scanning the GWR wagon register - a proposal

[drduncan]

Dear all,

 

For those interested in wagons and the GWR’s wagons in particular the wagon registers at Kew and York are a mine of useful information. But being so far apart consulting them is difficult for many of us (for example Kew is easily doable for me but York not).
 

So I was wondering if as a community we could collaborate in photographing the registers and sharing the resulting images.  

 

I imagine we’d need two teams, one photographing the Kew files, one the York.  We could then either set up our own io type group to host the images, or ask Andy York if RMweb could (but I suspect there would be too many files) or ask an organisation like the GW study group to host it in their io group.

In the much longer term we could transpose the information into an excell spreadsheet which would be easier to share and be searchable - but this would be a very long job.

 

What to people think?

 

Duncan

[Gordon A]

Interesting idea.

 

[ikcdab]

How does that fit with the ownership/repro rights of kew or NRM?

I think it's OK to make copies for private research, does this idea go beyond that?

I think it would need kew/nrm agreement,otherwise excellent idea.

[corneliuslundie]

Just to flag up that the GW  wagon registers were gone through in the past by the late Harold Morgan who listed all the wagons from many of the Welsh companies. The WRRC has his records, which have been used in compiling the relevant tables the two books of Mike Lloyd drawings.

Jonathan

[hmrspaul]

I don't want to be rude, but have you any idea what a colossal task this would be?

 

I have done some and found it necessary - with a goodish quality camera to be readable It was necessary to copy two or at most three entries at a time - and as these are spread across the two huge pages the two pages had to be done separately.  Partially this is necessary because the first use of numbers is crossed through and over written. It is the first uses which are of most interest - the second and sometimes 3rd uses are listed in the Atkins et al bible. 

 

Despite asking the WRRC I didn't get any help with looking for detailed info on some TVR wagons and did it all myself - fortunately when it is open the NRM is my local library, all of a 10 minute walk from the back door. 

 

Here is an example for https://PaulBartlett.zenfolio.com/gwrloriotlowmac/e90499c2  with information taken from Atkins et al and I'm not convinced it is interpreted properly and may be a later rebuild. 

 

[drduncan]

4 hours ago, hmrspaul said:

I don't want to be rude, but have you any idea what a colossal task this is.

 

Yes, that’s why I feel a collaborative approach that shares images and means people can avoid going over the same ground might be fruitful.

Duncan

[corneliuslundie]

Paul was unfortunate with the TVR as no-one has got around to transcribing Harold Morgan's records for that company. I shall probably make a start in two years or so when the Circle gets on to the  books of TVR drawings by Mike Lloyd. You can see from the attached scan of two pages, which is all I have, that even that is a fair task. And I agree that photographing the GWR books would be a massive and logistically difficult task. Because of the many amendments, good quality images would be needed.

Of course the absorbed company records are all mixed up in the books because of the way the GWR allocated numbers.

Jonathan

PS for the Rhymney, Barry and Cambrian I am happy to help with enquiries as I have transcribed Harold's lists - and of course for the first two the lists have been published.

HM-TV wagons pp4-5.pdf

[Chrisbr]

Great idea and and hmrspaul says a humungous one, but that doesn;t mean it can't be done.  In following up my interest in wagon number plates, I have made copies of 13 books already. Taking the pictures with a modern smartphone is relatively easy and provides good qualtiy photos - one photo per page of the register, so details of each wagon split across two photos.

You still have the issue of the axlebox column on the left hand page often being hidden by the binding/fold of the page but that can be true even if sitting in front of the book.

 

As I see it there are three big issues to overcome -

 

1 - Obtaining permission from NRM and Kew to publish the work - as has been mentioned above, all of mine were signed off as being for personal use only. I suspect permission would be available if approached correctly, not least to avoid us all thumbing through large, heavy, relatively old and fragile documents.

 

2 - Having somewhere to host the information. A few suggestions already and i can think of more (HMRS possibly being one) and the correct choice here would probably facitilate the permission needed above. I have no contacts with any of the potential groups and so can't speak to any that may interested or willing.

 

3 - Storing the information in a manner that allows you to find things - this is the elephant in the room as alluded to by Corneliuslundie - it's all very well having the data, but unless you are looking for a specific number, you have no way of finding the other information without transcribing it into some external data source, be that spreadsheet as suggested, or web based database as I have prototyped, which could allow various searches and also allows you to keep the original photos associated with the records so reference can be made back to the source material.

I have calculated that for the limited info I have (less than 10% of the fleet) it would take me 7 years to transpose the info assuming I did 5 numbers a day (which can be up to 15 wagons) every day of the year, year in, year out..... Which is why I haven't started!

 

All that said I believe it would be a great benefit to people interested in the history and the modelling of said wagons, so would be keen to see progress made and would help in anyway I could. Not to forget the coach and department stock books and the Lot registers......!!!!!

 

Chris

 

[drduncan]

With regard to the search issue, I’m told by a chap who is paid to understand such things the the open source Mayan document management software has an automatic ocr facility. I’m experimenting with Mayan at work and one of the tests will be to see how good the ocr facility is wrt jpeg images.  If it is all that is promised it may be the solution.

DrD

[hmrspaul]

9 hours ago, drduncan said:

With regard to the search issue, I’m told by a chap who is paid to understand such things the the open source Mayan document management software has an automatic ocr facility. I’m experimenting with Mayan at work and one of the tests will be to see how good the ocr facility is wrt jpeg images.  If it is all that is promised it may be the solution.

DrD

Quite honestly it is difficult for the human eye to read these. This is because there is an original entry in a very (to our eyes) italic type script which would be very difficult to read if it hadn't been crossed through and then over written by one, and sometimes two, new uses of the number. I'll send you a few of my photos of the register if you want to try this but I think it will be quite confusing. 

 

Now I'm on, another problem is the use of abbreviations - if you look at my entry above of 42009 I can't make out the original entry - which must have been a very long time ago because the replacement entry is for what became a Loriot - built in 1890. Apparently Cond in 1960 but "made into SV no. DW42009 OSR" My photo  https://paulbartlett.zenfolio.com/gwrloriotlowmac/h90499C2#h90499c2 more than 20 years later shows that it is fundamentally similar to the original condition shown in the Diagram books - including the very fine buffers but some up grading. But what is SV and OSR ? Just using this as an example - it was the first I had to hand but it does illustrate the complexity. Incidentally it mentions new lot 1059 which is for Mogos. 

 

Paul

[drduncan]

SV = service vehicle maybe?

DrD

[drduncan]

No idea about OSR.  Maybe the stationmaster or Miss Prism might have a scoob....

[bécasse]

On 23/08/2020 at 12:56, ikcdab said:

How does that fit with the ownership/repro rights of kew or NRM?

I think it's OK to make copies for private research, does this idea go beyond that?

I think it would need kew/nrm agreement,otherwise excellent idea.

You cannot copyright lists unless sufficient additional work is done in presenting or analysing them to create intellectual property rights.

Ironically, that means that the originals in the NA or NRM cannot be copyrighted but that the sort of list which I think is being suggested would be because of the vast amount of (intellectual) work involved in sorting the entries out.

[ikcdab]

17 minutes ago, bécasse said:

You cannot copyright lists unless sufficient additional work is done in presenting or analysing them to create intellectual property rights.

Ironically, that means that the originals in the NA or NRM cannot be copyrighted but that the sort of list which I think is being suggested would be because of the vast amount of (intellectual) work involved in sorting the entries out.

Well yes, I deliberately didn't use the word "copyright" as I understand the issues. The NRM/Kew cannot copyright the indexes, not only for the reasons you quote but also because any possible copyright will have expired. 

However, they do own the documents and can permit or deny scanning and republication as they see fit. Normally they allow reproduction for private research, but you would need their permission to publish copies of their documents.  I believe you would, however, be able to publish transcripts.

[drduncan]

So, to try to move this along:

1. Who already has images of the stock books from Kew/York AND (all other things being equal) be prepared to contribute them to the project? (Please detail references and piece numbers so we can avoid going over the same files).
2. Who would be prepared to photograph stock books in conjunction with others at:

a. Kew

b. York?
 

Volunteers only reply please, negatives are not required.  Also please let’s not keep going round the houses about what might or might not be possible. If enough people can support the project to make it plausible we can then go into the details...

Duncan

[SP Steve]

One possible way of deciphering details from images lies with zooniverse:

 

https://www.zooniverse.org/about

 

I found it by chance at the start of the lockdown as there was a project asking for volunteers to transpose old UK rainfall data into actual digital details that could be used in modern day applications - as I recall the project would have taken years but was done and dusted by June.

 

I seem to recall that each image was processed three times by individual volunteers to help eliminate incorrect entries.

 

Not saying it would be the answer but maybe worth considering.

[Adam]

In theory, this is an excellent and worthwhile project, which has lots of parallels in historical research so there are a number of established methodologies. I'm looking at this as a working university historian who has been involved in a couple of decent-sized transcription-based database creating projects. You can forget a research council funding this by the way! Whatever you decide, it would be a very good idea to talk to the relevant specialist at The National Archives (TNA) first. I'm not sure who that would be (I'm a medieval historian and the TNA is huge), but my friends at TNA are very helpful and knowledgeable. You might also find it useful to talk to the Head of Research at the Railway Museum who knows how to frame a research project - they're in no position to fund this either by the way.

 

So some examples. None of these are directly analogous, but should give you some idea of how this sort of project can look:

 

There are the census transcriptions that fuel the genealogy industry (many, for complicated reasons by the Church of Jesus Christ and Latterday Saints) - these are a bit rough and ready, but remarkably comprehensive (i.e., whatever the population if the time was), making mere wagon registers seem small. Check out Ancestry, Find My Past, etc.

 

There's the Anglo-American Legal Tradition project which photographs and puts online many hundreds of thousands of images of English legal documents (for free, for researchers with a modicum of guidance): http://aalt.law.uh.edu/ - note that this is supported by endowment and the images are licensed by TNA: it's not quite as simple as taking pictures systematically and putting them online. 

 

Transcribing Bentham: https://www.ucl.ac.uk/bentham-project/transcribe-bentham - 20,000 pages in various languages, quite a high editorial overhead on this project with lots of data checking. Of course the form in which the transcriptions are presented needs thinking about and there is the question of searchability and workability: this sort of volume takes a lot of server space and that has a cost. That said, it shows that old writing in English is perfectly readable once you've got your eye in.

 

The Railway Work, Life and Death project has a database of historic railway accident reports: http://www.railwayaccidents.port.ac.uk/. It is an academic research project  - and it isn't meant to be comprehensive except for certain sorts of records - and the Principal Investigator, Mike Esbester is someone I know and has a lot of time for the levels of very specific knowledge that exist among enthusiasts, modellers, etc.

 

So if I were contemplating this seriously the questions I would need to answer are these (I'm guessing to a certain extent because most of the published work on wagons that I have lacks meaningful archival references).

 

1. What documents are we actually discussing? i.e., what are the TNA series and item numbers that appear in the catalogue? (e.g., RAIL 254/256 - that's the one I found easily from a quick search of Discovery, so presumably the series is TNA, RAIL 254). How many are there?

2. What is the scale of the job? What form do the documents take? I'm guessing large ledgers? And how are these laid out? 

3. What - if any - work has been carried out on these documents to understand/show how they were put together and for what purpose? Has this been published? This would be enormously helpful in designing a project.

4. Is the information consistently recorded under the same headings for every wagon/class/group? Would you like to preserve this in any resource you create? If it is then a database-driven approach might be best, if not then transcription might serve you better.

5. Do you want this to be comprehensive? If so, how do you achieve that? Document by document is probably the best way to preserve consistency, but would that yield the result you want? Do you want a handlist to find out about particular wagon types for those with access to the documents (and that *is* valuable and does not exist as far as I know) or full coverage of the GWR wagon fleet from 1837 to 1947 (or later?). 

 

Sorry that this is a bit long, and apologies to Duncan if he's already thought of all of this, but the sort of scale and crowd-sourced research you're talking about is something where others have gone before and it's much better to work with them where you can.

 

Adam

 

 

 

Edited August 31, 2020 by Adam

[drduncan]

As I’ve had no volunteers come forward we can consider this thread closed (and locked if the Mods wish).

Duncan