Read the book View the pictures Use the index About Mr Gregory About this project

Building the website

Originally I had intended to recover the text from the printed page images using an OCR package but this has not yet been possible. So all 356 printed pages needed to be provided as images, as well as the 162 illustrations.

So we have 518 image files - or actually 680, because there is a thumbnail image for each photograph as well as the main image. In order to put these into a workable web site, its clear that the site design needs to make it easy for the user but has to be automated for the provider - the thought of hand-editing a change into each one of all of those web pages is horrifying!

The normal way of doing this is to automate at the web-server end using ASP or PHP. This wasn't really an option for two reasons - my previous web hosting package did not include it, although my current one does, and in any case I wanted to be able to put the book onto a CD for use where a web connection isn't available.

My first approach at automation was to do everything in client-side javascript - which was a really brain-twisting experience but eventually worked very well. However on coming back to the project after a while it became clear that although well-commented the code was difficult to maintain. It also ran the risk that the site would not work at all if the user had disabled javascript; and also was not accessible to search engines as it was actually a single web page despite looking different to the user.

The revised approach has been to set up an Access database with the image and chapter details in it, and to write VBA code to generate the web pages from these. Maintenance is now a case of "modify the HTML in a sample page to check it works; modify the VBA routine so it will generate that HTML; click a button to regenerate the web site." Each page now has its own keywords and description for search engines to use.

The appearance of the site is partly driven by the HTML code but mostly by the use of a standard CSS style sheet to set colours, fonts and, particularly, positioning.

Future plans? I hope to get better images of the plates from the book, as I now have a better camera (Fuji FinePix 5600). It would also be good to try to recover the original text rather than pictures of the pages.... time will tell. ..

Adding the Index

The original release of the website lacked a searchable index. To be of real use to local history researchers some way had to be found to index the site. Mr Gregory's original book includes over 20 columns of closely spaced index entries which I copy-typed into the Access database used to generate the site. Then I wrote routines to extract the keywords and build relational tables linking keywords, pages and index entries. After some careful design it was possible to generate the index pages. Formatting is done entirely using nested HTML definition lists and CSS.

Site design
©Roy Ayres, 2003-2006

Valid XHTML 1.0 Transitional Valid CSS!