A database of GWR photo references

[Clearwater]

Another point to consider is the same photo appearing in multiple books - not sure how to this could be indicated.

I think that's ok? After all not every wants to buy/find every book. From a practical perspective, unless the same person makes the entry, they may not realise it's a duplicate.

 

There's also scope for confusion as some locos were very region specific and hence photographed a lot in the same place. Not to mention the celebrity engines, how many photos of KGV alone!

[Stubby47]

True, but if the spreadsheet shows there are 3 photos of XYZ in three different books,  there might be the temptation to purchase all the books expecting different images.

 

Maybe include the photographer/photo collection ?

[Clearwater]

True, but if the spreadsheet shows there are 3 photos of XYZ in three different books, there might be the temptation to purchase all the books expecting different images.

 

Maybe include the photographer/photo collection ?

If it's given..(I know it should be credited but I wouldn't wager it's universal)

 

.tbh, the temptation to purchase the books is probably already there by that point!

[kevinlms]

If it's given..(I know it should be credited but I wouldn't wager it's universal)

 

.tbh, the temptation to purchase the books is probably already there by that point!

Acknowledging a source should always be given, if you don't know use 'Anon', as it's better to search for something rather than a blank space.

 

Edited to make more sense!

[kevinlms]

I agree Kevin. The suggestions so far are very loco-oriented. Having gone to so much trouble extracting data from potentially hundreds of books it would be a pity if, for instance, one cannot list all photos of, say, the Barnstaple branch. (A favourite of mine). Yes, I know I could do an individual search for each of the 12 stations & halts, but then there are the 2 major viaducts (both of which go by 2 or 3 names of which I've got to be aware) & there's always the possibility of photos between stations. No doubt others could come up with other requirements.

It isn't just stations. A search for 'Barnstaple' in such a database, ought to come up with a couple of stations, LSWR & L&B, as well as a well known West Country 4-6-2, including Hornby-Dublo/Wrenn models. That is something I know, without knowing anything specific about the area & its railways. Indeed from the last couple of sentences, its probably clear that I have a very sketchy knowledge!

 

A database as I have suggested can & does easily sort out the unwanted items from within it. Not sure than an Excel spreadsheet can do so.

[kevinlms]

This is a great idea but I think the essential thing is to keep it simple.

Many years ago I was involved in the HMRS Photocat project. Three of us spent a great deal of time defining locations in South Wales. I still have the file and it is enormous (and was never used in the end). That was intended to ensure that, for example,  photos of a signal box in the country would always be attached to the same location. And also to standardise the names of locations, as they often changed over time. So to formalise locations for all of the GWR alone would be a truly massive job and almost certainly not worth it. An awful lot of work for relatively little benefit.

I have also written numerous indexes for publication.

But there does need to be a protocol about basic things such as how you insert loco classes, wheel arrangement etc or searching becomes laborious. So is it 4-6-0 or 460? Collet goods or 2251 (a problem here as you can't use 22xx and 2251 is also a specific loco).

Not relevant at this stage but I have been trawling sources for relevant photos of Cambrian Railways locos - not just photos in books but also sources of photos and drawings. At the moment it is a text file but it runs to many pages and is not complete. I intend to do the same for other rolling stock but I suspect that the lists will be much shorter. That is probably not true of the GWR proper. I also have an Excel file of private owner wagon photos relevant to South Wales. That has hundreds of entries and is not complete as I still have at least two books to add. This not meant to be boasting, but to point out the magnitude of the task. And there are other such files available on the web.

But it does raise the question of including photos in collections where copies can be purchased. My feeling is not to as it would add greatly to the task and the databases are usually searchable, but it would I think in the preliminary sections be very useful to add a list of such sources. Such lists as the HMRS, RCTS, RCHS, WRRC, other relevant Societies, Lens of Sutton Association, Kidderminster Railway Museum etc.

Incidentally, when doing the Cambrian list it was noticeable that some photos came up repeatedly. There needs to be a way of coping with this in the database.

Jonathan

The answer as to how to enter entries, is to use the most common format every time, including spelling! If a name has 2 spellings, use both, then a search will find an entry, which may prompt someone, who doesn't know about the alternative spellings, to look further. However (a pet hate of mine coming up) its confusing when people use 'Garrett', when they mean 'Garratt' in reference to various articulated locos, built by Beyer Peacock.

If talking about traction engines & fittings to suit, then 'Garrett' is more than acceptable!

 

An example for what you asked earlier 'Is it 4-6-0 or 460'? ALWAYS use the former on a written format, its what most enthusiasts use (modellers and/or prototype). Also '460' could be part of a loco number (or all of it)

[Clearwater]

Kevin

 

A couple of questions:

1) how much does Microsoft access cost for a user to install?

2) how long would it take someone who knows what they are doing to create a relational database as you propose?

 

Thanks

 

David

[kevinlms]

Kevin

 

A couple of questions:

1) how much does Microsoft access cost for a user to install?

2) how long would it take someone who knows what they are doing to create a relational database as you propose?

 

Thanks

 

David

David

 

For the first question, you will have to go to Microsoftstore.com and find your local price. They do have it on special from time to time. I have Access 2013 & it didn't cost me much at all a couple of years ago. It comes 'free' with some versions of Office.

 

There is a free Access viewer around.

 

For 2, I don't know. I'm certainly no expert & I muddled my way through.

 

As always, its easier to modify an existing database structure, than start from scratch.

[Clearwater]

David

 

For the first question, you will have to go to Microsoftstore.com and find your local price. They do have it on special from time to time. I have Access 2013 & it didn't cost me much at all a couple of years ago. It comes 'free' with some versions of Office.

 

There is a free Access viewer around.

 

For 2, I don't know. I'm certainly no expert & I muddled my way through.

 

As always, its easier to modify an existing database structure, than start from scratch.

Thanks Kevin. On 2, that's my point. I've done enough in excel, and access, to know that building a robust functional database is a complex task. It's not impossible, by no means, but the pitfalls you recognise and others also highlight in my view make it impractical for what's proposed here. Excel is dead easy to use. You can search and skip term to term with "control f", you can add filters to columns in moments. If you're sophisticated, it's not hard to build analytical tools to , e.g. Count number of times a term occurs. It's why it's so widely used. In my professional life, which is in a data heavy world, I don't use access, only excel. That's been common in all of my jobs.

 

Nothing's stopping you creating such a database if you want to but I'd wager that as you get into the complexities of defining fields and what you want it to capture, it will become a monster.

 

David

[martinT]

Excel allows filtering on multiple options - as there is already a column for Location, a simple filter search will produce all images for a line.

 

Don't over complicate the columns, use the functionality.

How is this going to work? Unless there is a field defining 'Line' how does Excel know that Venn Cross (say) is on the Barnstaple branch? If a user wants a listing of all photos on the branch he shouldn't have to work out all the possible locations, picking from a very long list - that would only be acceptable for a short line with which the user is familiar. No one for a moment would expect to have to select all 171 Castles numbers just to provide a listing of all the Castle class photos. As I've said in previous message this is a very loco-centric - it would be a pity if the considerable efforts of what I hope will be many of us are not fully utilised.

 

Martin

[Clearwater]

How is this going to work? Unless there is a field defining 'Line' how does Excel know that Venn Cross (say) is on the Barnstaple branch? If a user wants a listing of all photos on the branch he shouldn't have to work out all the possible locations, picking from a very long list - that would only be acceptable for a short line with which the user is familiar. No one for a moment would expect to have to select all 171 Castles numbers just to provide a listing of all the Castle class photos. As I've said in previous message this is a very loco-centric - it would be a pity if the considerable efforts of what I hope will be many of us are not fully utilised.

 

Martin

Martin

 

If, as miss p suggests, there is an optional line column, you vcoild use that. However, as a pointed out before, I don't think line is an unambiguous concept. Castle class clearly is (hence the field for both loco and class.). In any event, the search is only as good as the input data. I'd be wary on relying on a search to be accurate. In excel, you can filter in a choice box at the top of a column. If you wanted specific locations, you'd have to put a tick in each box of the location you want. Ultimately, any tool like this is not going to do everything for you - you'll still need to y some judgement and apply some research techniques. The point is that it massively speeds up a search to use an electronic tool rather than plough through the indices of 50 odd books (for example).

 

I'm well aware that there are some fantastic database products out there - any hotel website is a great example. But those people will spend many millions on maintaining and developing the software to allow you the type of precise search you mention.

 

I'm also not sure what I produced was "loco centric" - you'll note it included columns for location and train formation including carriage diagrams.

 

David

[kandc_au]

For those who don't have Office on their computers , there is always Open Office which is free on the net.

You do have the option of saving your work in either the standard formats, that is Open Freeware formats OR saving your work in Microsoft Office formats.

 

Khris

 

Edited for spelling.

[Stubby47]

If we're talking line-centric photos, there might not be a loco or any rolling stock in the image.

 

This view of Shepherds is quite well used, but would not easily fit into the proposed format of data requirements

 

http://goonhavern.com/wp/gallery/goonhavern-district-o-c-s-gallery/stepherds-station-showing-treamble-line-off-to-right/

[Clearwater]

Stu

 

There's a few general views in the GWR in the 1930s albums. I entered those as loco "n/a" but then filled the location field with, e.g. (And from memory), Exterior Bristol Temple Meads.

 

Edit: if you look at colour-rail.com they have a relatively simple structure that groups lines into 22 broad categories. I.e. To use Martin's example, you'd have to know a)which area the Barnstaple Branch fitted into and b) then search through each result. You could search on an invidual location as well but then you'd have to separately input each place you wanted to search. I've not tried to see whether if you separated variables by comma or semi colons you can do multiple searches

 

David

[Miss Prism]

I'm not disagreeing with you, Martin, but there are potential problems with the 'line' description, whether in a separate column or not, e.g.:

 

- a contributor might not know (or be interested in) the 'line name' when making an entry;

- line 'names' are often not standardised (and steeped in historical complexity);

- junction locations can be confusing, e.g. Dulverton (on both the 'Barnstaple line' and the 'Exe Valley line')

 

The centricity and centricities of the list will reflect the focus, knowledge and effort of its contributors. None of the fields on the sheet are mandatory. Thus any loco-less picture, like Stubby's Shepherds one, can simply have a location entry (and possibly date and any other things in the notes column a contributor might want to add).

[Clearwater]

I think for anyone using such a spreadsheet, there are other workarounds to make searching more efficient. E.g. Book title often gives a major head by - e.g. Middleton press title their booos in the form "the so and so branch" or "the withered arm between the wars". Using such commonsense filters helps narrow a search

[Stubby47]

I'll be doing the 'Branchlines to Newquay' (Middleton Press) book

[Stubby47]

Question - if a photo has more than one loco, should there be 1 row per loco added to the spreadsheet ?

 

I would suggest 'Yes'.

 

( makes note to avoid books on loco sheds )

[jonny777]

There may be a problem with the updating process. If everyone can access the database to add further books, what safeguard is there to prevent someone deliberately (or accidentally) erasing some or all of the info already there? Or, maybe just entering incorrect information.

 

I apologise for the negativity, but it comes about because the idea is on the face of it, brilliant, simple and popular; but on closer analysis of the complex requirements and the myriad of search-string possibilities, it suddenly becomes a vast undertaking.

[martinT]

I'm not disagreeing with you, Martin, but there are potential problems with the 'line' description, whether in a separate column or not, e.g.:

 

- a contributor might not know (or be interested in) the 'line name' when making an entry;

- line 'names' are often not standardised (and steeped in historical complexity);

- junction locations can be confusing, e.g. Dulverton (on both the 'Barnstaple line' and the 'Exe Valley line')

 

The centricity and centricities of the list will reflect the focus, knowledge and effort of its contributors. None of the fields on the sheet are mandatory. Thus any loco-less picture, like Stubby's Shepherds one, can simply have a location entry (and possibly date and any other things in the notes column a contributor might want to add).

• Equally contributors might not be interested in the loco's livery or the coach diagrams but fields are provided inviting an input.
• There are always going to be problems classifying photos - is a Dukedog a Dukedog, a 90xx, a 32xx, or an Earl? All are feasible but the first 2 most likely to be used. When it comes to lines I would suggest the official GWR descriptions as used by R A Cooke in his 1947 Atlas. Some modification will probably be needed for very long lines. Thus the GW's official 'Main Line' (Paddington to Penzance via Bristol & Plymouth Millbay) I would suggest should be split into: Bristol ML, Bristol & Exeter ML, South Devon ML, Cornwall ML, & West Cornwall ML.
• Yes, junctions are a problem & I think some knowledge of the GWR's geography is inevitable!

 

 

Can I raise the possibility of duplication of effort. I've started indexing 'The Early Years of Western Region Steam: An Album of P M Alexander's Photographs' (compiled by P Whitehouse, Wild Swan 1983). It's going to take some time - & I'd hate either to waste someone else's time or have my time wasted. Please can everyone announce what books or other publications they intend to do.

 

Martin

[Miss Prism]

Stubby47's and MartinT's offers of book coverage are now earmarked on the sheet.

[Miss Prism]

There may be a problem with the updating process. If everyone can access the database to add further books, what safeguard is there to prevent someone deliberately (or accidentally) erasing some or all of the info already there? Or, maybe just entering incorrect information.

 

I am conscious of these significant potential problems.

 

Currently, updating of the whole file by a contributor could lead to inadvertent or deliberate alteration of existing data, but it does has the virtue of providing the overall (column) architecture, which will save the contributor some work. The alternative submission mechanism (i.e. sending in a separate sheet) would leave me only to cut and paste the new material into the server file, and safeguards against the unauthorised change problem. The alternative submission mechanism will also cater for the situation where the current server file is being updated, at the same time, by different contributors. It is therefore possible the alternative submission mechanism might become the only one available in the future. Let's see how it goes.

 

Light-touch regulation and unobtrusive change-control are I feel desirable at this stage. We are proceeding here on the basis of best mutual intentions. We need to trust each other.

[firebirduk]

I am conscious of these significant potential problems.

 

Currently, updating of the whole file by a contributor could lead to inadvertent or deliberate alteration of existing data, but it does has the virtue of providing the overall (column) architecture, which will save the contributor some work. The alternative submission mechanism (i.e. sending in a separate sheet) would leave me only to cut and paste the new material into the server file, and safeguards against the unauthorised change problem. The alternative submission mechanism will also cater for the situation where the current server file is being updated, at the same time, by different contributors. It is therefore possible the alternative submission mechanism might become the only one available in the future. Let's see how it goes.

 

Light-touch regulation and unobtrusive change-control are I feel desirable at this stage. We are proceeding here on the basis of best mutual intentions. We need to trust each other.

 

 

Please count me in on this project - what a boon.

 

There have been several posters not particularly happy with Excel - yes it has its limitations but has been stated the alternatives for various reasons (mostly complexity and cost) are a non-starter.

 

Forgive me if this has been stated before but what we need is a standardised spreadsheet formatted with all input inquiries that can be downloaded from gwr.org.uk. No doubt that this will spark the usual negativity!

 

Mike Spence

[Miss Prism]

Forgive me if this has been stated before but what we need is a standardised spreadsheet formatted with all input inquiries that can be downloaded from gwr.org.uk.

 

I had been contemplating doing exactly that this morning, Mike.

 

Hang on a moment.

[Miss Prism]

A blank, formatted, sheet is now available from http://www.gwr.org.uk/no-photo-database.html