What year was ORCATS introduced?

[alistairlees]

There's a few threads on here about ORCATS, referring to it being a BR creation. But I can't see anything specifying when it was introduced by BR. Does anyone know?

ORCATS = Operational Research Computerised Allocation of Tickets to Services (according to Wikipedia).

 

[Chris Butler]

I worked in the OR group when ORCATS was being developed and used. That was around 1980.

However, it depends what you mean by 'used'. The algorithms for estimating which trains passengers would travel on was not initially used for a routine allocation of revenue (at that time to sectors) but was used initially to understand how 'attractive' services were to passengers. The main,initial focus was on understanding the elasticity of demand to speed, frequency etc., especially related to the introduction of HSTs. A further focus was to try to evaluate the impact of services (primarily cross-country) where the impact of improvements impacted a plethora of different point-to-point passenger flows and so required an understanding of which services passenger would find attractive combined with the computing power to evaluate that for all the relevant point-to-point flows.

I don't think ORCATS was yet being used for routine allocation of revenues to sectors when I left in 1983, but it was close.

 

[alistairlees]

That's interesting, thank you. If you have any more history of ORCATS to share, I would be very interested in hearing it. Or, indeed, from anyone else who was there at the time.

 

[Dr Hoo]

The 'birth' of ORCATS was really just a continuation of developments in computer programming, cost and availability, data capture and so forth.

Even earlier BR had things like NPAAS, NFAAS and NCAAS. (National Passenger Accounting and Analysis System, or Freight or Coal). These had things like 'BUSEC' codes (Business Sector). Even earlier there was SAFGABS (System Analysis For Grant Aided Business Services), introduced to keep track of revenue, costs and subsidies for services after the Transport Act 1968. This last-mentioned was my first exposure to strange things like concertina-fold computer print-outs in days when armies of finance clerks used to wield round rulers and pencils on huge paper spreadsheets whilst spinning manual FACIT calculating machines.

Some people seem to think that it is new/odd/un-necessary to try and understand where revenue comes from and needs to be distributed to. Strangely enough even back to the dawn of railways, before pre-printed card tickets, clerks at stations used to write down ticketing transactions in ledgers - hence the term 'Booking Office'. This was in turn borrowed from stagecoach practices where a national network of services was administered from a couple of rooms above the Golden Cross pub near Charing Cross. This included reimbursement for costs of providing many hundreds of horses for journey legs between dozens of coaching inns across Britain.

 

[alistairlees]

The 'birth' of ORCATS was really just a continuation of developments in computer programming, cost and availability, data capture and so forth.

Even earlier BR had things like NPAAS, NFAAS and NCAAS. (National Passenger Accounting and Analysis System, or Freight or Coal). These had things like 'BUSEC' codes (Business Sector). Even earlier there was SAFGABS (System Analysis For Grant Aided Business Services), introduced to keep track of revenue, costs and subsidies for services after the Transport Act 1968. This last-mentioned was my first exposure to strange things like concertina-fold computer print-outs in days when armies of finance clerks used to wield round rulers and pencils on huge paper spreadsheets whilst spinning manual FACIT calculating machines.

Some people seem to think that it is new/odd/un-necessary to try and understand where revenue comes from and needs to be distributed to. Strangely enough even back to the dawn of railways, before pre-printed card tickets, clerks at stations used to write down ticketing transactions in ledgers - hence the term 'Booking Office'. This was in turn borrowed from stagecoach practices where a national network of services was administered from a couple of rooms above the Golden Cross pub near Charing Cross. This included reimbursement for costs of providing many hundreds of horses for journey legs between dozens of coaching inns across Britain.

Thank you Dr Hoo. I had never really thought about the origins of the word "booking" before. And the connections back to stagecoach days were entirely unknown to me.

 

[Taunton]

The 'birth' of ORCATS was really just a continuation of developments in computer programming, cost and availability, data capture and so forth.

This is true. There's an extended description of the BR revenue allocation process (and its misuse) in Stewart Joy's "The Train that Ran Away". Joy was Beeching's Chief Economist, also an Australian (therefore notably practical and plain speaking). Attributing revenue really came to the fore with the 1968 Transport Act, and formalised subsidy of individual routes, with through ticketing.

If you read my university thesis, possibly still in their library after all these years, you will find my notes on the inappropriate usage of this. Most management seemed to concentrate on the costs and their expenditure, probably because they found the revenue attribution side too difficult, and just hoped it would happen. I seem to have ploughed a lone path on the revenue side for a long time. Even Gerry Fiennes wanted his Basic Railway to just have its own fares, needing to book again when you reached the main network.

Even more important now we are in virus times is to concentrate not even on cost/revenue, but on actual cash in/out. There is a difference!

 

[30907]

Even Gerry Fiennes wanted his Basic Railway to just have its own fares, needing to book again when you reached the main network.

And he got it IIRC

 

[Dr Hoo]

And he got it IIRC

Yes, indeed, at least on the Eastern Region. Whilst keeping ticket issue on Paytrains simple it did, of course, completely obscure visibility of 'contributory revenue', increased ticket prices and inconvenience to passengers and sometimes even led to missed connections.

This is just the sort of thing that more sophisticated ticket issuing systems and revenue analysis programs (such as ORCATS) were designed to avoid but somehow they end up being tarred with the brush of being some ghastly precursor to privatisation with modern demands and expectations that they be abolished to "improve efficiency" or "reduce un-necessary bureaucracy".

 

[Bald Rick]

Fascinating thread, that answers a lot of questions I didn’t know I wanted to ask. Thank you all.

 

[Taunton]

Yes, indeed, at least on the Eastern Region. Whilst keeping ticket issue on Paytrains simple it did, of course, completely obscure visibility of 'contributory revenue', increased ticket prices and inconvenience to passengers and sometimes even led to missed connections.

Not completely simple. While it kept outgoing revenue visible locally, you could get inbound tickets to stations on the paytrain lines from other ticket offices or even on other Regions. I'm guessing they would be attributed based wholly on mileage, which was the old simple way of doing it. However costs are definitely not related to mileage.

One can go on for the rest of time about revenue allocation, but a general principle does seem that, whether on railway or international airlines, the major trunk haul has the senior and better placed managers who manage to bend the process to suit their side of the business. I know that Orcats tried to be completely impartial, done by an impartial team, but it still has issues. Of course, it was originally developed as an internal BR tool which just suited management accounting - being used for real cash changing hands between TOCs, Orcats Raids, and all that came later.

An interesting associated concept is the Cooper Brothers Formula. Again by an external body, this time the Cooper Brothers audit and accounting business (nowadays just the C in PWC), and devised as part of the 1968 Transport Act, it is what was added to the revenue of a line to assess how worthwhile it was to society, and to the government providing subsidy. I always liked this one (and nearly got a job with them, well after its development, although they all seemed to be formally qualified accountants). I always like to compare subsequent finance attempts, such as franchising, or what we have today, back to this approach.

 

[Chris Butler]

Most posts here are focusing on the (very long) history of revenue allocation, which has become the defining role of ORCATS. I do however want to emphasise that, at its birth, there were really three things that came together to make ORCATS possible/necessary.

First was a comprehensive set of data on ticket sales. At its birth the NPAAS data mentioned above by Dr Hoo was used. This data was pretty comprehensive, but far from perfect. Manual tickets, group tickets, rovers/rangers (to use today's parlance), and any number of tickets for which the origin and destination of the passenger could not be deduced made the data imperfect, but far better (because it was on computer tapes) than having to sample ticket sales from specific locations.

Second was the available computing power to do the calculations required. The processing needed to take all (significant) passenger flows that would be impacted by the service in question (e.g. Harrogate to Bath would be impacted by the introduction of CrossCountry HSTs) and calculate the overall timetable between those flows and how a change to the Leeds to Bristol service would impact the attractiveness of that timetable, and finally, aggregate all the revenue increases from those flows and 'allocate' it to the Leeds to Bristol service.

Thirdly, to understand how passengers reacted to changes in speed, frequency, number of changes in a services. Previously there were 'rules of thumb' that had been developed as 'distilling previous experience (especially the WCML electrification)', but it was the introduction of HSTs that gave the most significant changes at the time the data and computing power allowed systematic study of the passengers' reaction to the new service and understanding, primarily, the elasticity of demand to speed (journey time).

As I mentioned previously, the question of regular allocation of revenues was always an objective for ORCATS, but the initial focus of this part of the OR group's work was evaluating the various investment decisions in the Inter City sector, and (also a key focus) justifying that investment to the Dept of Transport and Govt.. As time went on (and I left the railway in 1983) the objective of routinely allocating revenues seemed more and more possible and so became a bigger and bigger focus.

Finally, this post gives some good context on the history of the revenue allocation process, including some great historic photos:

Meerkats, Mere Cats Or Orcats?

And where and what is this? It is in Seymour Street in central London; or rather it was. fbb has driven along Seymour Street c/o Google ...

publictransportexperience.blogspot.com

NEW POST BELOW - WRITTEN SEPARATELY, BUT AUTOMATICALLY MERGED BY THE FORUM SOFTWARE


This post is just to emphasise how rudimentary the understanding of passenger behaviour was at the birth of ORCATS.

The model used was simply to assume that passengers had an ideal departure (or arrival time) for a journey. They then 'evaluated' the timetable on the basis of the journey time, to which they added a fixed 'penalty' for each change required. They then chose the service 'nearest' to their ideal departure time. That was pretty much it. It sounds, and was, rudimentary.

We, of course, understood that there was an expensive list of shortcomings, but to estimate/calibrate the various possible parameters would have required data from far more significant changes in services than actually occurred in those early years of the work on passenger demand. Obviously the introduction of HSTs involved a significant improvement in speed, frequency and on-train experience. It was therefore difficult to separate the impact of those factors. I recall, for example studying closely a doubling of services stopping at Dewsbury, precisely because no change in rolling stock, fares or journey time took place at the same time and so it was a rare opportunity to study the impact or a 'pure' change in frequency in isolation.

I am sure, over the forty of so years since, the understanding of passenger demand has improved immeasurably.

Finally, just to mention that work had also been done on a 'cross sectional' approach to the question of passenger demand. The work described above was essentially time-series analysis; it looked at how demand changed in the weeks/months following a 'step change' in service. The alternative was a more fundamental approach of asking, based on the data on rail travel between every city in the country, how much demand would be expected between two towns/cities with populations P1 and P2, a distance D apart and with an average rail journey time of T1 as compared to a road journey time of T2. The results of that analysis had such enormous 'margins of error' as to be essentially useless for any 'real world'applications.

 

[Taunton]

As I mentioned previously, the question of regular allocation of revenues was always an objective for ORCATS, but the initial focus of this part of the OR group's work was evaluating the various investment decisions in the Inter City sector, and (also a key focus) justifying that investment to the Dept of Transport and Govt.

It would indeed appeal to Treasury officials, a bunch of highly numerate (though I'm not commenting on their other attributes) staff who would look approvingly at such calculated numbers.

An earlier approach was associated with electrification. It was well understood, and proven in reality, that if you introduced an electrified service it would get more passengers than the same service provided by diesels. This was known quite formally as "the Sparks Effect". Even the Southern Railway had identified it in the 1920s. I actually suspect that the "same service" was otherwise not quite true, in terms of all day frequency and speed, but nevertheless there it was. I was struck speaking to a larger Home Counties housebuilder around 1990 (when Network South East were doing quite a lot of both extension and infill electrification) that they watched for and took into account any 'railway electrification' in their pricing model calculations for housing developments in the ring 60 miles around London.

You find the same elsewhere. New tram services commonly attract more passengers and revenue than any comparable bus service doing the same thing.

 

[Chris Butler]

FINAL POST - WHICH WILL PROBABLY GET MERGED WITH THE ABOVE TWO


Not only was the data available to calibrate our model limited, but so were the computer facilities.

When I joined, the OR group (which was about 20 people, of whom about a quarter worked on passenger demand) worked in the offices between Kings Cross Platform 1 and York Road itself. As an aside those of us in a union could use the small canteen which involved walking right to the down end of platform 1 and then crossing the then still in-situ line that burrowed down from the end of platform 1 to the York Rd Curve.

For computing power, we used the TOPS machine which was located in Blanford House, a 'hidden' building very close to the BR HQ at 222 Maylebone Rd.. The locations and TOPS computing hardware is described here: http://www.les-smith.com/software/tops.htm. In those days you didn't carry the computer in your pocket, you went to it onstead. So to use the computer we got the Circle etc. Line from kings Cross to Baker St and walked to Marylebone.

At that time we used punched cards to input our programs to the computer (or rather, we left them in pigeonholes and the computer operators fed them to the computer when time was available) and then had to wait (typically several hours) for the results to arrive back on fan-fold paper, often a pile several inches thick. A single typo (meaning a mispunched hole) could costs you a whole working day !

For ORCATS, as I mentioned, the NPAAS (ticket) data was stored on tapes and so the computer operators had to load them too.

Quickly the facilities improved. First we got a printer installed at Kings Cross so that we didn't have to wait at Marylebome for the output and later we moved to offices on Tavistock Place (adjacent to the British Transport Police, whose canteen we shared) where, for the first time we had terminals which (via TSO) gave us direct access to the TOPS machine without the need to travel the Underground, cards and paper output in hand.

Amid all of that, ORCATS was born. At that time, it never occurred to me that, forty years on, it would still be going and more famous/infamous than anything else we worked on other than the APT.

 

[Dr Hoo]

Fantastic post from Chris Butler. Thanks.

My very first home location as a BR employee was in the old RCH building in Eversholt Street, alongside Euston Station, as Seymour Street had become by then. I have worked in some of the rooms illustrated. (The high level filing cupboards, accessed by balconies and spiral staircases were no longer in use on safety grounds.) I think that the last BR department to be based there was the Shipping and International Services Division/Sealink.

 

[Chris Butler]

An earlier approach was associated with electrification. It was well understood, and proven in reality, that if you introduced an electrified service it would get more passengers than the same service provided by diesels. This was known quite formally as "the Sparks Effect".

Indeed. The 'Sparks Effect' was well known. The equivalent, equally pronounced, effect for HSTs was dubbed 'The Nosecone Effect'."

 

[Taunton]

My own experience is a parallel to Chris Butler, possibly at the same time. University had a computer department and course, but ours had no contact - until we were offered an informal side project. There was a computer program, actually developed by the National Coal Board computer department, which modelled the demand for their employees to move on from one mine to another as they opened and closed, with parameters for distance, time, cost, etc, across a network of origins and destinations. You can probably see where this is going, and its applicability to transportation demand.

Likewise, it was on 80-column punched cards, with data done by operators (most of the staff seemed to be graduates from the course) from our handwritten sheets. Students, after tuition, were then (grudgingly) allowed to punch their own cards, for which two card punches were provided. One printed the punched text at the top, the older one didn't have this luxury and you had to scrutinise the little holes very carefully for any errors. This was the first time any of us had used a Qwerty keyboard. 24 hour turnaround.

ICL mainframe, and a DEC PDP/11 for the cognoscenti. The program came from NCB on about 2,000 cards, with instructions to be VERY careful not to drop them.

We played about with the parameters, not really knowing what we were doing at first, but developed our own rules. All a bit pioneering stuff. Interesting.

 

[Dr Hoo]

Indeed. The 'Sparks Effect' was well known. The equivalent, equally pronounced, effect for HSTs was dubbed 'The Nosecone Effect'."

Well... Up to a point. Whilst it isn't exactly surprising that a 'better' product (faster, cleaner, quieter, more frequent, 'new' rolling stock) will lead to greater demand there is very little real evidence that BR appreciated this in terms of HQ economic planning until after it had been demonstrated, firstly with the Kent Coast and then more significantly with the WCML after 1966.

If you read the 1955 Modernisation Plan the net annual financial improvement of around £35m on passenger services was almost entirely expected from reduced operational costs and some rationalisation of rural services (pre-Beeching and actually nothing to do with the Plan). Authorisation of electrification projects never seemed to depend on growth assumptions.

Even on the Southern Railway passenger revenue hardly increased between 1923 and 1938 (albeit that there were obviously some very powerful economic forces at work in this period).

I find it ironic that the RCH, which was the only body that might have had the numerical fluency and appreciation of passenger movement across the entire network to have taken on a new role as a business intelligence and forecasting unit, was unceremoniously trashed by the British Transport Commission. When Beeching came along in the early 1960s he had to spend the first couple of years just capturing some half-decent granular revenue and patronage statistics. At this point it seemed to come as a surprise to people that around one third of the network generated only a couple of percent of the revenue.

There is no doubt that BR's understanding of passenger revenue and operational research came on in leaps and bounds after that but it was from a pretty low base. Amazing that the shrewdness of some of the early railway management back in the 1840s had to be re-created over a century later.

And I loved both Chris Butler's and Taunton's reminiscences of 80-column punch cards. They exactly match my own recollections. I remember dropping my own 'deck' once. Another day wasted!

 

[Chris Butler]

Well... Up to a point.

To clarify, I meant that by the 1980's the 'received wisdom' was that the WCML electrification had seen significant demand increases and a significant 'sparks effect'.

 

[alistairlees]

Thank you to all for the very informative and interesting answers so far. I had never expected there to be so much to discuss here. If anyone has any more information about when ORCATS was first properly used for revenue allocation, I would be interested to hear.

 

[Dr Hoo]

Thank you to all for the very informative and interesting answers so far. I had never expected there to be so much to discuss here. If anyone has any more information about when ORCATS was first properly used for revenue allocation, I would be interested to hear.

To try and give a more specific response to this very reasonable question (rather than more antecedent history), ORCATS was the bedrock of passenger sector management in BR. To have a meaningful 'profit centre' you needed to have a reasonable idea of both costs and revenue of a group of services - the latter deriving from ORCATS.

The 'business led' organisation came into effect in 1986. Perhaps most famous for the launch of Network SouthEast in that year the principles were applied across InterCity and Provincial as well. So, within Provincial there were several 'sub-sectors', of which ScotRail was one. Within ScotRail there were six 'profit centres' (Express, East Coast Locals, South West, Highland, West Highland and Strathclyde). Within each profit centre there were 'service groups' (for example in Strathclyde there were Ayrshire, Inverclyde, North Electrics, Queen Street Local, etc.). Within each service group there might be 'service codes' (for example splitting Inverclyde into Glasgow-Gourock and Glasgow-Wemyss Bay). A high level of detail and disaggregation.

I joined ScotRail in that era and immediately became accustomed to working with management information derived from ORCATS.

Obviously on 'shared' routes ORCATS could even split flows between service codes. So between Larbert and Dunblane for example there were Strathclyde (Queen Street Local), East Coast Local and ScotRail Express trains and even the odd InterCity.

The concept of the 'ORCATS raid' quickly developed because if the spacing of trains around a clock face hour changed so did the revenue allocation. Quite a few apparent increases or decreases in revenue on a particular service code were explained by understanding how the timetable had altered.

By way of an anecdote, the long-established InterCity Euston-Inverness 'Clansman' ceased in May 1992, leaving a gap on the Highland Main Line. The ScotRail Express service had been re-vamped with Class 158s and more use of the Edinburgh-Ladybank-Perth route but we suddenly discovered that InterCity intended to run a residual Inverness-Edinburgh daytime round trip (1015 ex Inverness, 1440 ex Edinburgh) wholly within ScotRail territory. This didn't last long and came off at the end of September. It did not re-appear the following summer. This train was commonly called 'The ORCATIAN' in ScotRail commercial circles (in a nod to the pre-war 'Orcadian' 1025 Inverness-Thurso/Wick restaurant car service; a name briefly revived when the Class 156 Super Sprinters were introduced with the finest catering trolley that the Inveroykel Catering Company could provide).

 

[Chris Butler]

The 'business led' organisation came into effect in 1986.

The original sectors (InterCity, London & South East and Provincial) were created in 1982 weren't they ?

 

[Taunton]

Orcats Raids have always amused/irritated me (viewing from a distance). I can't quite see how such an obviously flawed technique could be allowed to continue without correction. More recently we even have the DfT themselves engineering Orcats raids to suit a particular franchise award of the moment. A particular example of this seems to be Trans Pennine Express extending from Newcastle to Edinburgh, triplicating the service at the north end of the ECML where few Kings Cross trains, the majority, run at more than 50% capacity. Amusing in that, since Virgin handed back the East Coast franchise, DfT now turn out to have effectively Orcats-raided themselves. Someone once wrote that Cross Country, rejigging their network with minimalist Voyagers, was "the biggest Orcats Raid ever".

 

[Chris Butler]

Thank you to all for the very informative and interesting answers so far. I had never expected there to be so much to discuss here. If anyone has any more information about when ORCATS was first properly used for revenue allocation, I would be interested to hear.

As I mentioned before, it depends what you mean by 'properly used' . Sectorisation started in 1982 and ORCATS (or, to be more precise, the tools and concepts underlying it) was used right back then to allocate revenues for specific management information purposes, such as investment decisions and service planning (at a strategic level). I'd call that use 'decision support'.

I don't recall it being used for routine/regular management reports (essentially profit centre accounting) at that time. I am sure it was being used for that by 1986. Between those two dates I would be pretty confident at guessing that it was used more and more for more routine reporting and less and less only for specific 'decision support'. I'm doubtful that there is any single implementation date.

Of course, it was not till privatisation that there was any need to actually allocate cash and determine the amount of payments between independent companies. In that sense, 'properly used' occurred with privatisation.

Correction: Prior to privatisation there was a need to allocate cash between independent entities, namely between BR and the PTEs etc.. With the Government backing both pockets, that allocation didn't have such sharp 'commercial teeth'. For all sorts of reasons, ORCATS wasn't really the tool for that so I'm not sure its all that relevant, but there was 'real' money involved there.

 

[Dr Hoo]

As I mentioned before, it depends what you mean by 'properly used' . Sectorisation started in 1982 and ORCATS (or, to be more precise, the tools and concepts underlying it) was used right back then to allocate revenues for specific management information purposes, such as investment decisions and service planning (at a strategic level). I'd call that use 'decision support'.

I don't recall it being used for routine/regular management reports (essentially profit centre accounting) at that time. I am sure it was being used for that by 1986. Between those two dates I would be pretty confident at guessing that it was used more and more for more routine reporting and less and less only for specific 'decision support'. I'm doubtful that there is any single implementation date.

Of course, it was not till privatisation that there was any need to actually allocate cash and determine the amount of payments between independent companies. In that sense, 'properly used' occurred with privatisation.

Correction: Prior to privatisation there was a need to allocate cash between independent entities, namely between BR and the PTEs etc.. With the Government backing both pockets, that allocation didn't have such sharp 'commercial teeth'. For all sorts of reasons, ORCATS wasn't really the tool for that so I'm not sure its all that relevant, but there was 'real' money involved there.

Thank you for that important correction!

In several years of sometimes fraught involvement with a PTE over 'Section 20' claims I never thought of the commercial teeth as being particularly blunt. Bear in mind that the PTE was effectively paying out the local authorities' Community Charge (aka 'Poll Tax') pot at that time.

So far as Taunton's point is concerned, the problem is that if you don't recognise that revenue can properly be 'shared' on overlapping or alternative routes you tend to end up with a lot of short-distance shuttles and many changes needed at boundaries or alternatively the 'one-through-train-per-day' model.

At the end of the day any tool can be misused. The trick is just to keep an eye on people who start taking the ****.

 

[Chris Butler]

In several years of sometimes fraught involvement with a PTE over 'Section 20' claims I never thought of the commercial teeth as being particularly blunt. Bear in mind that the PTE was effectively paying out the local authorities' Community Charge (aka 'Poll Tax') pot at that time.

I think we're talking about different eras. My comments were about the 1980's, pre privatisation (and so pre section 20) and even pre Poll Tax. In those days BR and the PTEs were separate, but ultimately Govt owned and that relationship was the one I was describing. Indeed, I even think that the PTEs didn't exist then and their role was covered by the Metropolitan County Council. Very different times.

 

[Dr Hoo]

1989

Transport Act 1968 still in force, including Section 20 (covering the concept of PTE subsidy/claim) +
ORCATS firing on all cylinders +
BR Sectors fully established +
Community Charge/Poll Tax applicable in Scotland +
Strathclyde PTE in absolutely no way regarding itself as part of the UK government(!) =

Serious discussions about serious amounts of real money

All long before privatisation

One day someone might even write a book about the modest part that ORCATS played in the decline of the Conservative party's fortunes north of the border.

 

[Chris Butler]

1989

Yes, we do have different times in mind. I had in mind 1983 (pre- Poll Tax, ORCATS not firing on all cylinders, sectors finding their feet) and the period around/before that when the PTE/PTAs had been abolished and not recreated. Perhaps that abolition didn't happen in Scotland, I can't recall. Anyway that was the period that my perception was that the PTE (actually, at that time, Metropolitan County Councils) relationship with BR had a different feel than in the later years. My perception is that around 1985 things changed a lot.

But you are right, all my other comments about PTEs would have applied after 1983, and before privatisation and the relationship by then had changed, as you say.