"TOPS" loco class classification - when did this actually start?

[AdamWW]

I didn't imagine it would be changed from what Southern Pacific had, who had locos eg 3000, or with an A or B suffix, and never renumbered anything when they devised TOPS.

And possibly had to cope with power from foreign railroads with an appropriate prefix before the number.

Another clue that not using 000 was just due to people liking to start numbering things from 1 is that we know TOPS can do locomotive numbers ending in 00, but subclasses tended to start at, e.g. 401 when they clearly could have started at 400.

(And I don't believe there was ever a D0)

 

[randyrippley]

I have now found my copy of the history of TOPS by Robert Arnott.

This confirms that Exeter was the first trial site.

It also clearly states that the first cutover for live use of TOPS was at St Blazey and Plymouth in August 1973. This should have included Exeter but that was delayed by 4 weeks because of some local unofficial industrial action. These match the dates that are on file at Kew.

There is a short report in the September 1973 issue of Railway Magazine, page 472.

Whatever was being used at Westbury in 1971 was not TOPS. It was probably a version of a Fortran program that had been developed at Derby and was first used to manage the Aire Valley MGR trains in 1970.

One little problem........at the time we were told it was TOPS by the people using it .

 

[D7666]

One little problem........at the time we were told it was TOPS by the people using it .

In the 52 years from 1972 to 2024 I never ever had cause to question it was anything else.

Most likely a trial version if not 'live' until mid 1973; live can have all sorts of different meanings, the engineers and developers could ave been using real data for Mendip stone long before handover to 'live' operators.

One does not suddenly go from nothing at 23:59 to full live by 00:01 on something like that back then. Years of testing before 'live'.

OK I do not lie awake at night worrying over such matters, but pretty sure with the rise of web discussion other debate would have questioned this before.

 

[Magdalia]

One does not suddenly go from nothing at 23:59 to full live by 00:01 on something like that back then. Years of testing before 'live'.

Roll out of TOPS was very quick. BR bought TOPS in 1971 and had gone live in the Bristol Division of the Western Region by the end of 1973.

As I said before the first trial site was Exeter in May 1972. That was when the testing started. Exeter was closely followed as trial sites by Dringhouses at York and Brewery Yard in Manchester. Further trial sites were Millerhill and Norwood to give full regional representation, plus Radyr as the Western Region were rolling out first.

From first trial site at Exeter to first cutover in Devon and Cornwall was about 15 months May 1972 to August 1973.

Westbury was not a trial site. It went live on TOPS in November 1973, along with the rest of the Bristol Division outside of Devon and Cornwall.

 

[Taunton]

The lowest number in the Southern Pacific loco fleet when TOPS went in there was ... 3. A small switcher. So even single digits did not upset the system.

 

[Big Jumby 74]

My extremely dog-eared copy 1974 edition of the Ian Allan Locoshed book (LMR dates 16-03-1974 and ER 23-03-1974) has only 45 001-012 and 101-121 with their new numbers. The rest of the 45 are shown with their old numbers and had to be filled in as they were dealt with out of order.

Likewise the 1975 edition still has many 45's with original numbers. One of my crowd back then would buy one of the railway journals each month (RM?) and from what I believe was the March 1976 edition, I have noted Peak 125 being renumbered 45071, possibly the last to be renumbered. The 1976 Loco shed book has all class members carrying TOPS numbers.

Edited: have just confirmed via DerbySulzers web site, Peak 125 was on Derby works for an extended period following collision damage repairs, and finally received the number '45071' during December 1975 before release to traffic, the last main line loco to receive its new number (of the classes that were to receive same).

 

[randyrippley]

Roll out of TOPS was very quick. BR bought TOPS in 1971 and had gone live in the Bristol Division of the Western Region by the end of 1973.

As I said before the first trial site was Exeter in May 1972. That was when the testing started. Exeter was closely followed as trial sites by Dringhouses at York and Brewery Yard in Manchester. Further trial sites were Millerhill and Norwood to give full regional representation, plus Radyr as the Western Region were rolling out first.

From first trial site at Exeter to first cutover in Devon and Cornwall was about 15 months May 1972 to August 1973.

Westbury was not a trial site. It went live on TOPS in November 1973, along with the rest of the Bristol Division outside of Devon and Cornwall.

It would appear that this is another of those occasions where history as written and recorded doesn't agree with history as experienced on the ground. You're reciting from records. I actually went there and I know what I was told by the people actually using it.

 

[MadMac]

There’s a summary of DMU TOPS codes at Railcar.

 

[The exile]

And also didn't have the concept of a class number followed by a serial number.

I suspect this is the nub of the “000” issue - human mentality rather than conputer capability. There is nothing wrong with having a 47000 if it is seen as a single number (as per post 1948 steam numbering) - if, however, it is “class number” followed by “serial number within that class” then logic pushes very strongly that the “first” of anything is number 1 (or 001!). The way the numbers were displayed on the vehicles themselves indicates that thought behind it. It was, of course, fortunate that even the mighty class 08 didn’t require a “1000”.

 

[Magdalia]

It would appear that this is another of those occasions where history as written and recorded doesn't agree with history as experienced on the ground. You're reciting from records. I actually went there and I know what I was told by the people actually using it.

The official sources are how it was experienced by the people actually delivering the TOPS project. My sources are the BR Board minutes and papers, its sub-committees, and the regional boards.

The Robert Arnott report, which I referred to upthread, was written by the TOPS project manager, who subsequently became the BR Board Chief Operations Manager. His detailed account of the TOPS project delivery runs to more than 100 pages, and was written in 1979, when everything was still fresh in the memory.

Together, they set out the TOPS purchase, testing and roll out in great detail.

These sources are much more reliable than the half a century old recollections of a schoolchild, based on one visit to one location. The official sources are contemporary and document what was happening when it was happening.

You are right that a lot of railway history as written and recorded is not what actually happened. But those are the histories that are based on the received wisdom of the enthusiast community, based on memories like yours, not original research from original sources. In my research of the original sources I find time and time again that the received wisdom of the enthusiast community is quite different from what actually happened. This is just another example to add to the list.

All this does not mean that recollections like yours are of no value to researchers. What you say does add to the evidence that computers were used to manage freight before TOPS.

 

[AdamWW]

I suspect this is the nub of the “000” issue - human mentality rather than conputer capability. There is nothing wrong with having a 47000 if it is seen as a single number (as per post 1948 steam numbering) - if, however, it is “class number” followed by “serial number within that class” then logic pushes very strongly that the “first” of anything is number 1 (or 001!). The way the numbers were displayed on the vehicles themselves indicates that thought behind it.

Indeed.

Hence subclasses starting at, e.g. 401, but then contiuing through 500.

 

[Gloster]

You are right that a lot of railway history as written and recorded is not what actually happened. But those are the histories that are based on the received wisdom of the enthusiast community, based on memories like yours, not original research from original sources. In my research of the original sources I find time and time again that the received wisdom of the enthusiast community is quite different from what actually happened. This is just another example to add to the list.

Indeed. The number of times where I have found that accepted truths are different to reality, even to the extent of it being stoutly maintained that the official records are wrong (*), is considerable. It should always be bone in mind that records can be wrong, but so can memory be wrong and one person’s memory has fewer checks on it.

* - Even to the extent of being categorically told that I was wrong about a signalbox closure date, even though I had worked it.

 

[Taunton]

What you say does add to the evidence that computers were used to manage freight before TOPS.

I believe a pioneer railway computer application, possibly the first "real" such, was in the accountancy invoicing of freight charges to customers, and collection of payment. Such computer programs were available commercially from a very early time, and were often an automated extension of what procedures had been done previously with card indexes, and typing onto multi-part carbonised forms. I think there was an article in one of the rail magazines in the 1950s about this, doubtless based on a BR press release.

 

[etr221]

I didn't imagine it would be changed from what Southern Pacific had, who had locos eg 3000, or with an A or B suffix, and never renumbered anything when they devised TOPS.

And possibly had to cope with power from foreign railroads with an appropriate prefix before the number.

The AAR (USA-Canada-Mexico) standard is loco numbers up to 4 digits, occaisionally there was/is a letter suffix, prefixed as required (in interchange service) with railroad/loco owner's 'reporting mark' (two to four letters). Computer systems are set up for this, with a four digit limit for loco numbers: when the Union Pacific (normal reporting mark UP) went over ten thousand locos - or rather, went over the limit where four digit numbers 'work' - they had to start using alternative reporting marks (e.g. I think UPY), of which the UP has a number (of its own, or inherited from absorbed roads).

And also didn't have the concept of a class number followed by a serial number.

I suspect this is the nub of the “000” issue - human mentality rather than conputer capability. There is nothing wrong with having a 47000 if it is seen as a single number (as per post 1948 steam numbering) - if, however, it is “class number” followed by “serial number within that class” then logic pushes very strongly that the “first” of anything is number 1 (or 001!). The way the numbers were displayed on the vehicles themselves indicates that thought behind it. It was, of course, fortunate that even the mighty class 08 didn’t require a “1000”.

I think there had long, going back to 'big four' days, been - for those responsible for such things, in particular perhaps on/from the LMS - a desire that loco numbers should start with a class identifier or indication: but a wholesale renumbering to implement this was not, and could never be, seen as justifiable, until computerisation (TOPS, etc.) came along and provided a reason to justify it.

But I have seen references to the BR Standard steam locos as class 70, etc., and if you look at LMS/BR loco numbers with that in mind, the outlines are visible.

 

[Clarence Yard]

I used to work for the man who helped draw up the class designation scheme.

The then CM&EE (BRB) wanted a system wide classification of locos so as not to cause confusion (such as with the ER who had been doing its own thing) and, with an eye to the future, renumbering the locos into a common format so that when computerisation came in for the CM&EE maintenance records, everything was now neat. It was a scheme devised for diesels by hp (with gaps for future classes) although the CM&EE wanted the small sulzers grouped by build so they went 24/25 & 26/27 rather than 24/26 & 25/27 (as 24-7 respectively), if you catch my drift.

This scheme was particularly important as the National Traction plan envisaged that locos would be easily transferred between regions so eventually everyone needed to work to one common nomenclature and standardisation of maintenance and records was the overall aim.

However, TOPS came in way before RAVERS (BR could only fund so much computerisation at once) and, iirc, after loco records were put on TOPS, loco maintenance by TOPS hours (rather than days worked) cut over in the late 1970’s but it wasn’t until the late 1980’s that RAVERS took hold at depot level, firstly working through a non networked version called LVRS. At depot level you would enter the work done on any particular loco and that not only gave you a record, jobs were timed so it was used to calculate the men’s bonus, do the depot productivity analysis, as well as produce costings.

Just to give you an idea, in sector days each loco was allocated to a sub sector. Inputting a loco number automatically aligned it to a job number which then was debited to that sub sector. So the workshop staff time (on an equated hours basis) plus overheads would be allocated to that sub sector, any stores docket (through the IMACS system) would allocate the materials to that loco and then to that sub sector and anyone with the right permissions from the BRB chairman right down to the depot organisation itself could see the detailed periodic sub sector costs (by type of expenditure) through the AXIS II financial system.

At loco depot level, the first computer that ever came on the depot was usually the TOPS terminal, nearly always a small Ventech machine which was typewriter sized and just had a keyboard and printer. The print paper came in cylindrical rolls. The handy thing about it was that you could do away with having to use a separate telex machine because of it’s ability to send freeform typed “ZZ” messages to your intended recipient, as long as they also had a TOPS terminal!

 

[AdamWW]

Indeed. The number of times where I have found that accepted truths are different to reality

I have read and heard many times that Beeching didn't stop to think about the loss of main line traffic fed by closed branch lines.

But unless my reproduction copy of the report is somehow a fake, he did indeed consider it and dismissed it.

I think there had long, going back to 'big four' days, been - for those responsible for such things, in particular perhaps on/from the LMS - a desire that loco numbers should start with a class identifier or indication: but a wholesale renumbering to implement this was not, and could never be, seen as justifiable, until computerisation (TOPS, etc.) came along and provided a reason to justify it.

BR did have two chances pre TOPS when they could have done something a bit more systematic with diesel and electric numbering, yet seem to have chosen not to do so.

You can argue that the 1948 numbering scheme was limited by most of the number range being given up to existing steam locomotives.

But the "D" and "E" numbers could presumably just as easily have included class numbers had there been a desire to do so.

 

[D7666]

The official sources are how it was experienced by the people actually delivering the TOPS project. My sources are the BR Board minutes and papers, its sub-committees, and the regional boards.

The Robert Arnott report, which I referred to upthread, was written by the TOPS project manager, who subsequently became the BR Board Chief Operations Manager. His detailed account of the TOPS project delivery runs to more than 100 pages, and was written in 1979, when everything was still fresh in the memory.

Together, they set out the TOPS purchase, testing and roll out in great detail.

These sources are much more reliable than the half a century old recollections of a schoolchild, based on one visit to one location. The official sources are contemporary and document what was happening when it was happening.

You are right that a lot of railway history as written and recorded is not what actually happened. But those are the histories that are based on the received wisdom of the enthusiast community, based on memories like yours, not original research from original sources. In my research of the original sources I find time and time again that the received wisdom of the enthusiast community is quite different from what actually happened. This is just another example to add to the list.

All this does not mean that recollections like yours are of no value to researchers. What you say does add to the evidence that computers were used to manage freight before TOPS.

1979 ? Only EIGHT years after 1971 then. Memory can lapse in 8 years.


2 memories. Two of us recall this. Our dates align. Late 1971 / early 1972 are ingrained in my memory as the D800 / D7000 rundowns ended / started, and for many other reasons.


It has to be based on memory. We did not have any forum to record this stuff. Not just no web. There was no forum at all. The likes of RCTS and LCGB would simply have ignored this stuff (just as RCTS later ignored the entire sector pools for years) while rivet and spoke counting, and the commercial mags only print articles on what they receive officially and the like of GFA would more than likely ignored me if I had sent that in.


This is not "received wisdom". If this was something regurgitated over and over again over the past 50 yeas then yes. But it is not. It is a very specific recall from 2 different people of a very specific time and event that does not appear to have been recorded anywhere, and not connected with anything else you will find. To quote what I got in a private message "we were there", and, with respect, you were not.


Whatever one selectively chooses to believe, BEFORE Exeter went live, there will have been an extensive trial period, and before trials a test period, and before testing a commissioning period. And all that is an even longer period for a writer doing so in 1979 to have recalled. Even today you'd be talking a 2-3 years lead time, back then more like 4-5 year. Believe me. This is my day job (on modern systems); currently working n a system that "went live" 2 months off 6 years ago, and is a commissioned asset, but we are still not yet out of it all, we still testing things, and already now I can't remember or what event occurred when never mind 6 y ago and the 2 y of pre-commissioing work before that; yes I have a diary and notes, but you can't note everything, and the trivial point at the time that went unrecorded can be the one that unlocks something later.


Also, the alleged history must be flawed. It makes no sense to have just one TOPS location. What was the purpose of that ? Wagon control is all about recording what wagons went where in which trains. So what use was Exeter to anybody for outgoing traffic as there was no destination to receive wagon data, and no location to send Exeter data of incoming traffic. There must have been somewhere else or Exeter were typing numbers into the ether which tests or trials or proves nothing.

Lastly. One should always challenge anything and everything. You question us, with our specific memories of one item, yet appear to accept without question one single source. IME, a single source should always be questioned. Even with a "standard", one working ethos is "challenge the standard".

 

[Magdalia]

It has to be based on memory. We did not have any forum to record this stuff. Not just no web.

No it doesn't. The British Railways Board records are not based on memory. They recorded it, in the minutes and papers of the Board and its sub-committees, the project progress reports, and the separate regional boards. All of that was recorded as it happened. The minutes and papers of the British Railways committees are a much better record than anything on this forum or the web. That's how records were kept in those days, before the internet existed.

1979 ? Only EIGHT years after 1971 then. Memory can lapse in 8 years.

That was the lifetime of the project. Arnott clearly consulted his colleagues when preparing his report, all of their memories don't lapse.

2 memories. Two of us recall this. Our dates align. Late 1971 / early 1972 are ingrained in my memory

Childhood memories from half a century ago. You are demonstrating that it very ingrained, because you are unable to reassess it despite the overwhelming evidence to the contrary. Are you seriously suggesting that childhood memory is more reliable than the extensive written record of the TOPS project manager, having consulted his colleagues, or the reports prepared for presentation to the BR Board?

Lastly. One should always challenge anything and everything. You question us, with our specific memories of one item, yet appear to accept without question one single source.

Absolutely. I suggest that you try it with your own childhood memory, challenging it against what is recorded about TOPS in the extensive and contemporary official records.

It is what I have been doing here. My inference, on the basis of all the evidence, is that Westbury had a computer system for the stone trains in 1971/72, but it wasn't TOPS, the most likely explanation being that they were using a linear programming tool like that used on the Aire Valley MGRs.

 

[eldomtom2]

I think there had long, going back to 'big four' days, been - for those responsible for such things, in particular perhaps on/from the LMS - a desire that loco numbers should start with a class identifier or indication: but a wholesale renumbering to implement this was not, and could never be, seen as justifiable, until computerisation (TOPS, etc.) came along and provided a reason to justify it.

But I have seen references to the BR Standard steam locos as class 70, etc., and if you look at LMS/BR loco numbers with that in mind, the outlines are visible.

It was the GWR that used loco numbers for class identification, not the LMS. If you look at BR-era documents they're consistent with class identification, inherited from the Big Four - letter/number combinations for LNER and SR classes, numbers for GWR classes, power classification and wheel arrangement for LMS classes (with original company in parentheses if differentiation was required). The LMS style was adopted for the steam standards - I've never seen the Britannias referred to in official documents as "class 70" or a similar designation, it's always "Class 7 4-6-2", with "7" here referring to the power classification, not the number.

 

[Taunton]

The minutes and papers of the British Railways committees are a much better record than anything on this forum or the web. That's how records were kept in those days, before the internet existed.

So one wonders how their TOPS implementation handled it when they came across the entries in the formal records for the four Wirral electric cars that were in the rolling stock registers right to the mid-1980s, or whenever these "better records than anything" were transferred to TOPS - the four cars which, almost new, had been destroyed by bombing at Birkenhead in 1941, had replacements built in the 1950s, but which were never taken out of the records, year after year, in some accountancy depreciation fudge.

The four cars that, year after year, turned up in the Ian Allan spotters' books (and which novice spotters solemnly ticked off) right up to that time. Because the publisher just believed those official documents they were supplied with.

I do know that when the Southern Pacific made the transition to TOPS there were a couple of hundred freight cars that were never found - plus a few that they never knew they had, or had been written off long beforehand. The first of these arising led to a serious meeting at the Vice-President level with long faces. After a while it was just laughed at in the bar after work as "I wonder what else we will find".

 

[AdamWW]

So one wonders how their TOPS implementation handled it when they came across the entries in the formal records for the four Wirral electric cars that were in the rolling stock registers right to the mid-1980s, or whenever these "better records than anything" were transferred to TOPS - the four cars which, almost new, had been destroyed by bombing at Birkenhead in 1941, had replacements built in the 1950s, but which were never taken out of the records, year after year, in some accountancy depreciation fudge.

The four cars that, year after year, turned up in the Ian Allan spotters' books (and which novice spotters solemnly ticked off) right up to that time. Because the publisher just believed those official documents they were supplied with.

Seems a bit different to me to having minutes of meetings recording events having taken place.

i know that if I went through records of projects I've managed there would be missing details that were so obvious they didn't end up being written down, and inconsistencies over how many of such and such parts were made - similar to having rolling stock in the records but not existing.

But I would have a very good idea of when the project started and when it hit key milestones.

 

[43096]

1979 ? Only EIGHT years after 1971 then. Memory can lapse in 8 years.


2 memories. Two of us recall this. Our dates align. Late 1971 / early 1972 are ingrained in my memory as the D800 / D7000 rundowns ended / started, and for many other reasons.


It has to be based on memory. We did not have any forum to record this stuff. Not just no web. There was no forum at all. The likes of RCTS and LCGB would simply have ignored this stuff (just as RCTS later ignored the entire sector pools for years) while rivet and spoke counting, and the commercial mags only print articles on what they receive officially and the like of GFA would more than likely ignored me if I had sent that in.


This is not "received wisdom". If this was something regurgitated over and over again over the past 50 yeas then yes. But it is not. It is a very specific recall from 2 different people of a very specific time and event that does not appear to have been recorded anywhere, and not connected with anything else you will find. To quote what I got in a private message "we were there", and, with respect, you were not.

So on the one hand you reject as a source someone who was involved and documented it at the end of the project (8 years) and instead rely on the memories of two people who were not directly involved 50 years later. I'm sure you can see why that approach lacks credibility. Given that the records of those involved tally with contemporary records taken from the archives, the anecdotal sources you and others provide cannot be relied upon. As you say memory can lapse over 8 years. So what happens over 50?

 

[Adrian Barr]

Also, the alleged history must be flawed. It makes no sense to have just one TOPS location. What was the purpose of that ? Wagon control is all about recording what wagons went where in which trains. So what use was Exeter to anybody for outgoing traffic as there was no destination to receive wagon data, and no location to send Exeter data of incoming traffic. There must have been somewhere else or Exeter were typing numbers into the ether which tests or trials or proves nothing.

TOPS started out as a punched card based system for the people using it in yards. The quotes below are from the Robert Arnott history mentioned upthread:

IBM 1050 I/O terminal
This was a card oriented device with printout facilities, situated at depots and freight yards and was the main device for inputting messages to update the computer. The 1050 was an electromechanical machine consisting of an 80 column card reader and punch, printer and keyboard...

...Each card can have up to 80 characters of data punched in it, and the card readers can each read up to 1000 cards a minute

Note there is no mention of a computer screen, and from conversations with "old school" TOPS clerks in the past, the original setup did not involve outputs to a screen at all. This is a rough and very basic second-hand understanding of how the system worked initially, as best I can understand it:

Information would be printed out on paper (e.g. the formation of incoming trains to the yard) but there would also be an output of punched cards, each representing a loco or wagon. These were about the size of playing cards, and would be physically placed in racks with slots labelled to represent different local freight terminals, or specific sidings in a yard. The shunting of a train in a yard would literally involve "shuffling" these cards into a different order to keep track of a wagon's location in the yard and the sequence of wagons in a particular siding.

For a departing train, a loco card and the wagon cards in the correct sequence would be fed into the card reader and the information transmitted onward to the mainframe and then forward to the TOPS Office for the destination or next calling point where the train would be remarshalled. At this location, another set of cards would be printed out and the process would start again. This means that information about the precise formation of wagons on different sidings in a yard would, at least initially, have been based on the order of physical cards in a rack. It's interesting that there is an element of crossover in this system between "files" in a computer as we would understand it now, and data "filed" in the sense of cards physically filed in a specific order ready to be fed through a card reader.

Presumably certain events (e.g. when wagons were placed in a terminal for unloading) would require feeding the relevant cards into the reader, in combination with a keyboard input that specified the action being done, to transmit information to the TOPS Mainframe. It's hard to imagine using this system, and I don't understand some of the details such as how the inputs worked without a conventional VDU being involved, but punched cards and printers featured heavily.

During the trial phase (and later during the gradual rollout of TOPS nationwide) there was a requirement to inferface with locations that had not been connected to TOPS. It appears that the immediate pre-TOPS method of communicating the formations of freight trains was via Telex (ATI - Advance Train Information):

The trial sites contained the machines, the files, the cards...

At trial sites the cards were produced in one of two ways. If there was an ATI message for an approaching train a paper tape was also cut on a special device provided, in addition to the hard copy on the telex, and the paper tape was fed through a card punch to produce the cards. In the absence of an ATI message the clerks had to punch out the cards from a numbertaker's report.

The cards were filed as in TOPS, the yard file mirroring the disposition of wagons throughout the sorting sidings, the location file those through the TRA. Local reportings from outside staff were in accordance with TOPS procedures, i.e. wagons placed at customer's premises, loaded, unloaded, released back to traffic, crippled, held for specific reasons.

From reading the history, it appears that a significant part of the trial phase was the testing of these machines and the processes involved. A system of this type could potentially be to some extent independent of a central computer, and in theory could have been used merely to transit information from yard to yard and keep track of wagons locally (of course this would have defeated the main purpose and benefit of TOPS; to have centralised nationwide control of wagons).

The trial sites were limited (it was originally only intended to have three) and spread regionally - Exeter, Dringhouses, Brewery Yard (Manchester), Radyr and Millerhill. There would have been very limited interface between them, although the interface between two TRAs (TOPS Responsibility Areas) was tested on the Southern region:

The Southern to their great credit produced their own trial site at Norwood Yard by doubling it up as a TOPS training centre for the whole region, simulating two separate TRAs on two machines in the future AFC, thereby being able to practice inter-TRA type messages as well as local procedures.

(AFC = Area Freight Centre)

"Local procedures" sums up the majority of what a trial TOPS site would have been involved with.

The history mentions that by 1975, "Ventek" terminals with VDUs had been introduced to replace the earliest and most primitive machines, but as I understand it, the punch cards were still an ongoing feature of TOPS for some time. From speaking to people who had worked using the cards, the idea of a TOPS system which did not use punch cards ("cardless TOPS") was initially treated with the same kind of scepticism we would understand from the term "paperless office."

The TOPS system was upgraded over time, so that a slot full of cards representing wagons in a particular siding gave way to "user sets" stored as data on the mainframe, but there is still some hangover from the card era. For example, a message from TOPS that a wagon consist for an inbound train has been changed will still advise the TOPS clerk to "discard consist and request cards."

This 1978 film is of interest: https://player.bfi.org.uk/free/film/watch-using-tops-1978-online

Some other background history:
Telecommunications and computing before the internet: British Railways’ nationwide train operating system (PDF)

 

[Magdalia]

BEFORE Exeter went live, there will have been an extensive trial period, and before trials a test period, and before testing a commissioning period. And all that is an even longer period for a writer doing so in 1979 to have recalled. Even today you'd be talking a 2-3 years lead time, back then more like 4-5 year.

TOPS was an existing and proven software product. All of those trials, tests and commissioning had already been done in the USA. BR acquired existing software, instead of developing its own, in order to cut all of that out and deliver speedy implementation.

It makes no sense to have just one TOPS location. What was the purpose of that ? Wagon control is all about recording what wagons went where in which trains. So what use was Exeter to anybody for outgoing traffic as there was no destination to receive wagon data, and no location to send Exeter data of incoming traffic. There must have been somewhere else or Exeter were typing numbers into the ether which tests or trials or proves nothing.

TOPS's ability to match outgoing and incoming traffic at different locations was already proven, it did not need to be tested again. The key purposes of the trial sites were to test data capture and the telecommunications system to and from the centre. Someone in Exeter typing in numbers that can be read at the centre proves that data capture and telecommunications are working. Trial sites spread across the regions demonstrated that data capture would work under different local operating arrangements.

 

[Taunton]

Southern Pacific was headquartered in San Francisco, and it is no coincidence that Silicon Valley and all its pioneering computer organisations, and much of the TOPS development, was right there, and done by those companies and individuals strung along the old SP commuter line from there up "The Valley" to San Jose.

I'm sure I've seen old BR documentation with the TOPS logo on it, which had the same font with large capitals as the old SP logo.

Here's a start on detailing what they did in those early days, with some interesting commentary on the TOPS sale to BR. The "King of Wisconsin" reference is to Ed Burkhardt, head of EW&S here. The further link in there takes you on to a 500 page manual of the system, an interesting blend of old 1960s computer geekery and USA railroad terminology:

Southern Pacific Lines Coast Line Division “The Route of the Octopusâ€

 

[Magdalia]

Some other background history:
Telecommunications and computing before the internet: British Railways’ nationwide train operating system (PDF)

Thanks very much for this. I'd not seen it before.

It confirms some of the key dates.

British Railways Board approved the acquisition of TOPS in June 1971 and Government approval was obtained in October.

TOPS was implemented between August 1973 and October 1975. The first pilot “cutover” exercise took place in the Plymouth and St Blazey area

And I thought this particularly interesting, which will have facilitated a speedy roll out.

In principle, British Rail was bound by contract to slavishly copy all the American software and equipment.

 

[Adrian Barr]

I have heard many times that the "TOPS" class numbers were necessary as TOPS couldn't cope with the previous numbering system.

From previous discussions on this, the true situation appears to be that the class numbers were thought up before TOPS, but TOPS introduction gave a justification for renumbering locomotives to include the new class numbers.

Of course some locomotives ran into the 70's without TOPS numbers and it seems that it happily coped with this.

Various statements are made about what TOPS can and cannot cope with (such as numbering starts at 001 because it couldn't handle 000) but I think it turns out that the locomotive number is just effectively a free text field and anything can go in it, and numbers start at 001 because non mathematicians generally don't like starting to count things from zero.

It's interesting that TOPS was being used for locos with pre-TOPS numbers, such as Westerns. A loco enquiry in TOPS seems to be happy with any number of digits up to five. For example, if you try to do an enquiry on a loco number such as 3 or 23 it will tell you there is no record of the engine's existence, but if you enter a six-digit number it returns ?????????? indicating it is not programmed to accept that number of digits in a loco number.

I'm pretty sure that TOPS stores locomotive numbers in a different database (the "locomotive file"?) to coaching stock and wagons.

Yes, loco enquiries (E3) are different to vehicle enquiries (J6) which can include wagons, coaches and multiple unit vehicles. It's the same when entering consist information, on freight trains the loco number goes on an "06 line" while other vehicles go on "02 lines" (with 06 or 02 being a line identifier which tells TOPS what kind of input you are making). For example, if multiple unit stock is being delivered in a freight train, the wagons and EMU vehicle numbers can all be entered using "02 lines."

***************************

Notes on wagon numbering:

With wagon numbers in TOPS, an owner prefix is considered part of the number. For example VTG 57231 is a TTA wagon, while 57231 is part of Class 150/2 unit 150231

It appears that TOPS didn't result in much renumbering of the BR wagon fleet; for example the "B series" 6-digit numbers for vacuum braked wagons were in place well before TOPS. One minor change (perhaps not directly related to TOPS but occurring about the same time) was the introduction of another 6-digit number series, without any prefix, for wagons built with air brakes. The MGR wagons originally had a B prefix, with the first one being B350000, but the B was later dropped to put them in the new air-braked number series.

A bigger change was the introduction of the 3-letter TOPS CARKNDs, such as HAA, to distinguish wagon types. Before becoming known as HAAs, the wagon type was displayed as "HOP 32 AB" on the side of the wagons (32 ton capacity hopper, air braked).

Pre-nationalisation wagons had their regional prefixes (S, M, W, E) to prevent them conflicting with other number series. The biggest numbering change for wagons was in the private owner fleet. Previously, it appears the wagon owners would use their own numbering schemes. In a David Larkin book on "non-pool freight stock" it's surprising to see a photo of a modern-looking BP tank wagon (later TTA) in 1968 numbered 32. The specific wagon was owned by STS and on hire to BP Chemicals, and referred to in the text as "STS 32" to distinguish it from any number of other private wagons that might also have been numbered 32. The idea of an owner prefix was implemented in TOPS numbering for private owner wagons, and they were also renumbered with a four or five digit number (such as the example of VTG 57231 earlier).

The HTV and MCO wagons were marshalled into sets with the 4 digit set number stencilled on the side of each wagon. The MCO sets ran with a stencilled brake van at each end; gradually the vans gave way to fitted heads, about 3-4 MCV displaced the same numer of MCO at each end as well as the van. I am pretty sure those wagon set numbers were computer data related as they had no 8 or 9 in the number range from what I saw which suggests they were octal numbers.

Those were TOPS pool numbers, with some listed here: https://www.rmweb.co.uk/forums/topic/162893-tops-wagon-pool-identities/
In some cases a pool will effectively constitute a set, with a few maintenance spares.

This is a good example of an MSV displaying a prominent pool number for stone traffic: https://paulbartlett.zenfolio.com/ironoretipplermsv/h2163d027#h2163d027
And here is an HTV displaying pool 7509: https://paulbartlett.zenfolio.com/brrebuilthtohtv

I looked for photos of these wagon types in Mendip stone traffic taken around 1972 and managed to find a couple.

Here at Merehead in May 1972, HTVs display their pre-TOPS type code of "HOP 24 VB" along with weight information and their B series numbers: http://www.hatspics.co.uk/photo_expanded.php?id=7205009
There are some interesting 1970s photos on that website, with an index by year here: http://www.hatspics.co.uk/type_years.php?type=British Railways

Here in a rather nice 1972 sequence of photos at Merstham (with a warship unloading HTVs at a temporary terminal for M25 construction) the wagons are again in a pre-TOPS state displaying "HOP 24 VB," with "HOUSE COAL CONCENTRATION" revealing their former occupation: https://www.flickr.com/photos/26292271@N05/31866811214/in/photostream/

In a "Foster Yeoman: The Rail Story" booklet (probably from an open day at Merehead) there's a Feb 1972 photo of D1662 with MCVs at Merehead, which again are in their pre-TOPS state with lettering displaying STONE, no pool code, with weight info and the B series number.

This was all taking place before I was born, but I can't help thinking the Mendip stone traffic must have been a great sight behind a Western or Warship!
821 at Westbury with stone empties: https://www.flickr.com/photos/52554553@N06/10421811905/
D1005 with stone empties: https://www.flickr.com/photos/52554553@N06/10421711036/

 

[norbitonflyer]

Yes, loco enquiries (E3) are different to vehicle enquiries (J6) which can include wagons, coaches and multiple unit vehicles. It's the same when entering consist information, on freight trains the loco number goes on an "06 line" while other vehicles go on "02 lines" (with 06 or 02 being a line identifier which tells TOPS what kind of input you are making). For example, if multiple unit stock is being delivered in a freight train, the wagons and EMU vehicle numbers can all be entered using "02 lines."

Interesting. That explains why EMU unit numbers can duplicate individual carriage numbers (such as the 4505xx carriages in Class 720)

Although back in the 1980s a large number of LHCS and DMU cars were renumbered to avoid duplicating classes 25, 26, 50, 56, 60, 81, 85 and 86 (and some EMUs which were numbered in the Southern Railway series, despite having been built, in some cases, as recently as 1959 - indeed, some received new SR numbers as late as 1983, when converted from SUB to EPB specification)
Some of this renumbering seemed over the top - was it really necessary to renumber nearly 1,500 dmu vehicles when only 185 of them duplicated locomotives?

 

[AdamWW]

Interesting. That explains why EMU unit numbers can duplicate individual carriage numbers (such as the 4505xx carriages in Class 720)

Although back in the 1980s a large number of LHCS and DMU cars were renumbered to avoid duplicating classes 25, 26, 50, 56, 60, 81, 85 and 86 (and some EMUs which were numbered in the Southern Railway series, despite having been built, in some cases, as recently as 1959 - indeed, some received new SR numbers as late as 1983, when converted from SUB to EPB specification)
Some of this renumbering seemed over the top - was it really necessary to renumber nearly 1,500 dmu vehicles when only 185 of them duplicated locomotives?

I believe the rule was that anything built to pre-nationalisation designs got a pre-nationalisation number.

Interesting that whatever the reason is for not wanting coach numbers to clash with loco numbers, it doesn't seem to extend to unit numbers.

 

[Pigeon]

I don't actually understand why 000 would automatically require extra code, though you might choose to reserve it to describe a null entry or some such.

You've not programmed with strings in C, and then at some later date found yourself having to make the code work with strings that have nulls in?

It's an inevitable concomitant to the decision of whether or not to hijack a particular "unlikely" in-band value to indicate an out-of-band meaning. If you're dealing with data that doesn't have any "unlikely" in-band values, then you have to use a separate flag to represent the out-of-band meaning, and include code to fetch and test that flag in addition to the code handling the data itself.

In addition, if you can choose 0 as the value you hijack, you gain the advantage that it's extremely common for CPUs to be able to test whether some value is 0 or not-0 directly, without having to do a separate "compare" instruction. You could, for instance, imagine the TOPS bods deciding that 999 would be the most "unlikely" value, but then you have to do a "CMP x, 999" instruction (bit of pseudo-assembly there) before you branch, whereas with 0 you don't. And in those days, one extra instruction was significant. (Come to that, it still is in some cases.)

Having said all that, the implication from other posters' descriptions implies that TOPS probably was not hijacking in-band values anyway, but there you go.