First Loco (re)Numbering Using TOPS System ?

[pete_mcfarlane]

So why was there never 55000 or 56000 etc etc?

Possibly because locomotives have always been numbered starting from 1 rather than 0. So 47001 is seen as class 47 number 1, rather than locomotive number fourty seven thousand and one. Following this logic, having a locomotive number 0 wouldn't make sense.

[asmay2002]

having a locomotive number 0 wouldn't make sense.

 

 

Why not? They did it in South Australia. (2-6-2T No 0.) And of course many, many British steam loco classes started with the first engine number ending in 0.

 

I suspect the cuplrit is merely lazy programming on the part of whoever wrote the software.

[Downendian]

Aah yes but there is the rumour that BR only ordered 100 Hymeks, and schoolboy arithmetical error of stating they wanted them numbered D7000-D7100 ended up with a class of 101.

Caveat emptor etc.

Neil

[Pennine MC]

It's a nice theory Neil, but OTOH, 1960s diesel builds were often ordered in batches of (for instance, figures out of the air) 44 for Area A plus 23 for Area B, which some chappie had evidently calculated would be the figure required to dieselise the diagrams for a given area or service. When you think about it, it's mostly just the non-adopted Pilot Scheme classes that stayed with batches of 10 or 20 - everything else is an apparently random number (263 Brush 2s, 74 Westerns); it's just that 101 has that 'ring' about it. And of course had things been different, there'd have been some 300 Hymeks

[jamie92208]

I have vague memories of being told tnhat TOPS couldn't cope with a loco number of 0. Presumably the loco number is actually stored in the memory as two separate items, the class and the individual number. Therefore 55000 would be two linked records. Therefore the progmatic decisiton was made to make the '0' loco in each class into the next available number. I seem to remember that 6700 became something like 37123 and took the number of a withdrawn sister. As to when they were applied, I have memories of typing out lists of pre and post TOPS numbers when I was at Poly in Huddersfield in 1972/73 so they were definitely around then but I have no idea as to how many had been done by then.

 

Jamie

[caradoc]

First I remember seeing about re-numbering locos was in the Jan 1972 Railway Magazine, which implies that 76050 had already been re-numbered at Crewe Works.

 

As for why there was no 47000 etc, that's not now we humans number things ! Any list of objects starts with 1; eg 10 objects are numbers 1 to 10, not 0 to 9.

[frobisher]

I suspect that TOPS was fully capable of coping with 000 numbers (in fact it did with coaching stock and wagons). From a programming stand point, what could have motivated the "no zero locos" thing could have been the fact that you would be masking out the numbers to search for a particular class through modular arithmatic.

 

For instance, to look for all Class 55 locos, you'd modular divide the TOPS number by 55000 a result of anything but 1 meant that it wasn't a Class 55, and the modular remainder of the calculation (1-22) would give you the loco number. It may be that the system used the remainder as the sucess/fail condition so a zero numbered loco would have been missed consistently.

[Peter Kazmierczak]

Actually, come to think of it, there was a loco ending in 000.

 

When the prototype HST power cars were renumbered in the coaching stock series they became 43000 and 43001. But later HST power cars became redesignated as locos again. But by then the two prototypes were in the departmental series, so maybe I've demolished my own argument.

 

On the original HSTs the number was applied as 41 001 (with a space between the 1 and 0) which possibly implies two fields; one for the class and the other for the specific vehicle.

[Natalie]

Actually, come to think of it, there was a loco ending in 000.

 

When the prototype HST power cars were renumbered in the coaching stock series they became 43000 and 43001. But later HST power cars became redesignated as locos again. But by then the two prototypes were in the departmental series, so maybe I've demolished my own argument.

 

On the original HSTs the number was applied as 41 001 (with a space between the 1 and 0) which possibly implies two fields; one for the class and the other for the specific vehicle.

 

I don't think coaching stock and multiple units were included on TOPS until after 1983 (?) by which time the two prototypes were in departmental use with appropriate numbers.

[Natalie]

Interesting Mike.

 

So why was there never 55000 or 56000 etc etc?

 

And if renumbering introduces errors, why renumber?

 

I think the 'no 000' also applied to sub classes too- the first 47/3 was 47301, 47/4- 47401, 47/7- 47701 etc but it was deemed ok to have a number ending in '00' if in the middle of a subclass- such as 31200, 31300, 47200, 47500 for example.

 

One exception to this was 31400 which on rebuild/conversion from 31161 was not allocated the next available number (31401 already existed) which would have 31470. Maybe it was electrically more akin to the original 31/4 ETH conversions?

[Downendian]

And of course had things been different, there'd have been some 300 Hymeks

Or 301 !

Neil

[drjcontroller]

Getting back to the original question, I saw & photographed 45104 at Derby Works on 16/09/72, just out of shops. Later the same day, I saw 83012 & 84008 at Bescot.

 

Douglas

[Pennine MC]

As to when they were applied, I have memories of typing out lists of pre and post TOPS numbers when I was at Poly in Huddersfield in 1972/73 so they were definitely around then but I have no idea as to how many had been done by then.

Getting back to the original question, I saw & photographed 45104 at Derby Works on 16/09/72, just out of shops.

 

Just to clarify before the thread takes the usual route of muddlement or repeating incorrect information - the point has already been established that diesel renumbering, with the class 45 Peaks, started in 1973. My clear recollection is that with 45s, it started that spring, in advance of the rest of the diesel fleet, and is confirmed by Roger Harris' allocation history

[leopardml2341]

.......... However many places took this to mean that they should also remove existing D prefixes. This resulted in many locos just having them painted out........

 

I distinctly remember many locos having the whole number over painted and having to decipher the 'real' number by looking at an oblique angle along the bodyside............ before Ian Allan caught up !

[ForestPines]

I suspect that TOPS was fully capable of coping with 000 numbers (in fact it did with coaching stock and wagons). From a programming stand point, what could have motivated the "no zero locos" thing could have been the fact that you would be masking out the numbers to search for a particular class through modular arithmatic.

 

For instance, to look for all Class 55 locos, you'd modular divide the TOPS number by 55000 a result of anything but 1 meant that it wasn't a Class 55, and the modular remainder of the calculation (1-22) would give you the loco number. It may be that the system used the remainder as the sucess/fail condition so a zero numbered loco would have been missed consistently.

 

I doubt that that is how it worked internally.

 

According to the information on the internet, TOPS ran originally on the IBM System/360 computer architecture and was written in assembly language macros. Based on that, I would assume that it stored loco numbers in binary-coded decimal and accessed the subfield for class or number by using a bitmask and bitwise boolean instructions to select the required nybbles - a System/360 machine with the right hardware options could do those operations very quickly, as the BCD architecture was always common in mainframe computing (it still is today in mainframes, but didn't make the transition to PCs). Division would be extremely slow and unnatural by comparison.

 

If you want a potential software-based reason, a more likely idea would be something like: class records being stored in blocks in a file, with the first record (record number 0 - most computers prefer to start counting that way) holding class-general data and the subsequent records holding loco data. That way, the three "loco number" digits would be used as an index into the record block. However, this architecture wouldn't work very well if you had missing numbers in the sequence (unless the System/360 supported "sparse files", which is where empty sections of a file don't take up any storage space - I don't know if it did). However, "because someone decided it" is just as likely a reason as any other.

[jamie92208]

I suppose that you would have to find a programmer from the early 60's who worked on the opriginal system for Southern Pacific. Their locos all had integer numbers of between 1 and 9999. That is still a major drawback to the UMLER system that runs the US railroads today. That allocates locos by reporting marks eg SP, UP etc and the number. The 9999 limit wasn't a problem until the big SP/UP merger that created a need to 5 figure numbers. Up have created a new reporting mark of UPY for their switchers and have kept the UP series for road locos.

 

Jamie

[The Stationmaster]

I suppose that you would have to find a programmer from the early 60's who worked on the opriginal system for Southern Pacific. Their locos all had integer numbers of between 1 and 9999. That is still a major drawback to the UMLER system that runs the US railroads today. That allocates locos by reporting marks eg SP, UP etc and the number. The 9999 limit wasn't a problem until the big SP/UP merger that created a need to 5 figure numbers. Up have created a new reporting mark of UPY for their switchers and have kept the UP series for road locos.

Jamie

Possibly - although by 1974 BR staff were re-writing or making additions to TOPS software and I think - but am far from certain - that the change from a 4 digit field to one which allowed longer numbers for locos was written by BR (maybe possibly in conjunction with SP?). Later changes, such as adding coaching stock, were definitely written by BR programmers.

[Pugsley]

my memory/impression of things was that TOPS was altered to accept numbers longer than 4 digits.

That would make sense, as the SP only used 4 digit loco numbers, in common with other US Railroads.

 

Edit - Bah, must read the whole topic before posting

 

Jamie - I didn't know that, but that would explain why there are no US locos with numbers greater than 4 digits.

[keefer]

re: the 76s

 

plenty of pics of them being renumbered 'on the cheap' - whether in original serif or 'BR' corporate characters, simply paint over the 'E2' and plonk a 7 on top (sometimes with a modern 7 + serif others)

[frobisher]

I doubt that that is how it worked internally.

 

According to the information on the internet, TOPS ran originally on the IBM System/360 computer architecture and was written in assembly language macros. Based on that, I would assume that it stored loco numbers in binary-coded decimal and accessed the subfield for class or number by using a bitmask and bitwise boolean instructions to select the required nybbles - a System/360 machine with the right hardware options could do those operations very quickly, as the BCD architecture was always common in mainframe computing (it still is today in mainframes, but didn't make the transition to PCs). Division would be extremely slow and unnatural by comparison.

 

Except of course where it is implement exactly as you describe

[royaloak]

I seem to remember that 6700 became something like 37123 and took the number of a withdrawn sister.

Jamie

6700 became 37119 (which was at the end of the split headcode locos) and 6819 (which was the first of the centre headcode locos) which should have become 37119 became 37283 in the place of withdrawn D6983.

[drjcontroller]

Just to clarify before the thread takes the usual route of muddlement or repeating incorrect information - the point has already been established that diesel renumbering, with the class 45 Peaks, started in 1973. My clear recollection is that with 45s, it started that spring, in advance of the rest of the diesel fleet, and is confirmed by Roger Harris' allocation history

 

I know what I saw, I still have the photo and I am certain of the date. The downside is I have no way of proving it, so I have to bow to conventional wisdom.

[scorpion]

my first encounter with a tops number was at crewe works in 1972 where the prototype HSDT power cars were under construction, the guide told us they were 41001 and 41002, at the time we assumed it was a new number series for multiple units

 

around the same time some of the AL3s in doncaster works had strange cabside chalk markings like 83.011

 

only when railway magazine showed a picture of 76 050 did all these strange things make sense.

 

shortly after i saw 83.011 fully painted and numbered, with full stop and numbers on all 4 corners.

 

not until 16/4/73 did i encounter 45 102 and 45 104 both ex works at derby

 

on 24/6/73, of 15 class 76s on wath shed only 4 had tops numbers

 

from autumn 1973 tops numbers appeared on more and more ex works locos but my notebooks show that from 1st feb 1974 the eastern started renumbering everything on shed including green locos

[leopardml2341]

shortly after i saw 83.011 fully painted and numbered, with full stop and numbers on all 4 corners.

 

 

 

I'd forgotten about the full stop..............

 

I did see some like that, but didn't have the presence of mind to record which ones, as I said in OP only a young 'un at the time.

 

Edit - In fact the more I think about it, the more I'm convinced that 45101 was in fact 45.101 ?

[emperordalek]

Hello All,

 

Don't know why this popped into my head, but it did and I've spent a bit of time searching this site and couldn't find an answer - so sorry if someone else does.

 

What (and when) ? was the first loco to carry TOPS numbering - I remember my first sighting 45 101 - I guess after ETH conversion and probably sometime mid '73 ? but I was only a young'un then.

 

Thanks in anticipation.

 

Andy Evening Sir! I answer to you query - as soon as I read it a class 76 sprang to mind - and lol & behold 26032 becoming 76 032 on 19 Jan '72 could be the lucky winner - i haven't read the 48 replies or whatever is quoted. 45 101 was March '73 btw. Hope this helps. Malcolm