Many of our readers have suggested that the TTC go back to providing a phone number attached to a transit stop. When you called the number, an automated service would tell you when the next bus or streetcar was to arrive. The TTC’s system was known as Timeline which went out of service on Dec. 31, 1999. Bob Brent, a former TTC senior manager, sent us a little history lesson on Timeline.
– – – – – – – – – – – – – – – –
Timeline was introduced around 1989 at a cost of $3.3 million with the Provincial government paying 75% of its cost (along with all other TTC capital spending).
For those of you unfamiliar with Timeline, you called a unique phone number found on the TTC sign pole at transfer stops and it would give you a digitized voice of the current time and the scheduled arrival time of the next three vehicles.
When I joined the TTC as a senior manager in January 1997, Timeline was an established service (about 1,000,000 calls/month) but the writing was already on the wall. It was running on obsolete technology that was literally falling apart with frequent outages and increasingly expensive service calls.
In 1998 an IT company, whose principals had been involved in the original Timeline implementation, came in to give a beta demonstration of a new Y2K-compliant Timeline customer information system running on a MS NT server. It was a mere shell than a mature application. It crashed a few minutes into their demonstration.
The alarm bells went off and tough questions were asked if it was ready for prime time. They admitted it was just a prototype shell to demonstrate what was possible. In fact, they wanted the TTC to be the guinea pig first installation and fund its commercial development in “partnership.†They estimate a stripped down version would cost $2 million.
It was known that Timeline could not operate after Y2K, without substantial reprogramming. Unfortunately, it was written in one of the more esoteric programming languages. TTC IT therefore advised it would be hard to find a consultant to make Timeline Y2K compatible, notwithstanding the fact they would be extremely expensive given there was such desperate demand for freelance IT contractors to make mission-critical corporate applications Y2K compatible — with money no object.
The icing on the cake for Timeline, however, was that even after being made Y2K-compliant (if possible), it was operating on only one-half of the in-bound telephone lines, as one of the two analog “phone banks†or cards had failed completely and no spares were available anywhere — new or used — at any price.
In fairness, given the popularity of Timeline I prepared a presentation to Chief General Manager David Gunn. Mr. Gunn and his GMs objectively outlined the TTC’s options. They rejected spending $2 million on Timeline Y2K software development, especially on speculative software. This was not unreasonable given their own sorry history with internal TTC IT projects like ATOS and CIS that were millions of dollars and years over budget and never delivered as promised.
The timing, however, couldn’t have been worse as Timeline Y2K would have to proceed without the 75% provincial capital subsidy — the cancellation of the subsidy having just been announced by the Harris Tories in 1998. As a result the once proud TTC was forced into rebuilding 18-year old GM “carcasses,†to quote CGM David Gunn, who was immensely proud of the rebuild program that delivered the most reliable TTC bus — the driver’s favourite — for half the cost of a new Orion carbon frame bus that was rusting after just 7 years. In this context, spending $2 million-plus on Timeline, not surprisingly, was seen as an unaffordable extravagance and went over like a lead balloon.
In June 1999 I made a similar presentation to the Commission. They reluctantly supported Staff’s recommendation on the understanding that customer information would be radically changed and expanded to soften the loss of Timeline (via the website, a new Nortel ACD switch to better route info and complaint calls, new state-of-art voicemail software that allowed route-specific information, plus a faxback capability: all for about $225K in total). We had 18 months to plan and carry out the upgrade, hoping for the best, preparing for the worst!
January 2000 was the first month of Y2K and the first without Timeline. We held our breath, anticipating a deluge of complaints calls. We received 182 complaints that first month. Frankly, we were amazed the Complaints hadn’t gone through the roof, given Timeline’s popularity! January was the complaint peak after which they steadily declined to virtually nothing within a year.
The website was really the unsung hero. It seemed to come of age at just the right time, tripling from 1 million hits/month in Dec. 1999 to 3 million hits/month in Sept. 2000. Thanks to the website and other new info offerings, the TTC dodged a bullet by developing a more diversified, expansive offering of customer information.
That being said, it was probably the most unpleasant decision I had to implement as TTC CMO. Internally, I was viewed as the uncaring “Timeline Assassinâ€. In fact, it would have been much easier to fork over the money to the developers and blame them if Timeline didn’t work as promised.
In truth, I was neutral on the issue. While I wasn’t emotionally tied to Timeline, if Timeline Y2K was affordable, I was more than willing to recommend it. With the province having just cancelled the TTC’s 75% capital subsidy in 1998, and capital dollars so scarce, developing Timeline Y2K was really a non-starter.
Fast-forward to today. Yes, it would be nice to still have Timeline’s capability. With traffic congestion so bad, however, I think “scheduled” stop announcements are passé. We need a not only a vastly improved TTC website but real-time next vehicle information by stop and real time delays by route and mode and system (take a trip on YRT’s ViVA to see real-time next bus information in action).
I think in retrospect, the Timeline Y2K decision benefited the TTC in the long term as it forced a major rethink of customer information on the organization.
– – – – – – – – – – – – – – – –
Read the comments section for clarification on some of the acronyms used in this post.
– – – – – – – – – – – – – – – –
photo from Toronto Archives: series 71, item 7085
30 comments
That’s something I always wondered about. Timeline didn’t really do anything — or convey new information — that looking at the (admittedly) convoluted schedule posted on the nearby pole did.
Without being attached to realtime information of where the next bus/streetcar is, it would be kind of useless today.
Why was it so popular? Did people just hate reading those schedules? I don’t blame them.
Acronyms, Acronyms, Acronyms
ATOS~the automatic operator sign-up for each of the TTC’s 13 or so regular plus holiday “Board Periods” where TTC Service Planning fine tunes service to match demand. I’ll leave it to Steve Munro to tell me what the “T” stands for! 😛
CIS~the communications & information system frontend is the green screen box and phone to the right of the driver behind the farebox on buses and streetcars. It allows communications with Transit Control and allows Route Supervisors in the Divisions to monitor each vehicles location along a route (via archaic wheel sensor, direction, speed, distance algorithm… not GPS).
P.S. Coupler~the TTC’s internal monthly newsletter~has a May 2000 article “Expanding TTC’s information offerings” that shows the dramatic change in TTC Customer Information between 1997–2000 @ http://www.toronto.ca/ttc/coupler/0500/brent.htm
Coupler~ is archived on the website http://www.toronto.ca/ttc/coupler/coupler.htm. from April 1998~January 2007
N.B. One more thing… Any time you want to look up an unfathomable TTC acronym or term go to ttc.ca and click on upper right on page and enter the TTC acronym or word and choose the pull down menu and it will search both Coupler and TTC Commission Reports and likely explain it.
You couldn’t see the schedule on the pole from the comfort of your home or office where you were waiting for the bus.
Many people knew the phone numbers of the stops that they most frequently used and a quick call could let you know whether you had to hustle out to the stop now or if you could wait another half hour. This became especially valuable after the cutbacks of the ’90s left some routes with 40 minutes or more between buses at night.
Also, until around the time that Timeline was discontinued, those route schedules were generally only posted at major intersections, not at the majority of the stops along the route.
It is most useful in suburban areas where buses are 30 minutes apart, but pick up during peak hours.
I know I used the similar system in Mississauga when I was dating a girl near Square One. Two different routes would stop in front of her place so I was always playing them against each other so I could spend as much time in her apartment and not stand in the cold wind-swept bus stop.
This kind of situation would apply to Etobicoke, Eastern and western North York, plus Scarborough.
hey,
that was really interesting. thanks for posting that! what a great site.
I think there’s definitely something to the idea of having a phone number on every stop for people to call and get automated information.
The ubiquity of cell phones means that people are willing and able to call a number to get more info, whereas devices to surf the web are not as available as yet, or more explicitly, people are still not actively surfing the web with their phones.
Open source telephony software such as Asterisk would make setting up a new “Timeline” relatively easy. Using Linux and open source databases like Mysql and Postresql, I could imagine a modest bit of funding going into a volunteer effort to build the software for the city. Just look at the Murmur project.
Three more things:
1. Erratum… ie.. I goofed! ;P Unique Timeline telephone numbers were assigned to ~all~ 10,000+ TTC stops (on route ID pole signs) while InfoPosts were only installed on the smaller subset (≈2,500?) of TTC transfer stops where two surface routes intersected e.g. College & Spadina.
2. Acronym omission: ACD~ is an automatic call distributor… a sophisticated telephony switch (Nortel) that routes calls to the proper extension (e.g. faxback, fares, live info or complaints operator) and keeps them in proper first in~first answered sequence in the call-waiting queue.
3. In my “N.B. One more thing…” comment above I inadvertently used HTML code (that ignores what is between it ” around “Toronto Search” so it reads: “…click on upper right…” when it should read “…click on Toronto Search on upper right on page…”
It’s also worth noting (or maybe it isn’t, but I’ll note it anyway) that although every stop had a number, not every number was unique to that stop. On my home route (Mortimer 62), two or three stops in a row would generally have the same number. The stops were as little as 150-250m apart along much of the route, so it was neither practical nor necessary for every stop to have a different schedule from its immediate neighbours.
Val, you’re right… TTC Service Planning’s schedule drove not only actual service, but Timeline, Infoposts, Telephone and Internet schedules too… and a surface vehicle was considered “on-time” if it arrived at a stop +/- 3 minutes of the schedule time posted.
I live on Steeles West and similarly there’s no difference in the Infopost/Internet times for 60 Steeles West for the two stops a block east and less than a block west of the primary stop at Bathurst and Steeles.
Val> Thanks. That makes sense. I also forgot this was pre-web (or pre-the-way-it-is-now-web) — I often do the online version of timeline now, check from home when the next bus on some route is coming, then go out when I need to. Or print out a sked for a friend’s place if they’re in the middle of “nowhere” so I’ll be able to quickly get on the TTC when needed.
So people would memorize/write down the numbers, and call them often, and remotely. Now it is useful (I arrived in TO in 2000, so i missed the Timeline era).
Jesse> Indeed — [murmur] runs on Asterisk. However, I will side with Joe Clark on this one, and say whoever makes the thing needs to get paid. Suggestions and think-tanking, meh — the TTC needs this injection of free and open creative thinking, but the making…the kids need braces.
Also, saw you on the CBC yesterday talking about online credit card safety. You made me feel safe and vigilant.
MattB> I wonder how much extra romance-time Timeline saved people in the 90s. How many 9 or 10+ year olds are there around today only because Timeline saved a few extra minutes.
I believe Timeline also provided information on all routes serving that stop, not just the trips on one route (like the infoposts). I used to use Hamilton’s version of Timeline (which is actually still operational) partly since I used an intermediate stop that wasn’t a timepoint, but mostly because there were 3 different routes I could take, all of which had different frequencies, and the phone call would provide next bus info for all three routes.
Calgary has had a similar system running for years.
You dial a number (one number for the entire system), and then an automated message ask you to type in the stop number (which is on the sign).
An automated messages tells how long it will take for the next bus, the following bus and the third bus. It is usually quite accurate, but it can be 2-3 minutes off.
Then again you could go to Vancouver’s system, where each bus is just a block away.
An Asterisk-based system with the same capacity as Timeline at its peak (1 million calls/month) could be built and run for a year for under $100,000 — maybe a lot less.
I remember the Mississauga Transit service too, from when I was a student at Erindale. It looks like it’s still in use too:
http://www.mississauga.ca/portal/residents/citylink
Mississauga, Montreal, Hamilton (HSR), Ottawa (OC Transpo) Waterloo Region (Grand River Transit) and Oakville still have systems like TimeLine. Montreal and Hamilton publish the stop numbers in pocket timetables, Mississauga publishes the numbers on its system maps at terminals and major transfer points. Most systems allow you to find your bus stop number online.
As Val Dodge mentioned, in Mississauga, at least, several nearby stops would have the same number – this is very common in Mississauga. At one time, each of these had their phone number – (ie 615-XXXX), later you now call one number one number (615-4287), then stop numbers would be entered much like a business extension. I memorized several – living in Brampton (but near the MT 19) my bus stop number was 2610 – Shoppers World.
I left Toronto shortly before 2000, so I didn’t know that Timeline was never replaced.
Having a separate number for each stop (or cluster thereof, as per above) must have been pretty expensive. Ottawa used to do it that way, too, but, like Mississauga, they now use a single number (613-560-1000) and then Bus Lady asks you to enter your stop number (7475) after you connect.
OC Transpo has a clunky-looking but extremely effective on-line trip planner which I use much more frequently than the “560” service:
http://www.octranspo.com/tps/jnot/startEN.oci
…it’ll even sketch you a map of your route.
It’s kind of shocking that Toronto has neither of these things.
On a nostalgic note, does anyone have sound samples of Timeline Woman reading off some bus times? I still remember her funny, mechanical delivery: “Next vehicle, to … Warden Station …”
Halifax has a system like this too. Each stop has its own number. I have not tried it, but assume that it would give you info for each route, as in the downtown core, 5-10 routes may serve a given stop. I think the difference is that with the possible exceptions of Vancouver and Montreal, the TTC is the only service providing a core of frequent-service routes seven days a week, whereas even in ottawa, only 3-4 routes operate at a rate better than 15-20 minutes off-peak.
Gabe, that’s pretty exciting. It’d be a really interesting project to take on, outside of the TTC aegis, and just request that the TTC make the data available necessary — basically, the bus schedules that are up at the stops, provided in a common electronic format, using constant identifiers per stop, all of which should be trivial for them — to feed the system.
Cost-wise, what exactly does the infrastructure involve? There’s a computer, there’s software (Asterisk and a database programme and some code to transform and load the data, I guess). The big cost is presumably the phone lines. Do you know much about this?
Features-wise, how much can Asterisk do? I have visions of being able to set up an “account” based on the server caller-ID-ing your phone. Or, once the phone number had been made famous, of being able not just to call that phone number and punch in a bus stop, but also to SMS the service with a bus stop number and get an SMS back.
Because, I dunno, I can see this being the kind of project where, once built, sponsorship at that kind of level wouldn’t be totally impossible. Imagine one of them new phone competitors thinking this was a good PR move…
I use GRT’s TeleRider all the time here in Kitchener – the stop number is right on the sign, you just flip open your cell, dial 519-888- and the four digits. You get the next two arrival times for every route that uses the stop. It’s even more useful these days with cellphones than when I first made use of it some 15 years ago!
I do wish the Cambridge customers could take advantage – it’s officially a legacy of Kitchener Transit, thus only works in K-W – and I can see the advantage it would certainly be in Toronto or any other city. The logical next step would be real-time feedback, something we’re supposed to have on our iXpress BRT route but has yet to materialize.
When exactly did timeline start system-wide?????
Timeline was first tested in Etobicoke (NOVEMBER 22, 1981) on eleven routes
(66 PrinceEdward/80 Queensway /110 Islington South/76 Royal York South
507 Long Branch/38 Horner / 87 Mall Circle/15 Evans/49 Bloor West
44 Kipling South & 45 Kipling
(Source : TTC Vol 1 Issue 1 TIMELINE November 1981)
Inquiring nerds want to know: exactly what “esoteric programming language” did it use? (Or is that a state secret? :])
I don’t want corporate sponsors telling me when my bus is coming. Timeline already did that when it announced the current time: “Pizza Pizza time is [anytime?] 11:37.”
The block containing the details got corrupted a long time ago but years ago I had an interview for a job at the company that made the Timeline system and IIRC the system in Kitchener. Their offices were at Front and University in the same building as the Fish House. One of the complicating issues in the whole thing was that they went bust in the late nineties. The other issue that Bob didn’t mention was that at the same time as this was happening the TTC was also developing Trapeze so anything built to work on Jan 1, 2000 would only have been used for a few months before having to be massively modified to interface with Trapeze.
Jon: I don’t remember what “esoteric” language it was, it was a commercial language but I don’t think I’d heard of it when I was told what it was. Everything I wrote on Timeline was from memory so it’s a bit rusty now after 7–8 years. A TTC Spacing reader will have to give us the answer!
I’m a little curious myself, reading the comments how other GTA properties made their Timeline code Y2K-compliant when it was so (2-digit) date intensive. They certainly didn’t have the call volume as the TTC in 1999 so perhaps they were able to re-write the code on a more leisurely basis once the Y2K scare had passed.
Remember, the Y2K “bug” is potentially only a problem if you are comparing dates to calculate the age of something. For something like Timeline it’s somewhat possible that it would run just fine with two digit dates, or not, depending on how the schedules were stored. Now it might not have been able to cope with the actual rollover at midnight but many systems could have been shut down at 11:00PM and cold started at 1:00AM on 01/01/00 and rolled along happy as a clam until 2099
First of all, thanks Bob for filling in the history of the Timeline service and the history of decisions that the TTC had to make.
Often, we (the public) bitch because we’re frustrated and can see how things can be made to work better, but don’t/can’t know what the available options were at the time.
One thing that would help the TTC on these matters is to talk about the history of the decisions they’ve had to make publicly. Understanding cuts down griping in a hurry; as does asking for involvement in working towards a common solution.
New technologies makes it possible to rework old solutions at a lower price. The key thing to make it happen is planning, a dialogue between implementer and user, a stable revenue stream and some scope for creative solutions.
That was a good read.
I remember back in 1998 when lots of talk over Y2K compliance was the worry, and that Timeline would not live past 12:01 on January 1st. It was bittersweet, because I loved the fact that I could leave my house and walk to the stop with enough time to catch the bus, since I knew exactly when it would arrive (that was when the TTC *used* to stick to schedule), and I also really liked that it was such a simple service to use.
But I also became concerned with the cost of replacing the system with something more compliant (or to even consult a Y2K geek like myself once-upon-a-time to figure it out). The website was indeed a definate improvement, and it was obvious the TTC was ramping up for a replacement of Timeline, by promoting all the different ways you could get route and schedule information from the TTC (without using Timeline). I think it’s the first clear result of the TTC working hard to make a customer-serving improvement in a long time.
Would I like to see Timeline again? Well, to be honest, it matters little to me since I don’t use the TTC anymore (I’m one of those pests who own a gas-guzzling automothingy). But the real challenge now is to take what became an instant success (the website) and work it into something really powerful. It always amazed me that governmental web properties were such a mishmash of information that didnt a) look good and b) make things easy to find. I don’t like Giambrone (at all), but I credit him and congratulate him for opening up the website can-o-worms.
For inquiring minds who want to know what the language used to code something like Timeline probably was: I’d venture Cobalt was probably used, because it was very popular back up until the mid-to-late 80’s. But really, any code language could be used to generate non-compliant software.
Maybe Timeline was written in AWK…
I agree that the TTC did its best to mitigate the decommissioning of Timeline by enhancing its customer service information, including the website, but come on — the website design would have been antiquated and nearly unusable even in 1999. I’m surprised that the TTC didn’t get it right back then, and worse still, that we’ve suffered with the horrible information architecture that is ttc.ca for so long.
Almost all Timeline-type systems now have a central number and individual stop numbers. Montreal’s STM has easily the most memorable – AUTOBUS (514-288-6287), which about 2 years ago replaced the old Telbus system that had separate phone numbers for each stop (or cluster). The general information line is now STM-INFO (514-786-4636) – the same as their web site http://www.stm.info – which has a home page link to AUTOBUS. You can type in your stop number on the home page, and get your bus times.
I was the Development Mgr for the TTC’s TeleRider procuct, so if you have any questions, contact me
at eg-dad2005@rogers.com, and hear the “other” side
of the story.
RandyF