outdated, or you can't find people familiar with the technology. All of these drive up the costs of maintaining the current situation.

On the other hand the newer technology doesn't have the proven track record as the old technology, suffers from different types of failure modes (some of which could be caused by external factors), or may not have the longevity of older technology.

Either way there needs to be a cost-benefit analysis done. What happens when paying someone to maintain legacy hardware becomes cost prohibitive? What happens when trying to buy replacement hardware is either impossible because it doesn't exist, or it's way too expensive? What happens when the failure rate of new technology is too high?

I'm pretty familiar with using outdated technology and the struggles that it can bring. One of the instruments for one of the telescopes I used to work for ran its software on a CP/M machine. We had a spare floppy disk for it, and only one floppy drive. When we decided this wasn't very good in case a disk or the drive broke, we tried to find replacements, and we ended up having to go on eBay to find it. While probably cheaper than traditional routes, government institutions buying mission-critical hardware off eBay probably isn't how things should be done on a regular basis.

But upgrading the software to run on a modern operating system would have been cost prohibitive, so we didn't do that (most observatories aren't exactly swimming in cash).

You also gotta take into account that todays technology, at least consumer technology like laptops and phones and TVs and video game consoles and cars, are designed to be obsolete after the successor gets a successor. Tech manufacturers want you to buy the new one and they make them difficult and expensive to repair, and damn-near-impossible to repair at home without taking a 50/50 chance of breaking it or something going wrong or the software detecting a change in the hardware or even that the case was opened and prompting you to take it into a repair centre, at which point it's likely the same price and/or cheaper and definitely takes less time to just buy a new one.

I have a Kenwood KR-7400 from the 1970s and I found a service manual, complete with specifications and parameters of the individual electrical components and circuit diagrams of the individual circuits, online and I am using that to restore it. It works like a charm and it actually much better for my needs as opposed to those digital receivers, without the unneeded complexity, because all I need is a connection from my computer and a left and right speaker. I don't need to switch inputs or have special modes for sports or movies or whatever.

While I agree that the system should be updated to more modern standards, I strongly disagree that we should change the technology to wireless because that also means we likely have to overhaul the rolling stock and do a lot more testing and do more training on the maintenance, which make it more expensive, as well as increases the amount of things that can go wrong.

[Sheba]

As long as we're totally off target here... Mi Amiga: One Michigan School District's Three-Decades-Old Hero Computer That Still Manages HVAC Today

Getting slightly back on track, which line do you think Corrigan would prioritize, Broadway or Surrey? Personally I'd go with Skytrain over LRT, so Broadway would be first of those two. I also think Skytrain down Fraser Hwy is higher priority than the L Line, and as the Fraser Hwy line would be elevated it could be broken down into phases - then I would do a first phase at the same time as Broadway (which being a subway is going to take a long time).

[MIPS]

Originally Posted by swimmer_spe

It is interesting. Because of my field, I have learned about how Nuclear Power stations in Canada are having a problem. The controllers they use for the systems by law must be proven for at least 10 years. The problem is, there are no companies that keep controllers that long anymore.

I seem to recall this issue was raised when we had that first big breakdown a few years back. While we do maintain a stockpile of replacement parts, the operators BCRTC contracted are typically only there for a few years before moving on. As a result when the COMM board failed there was nobody who knew immediately what the lamp status (blinking at a different speed) of a failed card looked like, which cost precious time in troubleshooting the fault and may of ultimately cost TransLink the day to shutdown and manually restart the entire system.

While using old systems may seem like a good idea, when they start to need parts, where do you get them from? I am ex-navy. We had a challenge getting parts as supplies don't make the 30 year old parts or those companies no longer exist.

So, this upgrade may be out of necessity.

This is also true. While I believe that it's cool to keep some systems around forever there are factors you have to recognize in how cost effective it may be to maintain it.
One example was how fancy the control room looked in the 80's as opposed to now. It was cheaper to rebuild the entire room and all the consoles than to constantly modernize it with equipment retrofitted into the old desks and console. The original VCC's were also upgraded at one point, then again. I think right now they exist solely as virtual machines running on commodity servers. Another example was the high voltage rectification module in the Mark I. The original module was obsolete by the end of the 90's and it was cheaper to design a newer unit that was 25% smaller but allowed the MK I to remain in service. By the way, it has since been replaced again with a newer design that is 50% smaller than the original module.

Leaky coax is very straightforward. In a nutshell you are just running an RF signal through a length of cable that has no shielding, allowing the signal to propagate out of the wire and be picked up by an antenna six inches away. The cost of designing new repeaters for the cables is likely far cheaper than reimplimenting the signalling system.

[swimmer_spe]

Originally Posted by CanSpice

Could it be that both sides are right?

On one hand you're relying on technology that isn't manufactured any more, or is expensive to manufacture because it's outdated, or you can't find people familiar with the technology. All of these drive up the costs of maintaining the current situation.

On the other hand the newer technology doesn't have the proven track record as the old technology, suffers from different types of failure modes (some of which could be caused by external factors), or may not have the longevity of older technology.

Either way there needs to be a cost-benefit analysis done. What happens when paying someone to maintain legacy hardware becomes cost prohibitive? What happens when trying to buy replacement hardware is either impossible because it doesn't exist, or it's way too expensive? What happens when the failure rate of new technology is too high?

I'm pretty familiar with using outdated technology and the struggles that it can bring. One of the instruments for one of the telescopes I used to work for ran its software on a CP/M machine. We had a spare floppy disk for it, and only one floppy drive. When we decided this wasn't very good in case a disk or the drive broke, we tried to find replacements, and we ended up having to go on eBay to find it. While probably cheaper than traditional routes, government institutions buying mission-critical hardware off eBay probably isn't how things should be done on a regular basis.

But upgrading the software to run on a modern operating system would have been cost prohibitive, so we didn't do that (most observatories aren't exactly swimming in cash).

What needs to be done has have a nationwide system that all transit uses. it needs to be upgraded about every decade.

That will make it overall cheaper for all transit authorities.

[cganuelas1995]

Originally Posted by swimmer_spe

What needs to be done has have a nationwide system that all transit uses. it needs to be upgraded about every decade.

That will make it overall cheaper for all transit authorities.

And what would the upgrades entail?

[swimmer_spe]

Originally Posted by cganuelas1995

And what would the upgrades entail?

The pieces that are not compatible. The pieces that are not going to be made anymore. The pieces that are a constant problem.

In short, when do you replace anything? When it is no longer financially feasible to maintain.

[SFUVancouver]

Originally Posted by Migrant_Coconut

Not sure if funny or sad - dude, people have been hacking DVD encryption since 1999. Blu-ray encryption fell just last year. And it's not like USBs and disks can't get worn out, corrupted or scratched just the same as floppies.

There's an old joke for this:

Ah, the apocryphal "NASA spent billions on a space pen while Russia used pencils" story.

NASA spacecraft originally used a 100% oxygen atmosphere and the mechanical switches for the controls could be bridged and shorted by free-floating conductors... like pencil graphite fragments. NASA astronauts originally used pencils in the Mercury and Gemini programs, as did the Soviets in their endeavors. Incidentally, the Soviets used a mixed gas atmosphere in their spacecraft and not 100% oxygen, so they never faced the same issues. To avoid an on-orbit Apollo 1 disaster, NASA needed to remove as many potential ignition sources and flammable materials as possible and set about systematically seeking out alternatives and methods of mitigation.

The zero-g pen itself was not developed by NASA, nor funded by US taxpayers. It was developed by the Fisher Pen Company with its own resources and then offered to NASA as a solution to its pencil problem. NASA tested it, verified it, and then began purchasing them as it did countless other items.

It's a fun juxtaposition - NASA spending money like its going out of style to develop a space pen while thrifty Soviet Union used a pencil - but it just isn't that straight forward.

[Migrant_Coconut]

Originally Posted by SFUVancouver

NASA spacecraft originally used a 100% oxygen atmosphere and the mechanical switches for the controls could be bridged and shorted by free-floating conductors... like pencil graphite fragments. NASA astronauts originally used pencils in the Mercury and Gemini programs, as did the Soviets in their endeavors. Incidentally, the Soviets used a mixed gas atmosphere in their spacecraft and not 100% oxygen, so they never faced the same issues. To avoid an on-orbit Apollo 1 disaster, NASA needed to remove as many potential ignition sources and flammable materials as possible and set about systematically seek out alternatives and mitigation...

... It's a fun juxtaposition - NASA spending money like its going out of style to develop a space pen while thrifty Soviet Union used a pencil - but it just isn't that straight forward.

Not to mention that said graphite shards floating through the spacecraft - and without any means of filtering or removing it - presents a serious health hazard. I'm aware that the joke's bogus, BUT it works perfectly as a metaphor, yes?

[Kisai]

Originally Posted by afreak

I am someone who works in an information security role that has a major part in an industrial control environment for a very large employer; I can safely say that you're wrong.

A stable industrial control system like SkyTrain does not need to be changed and if you're concerned about the use of floppies as storage material to operate the system, I am all ears on what you have to propose but I am willing to bet that it is inadequate and will be more likely to break. I am familiar with SkyTrain's requirements and they're not simple to explain but the way it is setup is sensible.

The system as it is is not perfect but it also works and you don't go and fix that. Banks still rely on software that is 40-50 years old on that note and you can still make purchases via Interac every day without many problems.

There are only two good reasons, and several bad ones to replace hardware:

1) it is impossible to source new old stock(NOS) parts. As an example, there are still millions of floppy drives out there, but nobody produces blank disks anymore. However, floppy-drive emulators do exist, and they can be produced cheaper than sourcing an OEM NOS floppy drive. Then you just stick an SD card or USB stick in that and you're golden.

2) Voltage and frequency differences in integration. Good luck installing old SCADA ISA expansions boards from a 386 or 486 into a Xeon E5.

Bad reasons

1) newer stuff is easier to manage. This is never, ever, true. Often people like accountants or marketing people like some fancy new system, that the underlying tech is actually no different.

2) newer thing costs less to run(eg less people or energy costs) as this often ignores the costs of switching and running two systems in parallel while the news stuff is trained.

3) someone sees obsolete things like floppy drives, tape drives, CD-roms, CRT screens, etc somewhere and wants to replace it because it's embarrassing to show the media and tourists.

[retro_orange]

Originally Posted by Kisai

There are only two good reasons, and several bad ones to replace hardware:

1) it is impossible to source new old stock(NOS) parts. As an example, there are still millions of floppy drives out there, but nobody produces blank disks anymore. However, floppy-drive emulators do exist, and they can be produced cheaper than sourcing an OEM NOS floppy drive. Then you just stick an SD card or USB stick in that and you're golden.

Don't speak too fast on that. On Amazon you can purchase newly manufactured and NOS 3.5" floppy discs in packs. You can also purchase newly manufactured floppy disc drives that attach to a new computer with USB. Same with NOS 5.25" floppies.

New Windows 10 compatible USB 3.5" floppy drive, only $19.99
https://www.amazon.ca/Neoteck-Floppy-Ext...d=1515561264&sr=8-5&keywords=floppy+disc

When I was touring a few Sears stores in the past few days they had bankers boxes full of unused unopened 3.5" floppy disc packs. There's alot of basements and storage rooms out there still at other businesses and companies to be found in the future.

[casper]

I think you are missing the point on understanding what is obsolete.

Your desktop computer motherboard has a commercial product life of 6 months, perhaps a year before it is obsolete and is taken off the market. The MTBF before it starts to fail may be 10 years.

The motherboard that goes into an ATM machine or another industrial PC type applications has a configuration controlled mother board. Usually the manufacture agrees to manufacture for 5 years and supply replacement for an addition 5 after it is discontinued. You pay a good premium for this and accept lower performance. The reason is is is excessively expensive to redesign your system every 6 months because you can't source what you need.

An industrial control system usually built from boards that are under tight configuration control, are available on the market for a decade or more. It is likely engineers for a much higher MTBF.

I use to work in Nuclear many years ago. It may have changed, but the issue was not that it is "proven for 10 years". The issue is that is environmentally qualified to last the operating life of the equipment plus one day. We would put the control gear on a shake table after being thermally cycled (simulated aging) to demonstrate that for its service life + one day, that it could function through an earthquate and still do what it needs to do. For the next 10-30 years the expectation is that computer system is going to do the exact same thing with minor tweeks day in and day out.

Compare that to desktop computer where the expectation is it going to play the newest video game to hit the market in 2-3 months. Completely different expectations.

I have no relationship to them, but the company that makes the data loggers in Skytrain and the control radios that go into many of these transit systems is actually based out of Victoria.

http://www.questertangent.com/

[cornholio]

Originally Posted by Kisai

There are only two good reasons, and several bad ones to replace hardware:

1) it is impossible to source new old stock(NOS) parts. As an example, there are still millions of floppy drives out there, but nobody produces blank disks anymore. However, floppy-drive emulators do exist, and they can be produced cheaper than sourcing an OEM NOS floppy drive. Then you just stick an SD card or USB stick in that and you're golden.

2) Voltage and frequency differences in integration. Good luck installing old SCADA ISA expansions boards from a 386 or 486 into a Xeon E5.

Bad reasons

1) newer stuff is easier to manage. This is never, ever, true. Often people like accountants or marketing people like some fancy new system, that the underlying tech is actually no different.

2) newer thing costs less to run(eg less people or energy costs) as this often ignores the costs of switching and running two systems in parallel while the news stuff is trained.

3) someone sees obsolete things like floppy drives, tape drives, CD-roms, CRT screens, etc somewhere and wants to replace it because it's embarrassing to show the media and tourists.

Great points and right on the mark. I have plenty of personal experience with this. The thing is paying and training people and usually sourcing custom parts (unless really impossible) is far cheaper even if expensive then a major system failure due to untested hardware/software or a risky complex system redesign. Stability in complex and critical systems comes first. People that want to replace functioning systems or hardware with something newer because piss me off in my work. Newer is not better. In fact it is generally worse as it opens up new and unnecessary risks.

Originally Posted by casper

I think you are missing the point on understanding what is obsolete.

Your desktop computer motherboard has a commercial product life of 6 months, perhaps a year before it is obsolete and is taken off the market. The MTBF before it starts to fail may be 10 years.

The motherboard that goes into an ATM machine or another industrial PC type applications has a configuration controlled mother board. Usually the manufacture agrees to manufacture for 5 years and supply replacement for an addition 5 after it is discontinued. You pay a good premium for this and accept lower performance. The reason is is is excessively expensive to redesign your system every 6 months because you can't source what you need.

An industrial control system usually built from boards that are under tight configuration control, are available on the market for a decade or more. It is likely engineers for a much higher MTBF.

I use to work in Nuclear many years ago. It may have changed, but the issue was not that it is "proven for 10 years". The issue is that is environmentally qualified to last the operating life of the equipment plus one day. We would put the control gear on a shake table after being thermally cycled (simulated aging) to demonstrate that for its service life + one day, that it could function through an earthquate and still do what it needs to do. For the next 10-30 years the expectation is that computer system is going to do the exact same thing with minor tweeks day in and day out.

Compare that to desktop computer where the expectation is it going to play the newest video game to hit the market in 2-3 months. Completely different expectations.

I have no relationship to them, but the company that makes the data loggers in Skytrain and the control radios that go into many of these transit systems is actually based out of Victoria.

http://www.questertangent.com/

Your nuclear experience is interesting. A close uncle of mine was a engineer at one of the Nuclear power plants here in Central Europe (Temelin) and I have visited it before. Actually even last summer I did another tour which included a tour of the duplicate operating room etc used for training (i.e. they have a perfect copy of everything where they train operators for years until they can enter the real operating room). They use old tech for the critical systems from the 70's early 80's and it keeps getting replaced by the same stuff. Only some of the desk monitors etc are new but the critical systems are still the same. Why? Because its tested and it works. Upgrading anything poses unacceptable risks in the nuclear power plant in the middle of Europe. p.s. as you seem to have a interest in nuclear if you ever go out there you can do a tour of Temelin. The virtual tour is quite good and covers the interior so if it einterests you check it out from the link. You can use chrome or what ever to translate to english...