Using FileMaker Pro
Over the next few weeks you will need access to your own computer and a connection to the Internet. If you havent already done so, you need to obtain a laptop computer. You can get onto the internet anywhere on the DMU campus, all you need to do is scan wireless networks and select DMU-Wireless, then just log in with your username and password.
During the class you will be shown how to open a new Filemaker Pro database file, how to create a simple table, including naming the fields and entering data into those fields to create a set of records. You will learn how to create and link multiple tables and files, methods for importing data and preparing data to make it easier to find and how to publish your database on the Web.
Download a 30 day free trial copy of FileMaker Pro 12 to your laptop computer from  (select the "try now" option).
By the end of this session you should be able to:
- Create a Filemaker database.
- Link together multiple tables and link Filemaker Pro records to external files.
- Import data into your database from external files.
- Create a controlled vocabulary to aid data discovery.
- Query your database for useful information.
- Publish your database to the Web.
Constructing a simple database
- Using the simple table of the Roger Fenton letters from the Crimea that you drew last week, create a single table database in Filemaker Pro.
- Working together as a team, populate the fields in the table with data provided in class to create a working database for the complete set of 25 letters.
As it happens we have made digital transcriptions of both sets of letters. Rather than laboriously retype all this information into your database it makes more sense to import the data if possible. In this exercise you will be shown how to:
- Import transcripts of the letters from a text file provided.
Constructing a complex database
The data that lie behind the Roger Fenton Letters from the Crimea Web site are more complex than the site suggests at first reading. The Web site is actually a compilation of two sets of letters. You can read about how it came about here. The letters were written by Roger Fenton mostly to Annie Grace Fenton, but then copied out by hand by Annie Grace and by another relative, Joseph Fenton. We do not have the original Fenton letters but one set each of the copies made by Annie and Joseph have survived. These two sets of transcriptions do not reproduce exactly the same set of letters and sometimes even within the two copies of the same letter there are variations in content. To a photohistorian, the Roger Fenton Web site is probably of only limited use as it stands. It might be useful to be able to compare the Annie Grace and the Joseph Fenton transcriptions with each other. We have got digital scans of the letters as copied out by Annie and Joseph, so we could put these on the Web for scholars to examine. The disadvantage of these is that they are images of the letters, so it isn't possible to perform any sort of digital computations on the data, such as searching for or counting particular words. So it might be useful to include digital transcriptions of all the letters in addition to the page scans. So, in total then, instead of just one composite transcription of the letters, we are now considering a database to hold two different sets of transcriptions and two sets of page scans.
- Using the linked tables you drew in class last week, construct a multi-table database in Filemaker Pro and link the tables using primary and foreign keys.
- Copy page scans of the handwritten letters to a directory on your hard disk and import them from there.
By now you have quite a few tables and files in Filemaker, but you don't yet have a working database because the files are not yet linked together. In this exercise you will be shown how to link together files to create a working database. Files are linked together using "keys". A "primary key" is a unique indentifier for a particular record. In the case of PEIB exhibits we could use the catalogue numbers as primary keys, except that PEIB comprises a collection of exhibits from a number of different exhibitions 1839-1865, so its likely that the same catalogue number refers to several exhibits. To ensure that each exhibit has a totally unique identifier its advisable to assign each a unique number.
|Exhibit ID No.||Cat No.||Title||Process||Photographer||Date||Location||Price mounted||Price unmounted|
|1||19||Scott Monument||Talbotype||Adamson, Robert||1859||Aberdeen, British Association||no data||no data|
You can imagine how tedious it would be to have to enter the exhibition details over and over again if there were a lot of exhibits in an exhibition. A way round this is to split the exhibition details off into a separate table that can be linked back to the exhibits table using another primary key, this time an exhibition ID number.
|Exhibit ID No.||Cat No.||Title||Process||Photographer||Exhibition ID No.||Price mounted||Price unmounted|
|1||19||Scott Monument||Talbotype||Adamson, Robert||1||no data||no data|
|Exhibition ID No.||Date||Location||Host|
We have seen how, by splitting tables, we can eliminate duplication or redundancy. We have also seen that primary keys can be used to distinguish between different records to eliminate ambiguity. Computers are not good with ambiguity (although there is a branch of computational intelligence called "fuzzy logic" that can cope with ambiguity). To see an example of this problem, take a look at the exhibit prices listed on this catalogue page from the 1896 RPS exhibition. The prices are the information listed immediately after the processes, eg. 15/- 10/6, 2 guineas, etc. If you dont know what these are take a look at these sites  and . The problem is that there are duplicate terms for certain currency values. A half crown is the same a 2/6 (two and sixpence); a florin is the same as 2/- (two shillings); a sovereign is the same as £1 (one pound); while a guinea is the same as £1.1s (21 shillings or one pound and one shilling). In order for the computer to search for exhibits at different prices all these different expressions of value have to be converted into a standardised set of terms.
Try using the advance search function in Exhibitions of the Royal Phototgraphic Society to look for exhibits priced at 10/6. Can you deduce how this search function works?
Another example from the same source is the names of the different sections within the exhibitions. If you compare the exhibition section titles listed in advanced search with the exhibition section titles in the page scans of the exhibition catalogues themselves you will see that there are a great many more variations in section titles than are shown in the advanced search drop-down menu. The database section titles have been standardised to ensure that if you search for one of the section titles in the drop down list, you are guaranteed to get some hits, even if the exhibition section titles you find are not exactly the same as the one in the list that you selected.
So, the way to deal with heterogeneous data is to standardise it. Coverting archaic monetary values to a decimal based system and simplifying the variations in exhibition section titles are examples of standardisation. Another is the notion of "controlled vocabularies". You have already encountered this idea. We discussed controlled vocabularies when we looked at Archives and catalogues.
Imagine you wanted to search the Exhibitions of the Royal Photographic Society database for a particular term, such as "Captain". What would you search for? Try searching for "capt", "capt.", "captain". The problem should be immediately obvious. There are ways around this problem such as wild cards, stems, and look-up tables and controlled vocabularies. To see some examples of controlled vocabularies, see the drop down menus in the advanced search in http://erps.dmu.ac.uk. Controlled vocabularies are usually based on lists of important and/or frequently ocurring words in a text, known as "concordances". Now, using the concordance list provided in class, create a controlled vocabulary of useful search terms for the Roger Fenton database.
The principal purpose of a database is to allow computational activities to be performed on the data, ie "queries" such as "How many letters did Fenton write from location X?, "Did he mention general Y in his correspondence? This session will show you how to add queries to your database.
Many of the events, people and places that Fenton described in his letters were also photographed by him and many of the photographs have been digitised and put online. In this exercise you will be shown how to:
- Go to the Fenton Crimean War photographs on the Library of Congress Website and download low resolution copies of relevant photographs illustrating events, characters etc in the Fenton Letters.
- Enter these photos and relevant data about them into the master database on the server.
Publishing your database
So far you have built a database that you can access while you are logged on to our Filemaker server. Understandably access to the server is controlled, so most people cannot view this database. In this final session you will be shown how to publish your database to the Web.
Use what you have learned to start thinking about how to build your own database for your archive.
Follow this link to the PRDA Contents page to return to the module contents.
Back to the MA: Photographic History Main Page to return to the Course contents.