DfT Incompetence: Routeing Guide

[glynn80]

I didn't want to further hijack the Woking to Slough thread so have started a new one regarding a current FOI request with the DfT, which I need some help on.

The FOI request I originally placed with the DfT was for the following:

I am requesting for all instances ATOC has written to the DfT to request for changes to the National Routeing Guide (http://www.atoc.org/rsp/Routeing_Guide.asp). I would like copies of all correspondence between yourselves and ATOC related to this issue and all decisions ruled by yourselves on said changes. Lastly I am requesting for the criterion used by the DfT when making decisions on change approvals to the Routeing Guide.

This has since been narrowed on cost grounds to all changes to the Routeing Guide since 2005 and the criterion used when assessing decisions on change approvals.

The problem I currently have is this latest reply from the DfT:

Dear xx,

Request for Information under the Freedom of Information Act 2000

Firstly, I would like to address once more for clarity, the scope of the formal requests you have made relating to the ATOC National Routeing Guide under the FOI.

As explained in the initial letter to you changes to the NRG may be needed from time to time, if errors are discovered, new rail routes opened or reopened, additional routes are to be allowed, or train operators propose to disallow a route. The ATOC document, the Ticketing & Settlement Agreement (TSA) allows changes to permitted routes to be made. The TSA requires ATOC to obtain approval from the Secretary of State for any changes, and that the SoS consults with Passenger Focus to ensure that passengers’ routing flexibility is preserved. Approximately 20,000 flows were amended in and around 2001/2 through this change process. These items would prove rather time consuming to retrieve and disseminate and there would be consequent costs.

Since the Department for Transport have been responsible for the change approval process to the NRG no such requests have been received at all from ATOC.

Moving on to address your general concerns regarding the integrity of the ATOC National Routeing guide I should point out that the guide on the internet functions on extremely complex software and is dynamic, being subject to routine maintenance by ATOC at the request of train operators and passengers. Many of these maintenance functions are driven by the need to make additional permissions for passengers, typically changes are made to allow additional routes during engineering blockades. Others involve the development of additional mapping to illustrate more clearly the actual routes which are permitted from one station to another. These routine functions do not require the use of the formal change process because they are either relaxations of restricted routes or clarifications of existing routes

As xx has explained to you, if you believe there have been substantive changes to the guide that has reduced passenger choice between two stations it really would be best to approach ATOC in the first instance providing them with the fullest details so that they can investigate. Given the necessarily tortuous complexity of the suite of software architecture that supports the guide mistakes can be made and I am sure they would welcome your positive input. If you are not satisfied with their response then a dispute on any flow and the permitted routes that remained outstanding would be formally referred to the Department by ATOC for a judgment to be made.

I have also spoken with xx at ATOC regarding this matter, he is the xx Manager, for National Rail Enquiries, 3rd Floor, 40 Bernard Street London, WC1N 1BY. He is keen to ensure the continuing accuracy and integrity of the guide and has told me he would be prepared to examine any written specific station to station list of submissions from you regarding the accuracy of the guide.


Yours sincerely

xx

Either the DfT are completely incompetent with regard to their functions with the Routeing Guide or they are deliberately trying to mislead myself.

So just to try to narrow some of the main points of the above:

1. The DfT have confirmed "all" changes to the routeing guide are subject to Secretary of State Approval: "The ATOC document, the Ticketing & Settlement Agreement (TSA) allows changes to permitted routes to be made. The TSA requires ATOC to obtain approval from the Secretary of State for any changes"

2. The DfT claim that since 2005 (when the received responsibility for the routeing guide) there has been no requests from ATOC to the DfT for approval on changes to the Routeing Guide: "Since the Department for Transport have been responsible for the change approval process to the NRG no such requests have been received at all from ATOC."

3. The DfT claim the routeing guide functions using complex software despite the fact it is a group of PDF documents: "Moving on to address your general concerns regarding the integrity of the ATOC National Routeing guide I should point out that the guide on the internet functions on extremely complex software and is dynamic, being subject to routine maintenance by ATOC at the request of train operators and passengers."

4. They claim the apparent complex nature of the software makes in inevitable errors will appear within the guide. "Given the necessarily tortuous complexity of the suite of software architecture that supports the guide mistakes can be made and I am sure they would welcome your positive input."

5. It seems they wish for me to contact ATOC to highlight anomalies within the guide and so that they can correct these. "He is keen to ensure the continuing accuracy and integrity of the guide and has told me he would be prepared to examine any written specific station to station list of submissions from you regarding the accuracy of the guide. "

Now what I need help on, is from those on the forum who remember a specific change to the routeing guide occurring, I do have some RG copies dating back to 2005 and can therefore look these up, but don't want to spend hours searching for them.

Thanks in advance

glynn80

 

[hairyhandedfool]

Newcastle-Carlisle? (but not Carlisle-Newcastle!)

 

[Bayum]

I just wanna know what it is with the

xx

At the end of the page =P

 

[First class]

I have an ATOC 2002 Routing Guide Folder with all original maps and routing points etc inside...

 

[theblackwatch]

I just wanna know what it is with the

xx

At the end of the page =P

He's simply removed the names of everyone from the letter to protect the innocent (or guilty!?). Nothing more than that!

 

[caughtonatrain]

Here's a copy of some recent correspondence I have had with ATOC re the disappearance of ZZ routes from the Routeing Guide. Although I don't have specific old copies of the RG to back this up, as you can see ATOC agreed with my assertions and promised to do something. As of the most recent update to the RG the changes (reinstatements) they promised have not materialised.

Dear xx

Thank you for your email. I am sorry that it has taken longer than
normal to reply. It was necessary to consult with the Train Companies
involved.

In noting your comments about a journey from Sheffield to Scotland using
the Caledonian Sleeper, I am pleased to confirm that we agree with your
views and will update the Routeing Guide accordingly.

Thank you for bringing this matter to our attention.

Yours sincerely

xx
Customer Relations Manager

-----Original Message-----
From: xx
Sent: 28 April 2009 09:37
To: Atoc Correspondence
Subject: Routeing restrictions from Sheffield involving Scotrail Sleeper

Dear Sir/Madam

I have recently been affected by seemingly illogical restrictions in the
National Routeing Guide, which determines route and train availability
between my origin and destination stations.

I regularly have cause to travel between Sheffield and points in
Scotland including (but not limited to), Upper Tyndrum, Fort William,
Inverkeithing and Aberdeen.

It suits my requirements to use the Scotrail Caledonian Sleeper service.

However I note now that the Sleeper is no longer permitted route for my
journeys. Consequently, rail enquiry staff are not able to help me and
it is impossible to find a ticket using online systems.

I used ATOC's Online Routeing Guide to check this and, indeed, sleeper
services ("map ZZ") are not listed for journeys where Sheffield is the
origin point. I am confident that they used to be, hence my ability in
the past to find timetable information and tickets for these journeys
involving the Sleeper using National Rail Enquiries and other websites.

However, in the Routeing Guide, Sleeper Services ARE available from
Doncaster - which seems highly illogical given than it is closer to
Scotland than Sheffield and on a more direct route (i.e, the East Coast
line).

Could you please clarify this matter for me, and reinstate permitted
routes using the sleeper for Sheffield departures.

With many thanks in anticipation of your response,


Yours sincerely

 

[tony_mac]

I wonder how much longer the routeing guide will be in existence for. It seems to be difficult for them to maintain it properly, and they obviously don't appear to be doing so.

The DfT seem to think that we are already using some computerised system (RJIS?), so I wouldn't be surprised if we are soon forced to. They have already removed the ticket restrictions from public documents, could they not also insist that we ring NRES to check our routes as well?

 

[me123]

If the routeing guide is to continue, someone needs to sit down and re-work the whole thing out. It's too complicated for a normal person to use, which defeats its purpose. And it's totally out of date and unworkable in practice. Logic really should start to be applied with rail routeing.

For example, Edinburgh to London is a crazy output. Why not just have via ECML or WCML? In reality, they're the only two routes that would be used; I don't see many people going Edinburgh to Carlisle to Leeds to Sheffield via Barnsley to London St. Pancras when there's a basic hourly ECML service. And I know there'll be more crazy ones out there.

 

[ikorodu]

I wonder how much longer the routeing guide will be in existence for. It seems to be difficult for them to maintain it properly, and they obviously don't appear to be doing so.

The DfT seem to think that we are already using some computerised system (RJIS?), so I wouldn't be surprised if we are soon forced to. They have already removed the ticket restrictions from public documents, could they not also insist that we ring NRES to check our routes as well?

I agree, for as long as its in the public domain, errors will be discovered and ATOC will finally accept that its not cost effective for them to maintain it.

It will become another seceret of the railways!

ikorodu

 

[krus_aragon]

Now what I need help on, is from those on the forum who remember a specific change to the routeing guide occurring, I do have some RG copies dating back to 2005 and can therefore look these up, but don't want to spend hours searching for them.

I have an ATOC 2002 Routing Guide Folder with all original maps and routing points etc inside...

I've only started looking at the routeing guide in the past few months, but map WS "London Marylebone to Wrexham" yells WSMR to me. Could this be one of the specific changes we're looking for?
--- old post above --- --- new post below ---
And another thing...

3. The DfT claim the routeing guide functions using complex software despite the fact it is a group of PDF documents: "Moving on to address your general concerns regarding the integrity of the ATOC National Routeing guide I should point out that the guide on the internet functions on extremely complex software and is dynamic, being subject to routine maintenance by ATOC at the request of train operators and passengers."

4. They claim the apparent complex nature of the software makes in inevitable errors will appear within the guide. "Given the necessarily tortuous complexity of the suite of software architecture that supports the guide mistakes can be made and I am sure they would welcome your positive input."

Going on my experience of computer science and programming in general, I'd disagree with the DfT's tome in the above. Finding valid routes from the routeing guide isn't at all difficult: it's a simple matter of graph theory, and there's no need for an elegant algorithm to find them: an iterative (brute force) jobbie will do just fine, if you're using your own computing resources. Clive (of CORE fame) wrote a perfectly serviceable database and front-end for this, which only stops looking for solutions to avoid using too much CPU time on a public web server. The easements are the difficult part to code, but that's because they're kludges bolted on to the system, without a regular pattern to them.

The complicated and 'tortuous' bit is exclusively on ATOC's side of things. The ~200 routeing points give over 40000 journeys, which they try to describe with combinations of only 100 maps. The low number of maps was because the original routeing guide was a paper beast to be used by human beings. With so few, there will be some routes generated that ATOC didn't really want in the system, but they made the darned system in the first place!

So how can they fix it? Either with more easements, or using a lot more small maps instead of the large ones that they use right now. Of course I think more maps is the better idea, but it's easier in the short term to slap another 'plaster' easement on the old system.

If they do this, more maps will make it harder for individuals to work out routes on paper. This would be a danger, because it'd give ATOC an even greater license to say "sorry, you misread the guide somewhere".

If would be helpful if ATOC were to clarify whether the PDF-based routeing guide is the canonical document that the NR Enquiries etc. are implementations of. If it isn't then we need to badger them to give public access to the canonical guide. Giving us NRE alone doesn't cut it, as it doesn't show all the permitted routes that we may want; it seeks fastest journeys instead.

 

[tony_mac]

Please feel free to ignore this, but as a computer scientist, I just couldn't resist!

Finding all the paths in a graph is actually very difficult (finding a longest path is NP-hard, finding all paths must include finding the longest path so is at least NP-hard - there is no known way of computing this in polynomial time).

Just out of interest, I recently ran a program to search all of the routes between Aberdeen and London. After about 10 days I stopped it; it had counted something like 450 million possible routes.

Restricting to something sensible (like twice the shortest route) would be possible.

 

[krus_aragon]

Please feel free to ignore this, but as a computer scientist, I just couldn't resist!

Finding all the paths in a graph is actually very difficult (finding a longest path is NP-hard, finding all paths must include finding the longest path so is at least NP-hard - there is no known way of computing this in polynomial time).

Just out of interest, I recently ran a program to search all of the routes between Aberdeen and London. After about 10 days I stopped it; it had counted something like 450 million possible routes.

Restricting to something sensible (like twice the shortest route) would be possible.

Sorry, I'd been thinking in terms of the simplicity of writing a program, not in terms of complexity. You're quite correct that it's an NP-hard problem: you don't know the full solution until you try every route. But with regard to Aberdeen-London, there are only two maps permitted for that journey, either AS or ZZ. Are you sure you found 450 million valid routes?

 

[yorkie]

For example, Edinburgh to London is a crazy output. Why not just have via ECML or WCML? In reality, they're the only two routes that would be used; I don't see many people going Edinburgh to Carlisle to Leeds to Sheffield via Barnsley to London St. Pancras when there's a basic hourly ECML service. And I know there'll be more crazy ones out there.

I strongly disagree. Why shouldn't they go via Leeds? What would you propose charging someone who went via Leeds? Would you propose banning Glasgow to London via the shortest route too?

 

[tony_mac]

Sorry, I'd been thinking in terms of the simplicity of writing a program, not in terms of complexity.

Yes, it's reasonably easy to write - but you can't brute force it; CORE gives up for a reason!

But with regard to Aberdeen-London, there are only two maps permitted for that journey, either AS or ZZ. Are you sure you found 450 million valid routes?

I have made a few mistakes entering the map, and I haven't checked for doubling-back that isn't at a routeing point, but everything else looks ok and I think that it's in the right sort of area - the output looks sensible.
I could give you the code (java) or output if you would like.