digitizationblog News on digitization in libraries and allied institutions

The beta version of Streetprint Engine 3.0 has been out since March but I've only now gotten around to taking a close look. Described on its website as "the world's most user-friendly free software for powering digital collections," Streetprint Engine 3.0 is written in PHP and uses a MySQL backend, and runs on any platform that supports these tools. I haven't looked at Streetprint Engine version 2.1, but the website states that 3.0 adds "CSS layouts, new ways to search, browse, and view collection items, customizable fields, blog-style news posting, a glossary feature, an OAI-compliant metadata repository."

Streetprint is a product of the CRC Studio at the University of Alberta, a unit that "uses and also develops digital media to conduct, organize, preserve, and communicate research in the humanities, very broadly defined, for researchers of all kinds in universities, schools, public and private institutions, and the wider community." The presence of syllabus and course page tools and a glossary tool illustrates Streetprint Engine's humanities computing and pedagogical roots. The Streetprint site lists about a dozen collections that use the software, and also provides a support forum. As an indicator of the Streetprint community's activity, as of August 13, 2005 there were forty-nine posts on the forums, the latest dated Feb. 26, 2004.

Installation is standard for PHP/MySQL apps: set up a database and the required access rights, unzip the .gz or .zip file, copy the resulting files to your web directory, go to the install.php page and follow the prompts. As with many other such apps, you are prompted to log in using the default admin account, where you can change the password from blank. I had the software installed on my Windows laptop and on my OS X Mac in about 3 minutes each (apache, MySQL, and PHP were already installed and working, of course). A professionally produced, fairly thorough 63-page users manual (currently available only for version 2.1) takes you through the steps of installation and configuration. Also very useful, particularly for people without experience in digitization, is a chapter on "managing your collection," which explains the basics of image capture and processing, adding items to Streetprint, making them publicly available, editing them, adding accompanying full text, and deleting items. There is also a chapter on using Streetprint skins to customize your site.

Configuration involves adding users, site details, and "types". Users can be of two varieties, editors (who can do everything) and data entry users (who can only add items and edit their own items). Types include document types (you can define documents to be posters, books, telegrams, photos, etc.), image types (front covers, back covers, and inside images are supplied as sample image types for book items, but you would want to define something like Single Page images for still image items), media types (MP3, Quicktime, PDF, etc.), and categories. Categories are intended to desribe the subject of an document. Unlike the other type options, you can assign multiple categories to one document. Currently Streetprint Engine only supports a flat category list and not hierarchical categories. All type options appear in the browse and advanced search interfaces.

Metadata fields are hard coded but the labels can be changed using a PHP include file (changing them via the admin interface is on the to do list). The fields supplied are Title, Author, Reference Number, Publisher, Document Type, Category, Date, Date Details, Dimensions, Pagination, City, Illustrations, Location, and Notes. The Roadmap posting in the Developers section of the forums mentions adding "Field sets" for different types of items but no version number or planned date is included. Metadata handling is one aspect of Streetprint Engine that could be developed further. phpoai2, a PHP OAI data provider framework written by Heinrich Stamerjohanns, is included with Streetprint Engine 3.0 but is undocumented other than an statement in the Metadata administrative page that programmers can "fine tune" an installation's DCMI metadata settings by editing the SPDublinCore.obj file in the Objects directory. A link to the OAI base URL is included on the footer of every page.

Adding documents to a collection involves entering the required and optional values in the metadata details page, uploading the accompanying image or media files, and approving the item for publication. I'll focus on the middle step since the first and last are straight forward. After you have saved your metadata record, you click on the provided "add images" link. A series of pages follows that breaks the image addition process into the required steps. The wording on the next page instructs you to "First, move your web-ready images for '[item title here]' into Streetprint's upload folder." At this point, you have to FTP your files into the directory named "uploadfolder" on the webserver's filesystem. At a minimum you need a thumbnail and a low (web) resolution version for each image in your item, although you can include up to four versions of each image. The next page lists all the files in the "uploadfolder" that have not yet been attached to an item. You select all of the versions of the the files you want to attach (one image at a time), identifiy which version each file is (thumbnail, low res, etc.), and then click on a button labelled "Upload" to finalize attaching the files to the record. After clicking on this button you are told that the upload was successful.

I have to admit that the word "upload" used in this way caused me considerable trouble in understanding the workflow. Since many people understand the process of FTPing files as uploading, Streetprint's developers would have done better to use the label "attach" in this instance -- first you upload the files, then you attach them to the item's record. Also, I have been unable to find a way to avoid repeating this process for each image in each document. Adding a 400-page book one page at a time (and at least two images at a time, the thumbnail and the low res web display version) would be very inefficient and susceptible to errors. Some sort of bulk "upload" feature would be a welcome addition.

Despite these rough spots, Streetprint Engine is very easy to use and provides polished, slick search and browse interfaces. I encourage you to explore some production Streetprint sites so you can see these interfaces in action.

Given what is involved in getting collections up in other open source, general digital collection management software, Streetprint Engine's bold claim that it is "the world's most user-friendly free software for powering digital collections" is arguable. If the metadata handling capabilities were upgraded substantially, and if collection creators had access to better tools than those that are provided, Streetprint would be serious competition for applications such as Greenstone and Fedora.