Friday, 2 May 2014

Typical Traveline Travesty [3]

Of course, the Traveline solution to a journey from Smallbrook Junction to Tesco (Ryde, Isle of Wight) was utter tosh. See "Typical Traveline Travesty [2]" (read again).
The whole Traveline journey takes two hours and twenty five minutes.

Traveline does, fortunately, have a less stupid answer.
This misses out the double run with Island Line in the wrong direction via Brading and replaces the Wightlink return trip to Portsmouth with a Hovertravel return trip to Southsea.
The faster Traveline journey takes one hour and forty five minutes.

One can speculate as to the reasons for this farce, a farce for which we taxpayers are (directly or indirectly) paying about £80 million a year. When Traveline was set up, the expert opinion was that it was "impossible" to manage a national database of bus services; this, despite the fact the Southern Vectis was displaying their xephos (impossible and national) system! Because the Vectis solution was "impossible", lots of people at Southampton University spent lots of public money designing clever (?) and complicated software to integrate separate regional databases. fbb well remembers to debates and lengthy reports outlining the proposed solutions for bus services which had the temerity to cross the boundary from one area to another.

Of course this wasn't a problem for journeys from the isle of Wight to Hampshire as they were both in the Traveline South East region. But then Hampshire, rightly fed up with the poor quality of its designated area, jumped ship and joined the (not much better) Western region.
This left the Isle of Wight disconnected from its mother! Add this unfortunate orphanage to the complications of multi modal journeys, and, as a world renowned technologist might say ...
"The (search) engines'll not take it Cap'n!".

Of course the journey is relatively simple.

Catch an Island Line train from Smallbrook Junction ...
... and alight at Ryde St Johns Road.
Leave the station via the steps ...
... cross the road and walk a few yards down Monkton Street to the bus stop.
From here catch a service 2 or 3 to Tesco.
But not a service 8!
[fbb did not have a piccy of a 2 or a 3!]

Here is the map ...
... and here is a typical schedule:-
The correct journey time fifteen minutes.


 Next bus blog : Saturday 3rd May 


  1. I wondered if it was possible to avoid the return trip to the mainland by removing "Ferry" from the list of acceptable means of transport. But no - Traveline then tells me that the journey is impossible, and that I should allow Ferry and try again. You can't beat the system!

  2. From what you describe, this problem is presumably due to some missing pieces of data in the system - such as the existence of walk links from St John's Road Station to Monkton Street bus stop and from Ryde Esplanade Station to Ryde Interchange. Have you reported this to Traveline using their 'Feedback' facility ? I generally find they put things right fairly quickly.

    Unfortunately, I think this 'data problem' is a perennial problem with ALL journey planning systems, whatever the technology. A few years ago there was an excellent publication known as the Great Britain Bus Timetable, which you may have heard of. A huge amount of effort was obviously put into producing the first edition, but despite this a number of errors still slipped through. Taking the view that I'd only have myself to blame if these weren't corrected, I reported over 300 of these to the publishers and got a very prompt and helpful reply. By Edition 3 (12 months later), slightly under half had been put right with others in subsequent editions - demonstrating that these things can take time and rely on customer feedback to produce a product which fully meets user needs. On consideration, perhaps the last comment applies to almost any industry !

  3. A much easier solution is to treat St Johns Road and Monkton Street as the same place. The near-death xephos would equate St Johns Road Station with St Johns Road Monkton Street and describe them as such. A pre-set 5 minutes interchange time would ensure that the correct answer always works. Traveline is simply over engineered. The distance between them is no more than from one end of a bus station to the other.

  4. Likewise Ryde Esplanade station and bus station. They are one and the same. No "walk link" or similar contrivance is required. NaPTAN has a lot to answer for.

  5. Quite possibly ! But surely, however it is done, someone has to tell the system that Ryde Esplanade Station is the same place as the Bus Station - and if no-one has done that the system will produce the wrong results. Whatever you call the 'contrivance' of telling the system they are the same place there is always scope for getting it wrong or missing it out altogether. Surely Xephos must occasionally have missed such bits of data, even if they were corrected fairly quickly when pointed out !

  6. There is still no excuse for taking the passenger past a place in the wrong direction, before returning him to the same place later. Be that Ryde Esplanade today or Smallbrook Junction in yesterday's example.

    It ought to be a fundamental and simple algorithm for Traveline to build into its systems that, as a general rule, passengers should never travel south through any place, if their subsequent journey takes them north through the same place. Or other points on the compass.

    The only exceptions to this rule might be on one way systems, such as Northampton's road layout, but even these cover the same route on board the bus, so don't need a change to achieve the perversity.

  7. Petras409 is absolutely correct - no journey planner should suggest a trip passing through a place twice in opposite directions. Traveline is no exception. In this case, Traveline suggests going north through Ryde Esplanade Rail Station and then passes south through Ryde Bus Station which it doesn't know are the same place because the Walk Link has not been included. If Xephos hadn't been told that they are the same place it would also suggest a double ferry trip. In the previous day's example, the northbound 1128 from Brading does not stop at Smallbrook Junction so doesn't count. This is no different from suggesting a journey from Watford Junction to Manchester Piccadilly via Euston passing through WFJ non-stop on the northbound journey. The point I am trying to make is that nearly all Traveline problems are caused by errors in the data which can be put right if they are pointed out. Fbb's point is that the NAPTAN data definition is in his view over-complicated - resulting in the chances of data errors being much higher than using the simpler approach adopted by Xephos.

  8. The difference between xephos and traveline is simple. The people running xephos knew where the buses ran (and still do!) whereas the over engineered traveline is based on "data", not experience. It was inevitable that this would be so as those creating Traveline were largely database "managers" and not public transport users. I was at one traveline meeting where the only person to come by public transport was the guy from Southern Vectis - it was in York, so hardly challenging!

  9. Anonymous @20:41 As a matter of interest, why didn't you use Public Transport to attend the York meeting ? Presumably the other Traveline database "managers" had similar reasons to yourself, no doubt perfectly valid.