Elias Tzoc blog!

A few days ago, we submitted the final report for the LSTA MiniGrant project we worked on this past summer.  We had five core activities: 1) Conversion to PDF and OCR of digital master page images; 2) Migration of metadata and new PDF content to DSpace; 3) Harvesting and indexing of metadata using MultiFacet; 4) Building a new public interface for the collection; and 5) Develop best practices and step by step instructions documentation.

I was responsible for # 4, DSpace provides two options for presenting data on the web: either using JSPUI or XMLUI.  Because of the need to create and customize the look-and-feel for individual collections in DSpace, we decided to use XMLUI –a new front end for DSpace developed at Texas A&M University.  Our previous experience in creating and implementing themes in DSpace was useful.  However, this project allowed us to test and implement three major customizations:

• embedding external files, for the new OpenZoom image-viewer, we had to create a separate copy of the JP2 files, an XSL template takes the filename of the primary bitstream in DSpace and calls an external PHP file which displays all the pages for the current issue.  We believe this method can be useful for OhioLINK, especially as institutions need to embed other types of files such as videos.
• displaying local metadata labels, digital collections often have a specific metadata fields, for instance the “Volumen No.” is only relevant for newspaper collections; having the ability to customize the front-end label will be a big advantage for OhioLINK institutions creating unique and special collections in DSpace.
• calendar view for browsing, we wrote a PHP script that generates a twelve months calendar view for browsing large collections based on date.  This feature can be especially suitable for projects such as yearbooks, magazines or newspapers.  The source data contains two metadata fields: a date in ISO format and an identifier for individual records.

Additionally, I helped with the implementation of a custom viewer within DSpace.  We implemented a dynamic image-viewer using open source software tools to provide an enhanced pan and zoom interface.  The OpenZoom viewer we used requires the use of Flash, which is not fully supported on mobile devices (e.g. Apple iPad); so instead we implemented a purely server side viewer that delivers standard JPG page images quickly to mobile devices. This technique seems to work perfectly fine for multi-page files such as a newspaper issue, the original final is saved as a DjVu and with the magic of PHP script (written by John Millard) we were able to convert the DjVu files into jpg “on the fly.”  This was the foundation for the Miami Student Newspaper collection for iPads.

The front page of the issue published on September 29, 1910 looks like this:

…and this month we’ll have a poster session about this project at the 2010 ALAO Conference in Columbus, OH.

.

For the new school year, the MUL’s Diversity Cluster will host some ongoing and some new activities.  The first one was a Multi-Lingual Greeting Card that we sent to all library staff, the goal was to highlight the number of languages spoken by people in the library.  It was an interesting exercise and we were surprised to learn that at least 10 languages are (truly) represented in our workforce.

For next year, it’ll be interesting to do something with all the languages represented at Miami for that, we’ll probably have to check the Office of International Education’s website -including this interactive page.

Early this month, we also invited Dr. Ron Scott, VP for Institutional Diversity.  He came to speak at our general staff meeting and shared with us the university’s goals/efforts to increase diversity on campus.

Other activities for the fall semester include:

• Hispanic Heritage Month, a LibGuides page where we’ll list selected authors, books, and related events both at Miami and on the web.
• Film Screening and Discussion, “The danger of a single story: Chimamanda Adichie” facilitated by Dr. Ray Terrell, Assistant Dean for Research and Diversity in the School of EHS.
• Social Justice Read-In, in collaboration with the Center for American and World Cultures, we’ll host a 2 hrs event as part of the Eighth Annual Human Rights and Social Justice Program at Miami.

…and on a separate/related note, the Diversity in Academe supplement from the Chronicle of Higher Education was just published.

I think this that’s it for now, nos vemos pronto!

.

Yes, today is the day!
Our first digital collection in DSpace is about to go “live” and I think there is reason to celebrate …for instance: it’ll be our very first digital collection in an Open Source System and with a dynamic image viewer (using OpenZoom); and perhaps most importantly, users now will have the ability to:
a) view/zoom individual pages using a full-screen option;
b) download a single PDF file for any given issue;
c) browse the collection using a calendar view; and
d) browse the collection using mobile devices (particularly the iPads).

Here is “sneak peak” of what it’ll be:

…now back to the final “tweaks” -before we officially switch to the new URL
.

For the LSTA MiniGrant project that we’re wrapping up this month, one of my tasks was (is) to customize/develop a “Manakin Theme” for DSpace … based on three years of web usage statistics, we knew that this (newspaper) collection has been heavily used with the “browse by date” option.  We also looked at some current newspaper projects online, and we found that a “calendar” option for browsing is quite popular -e.g. The Chronicling America.

Additionally, we were very happy with the results of OpenZoom for displaying dynamic (flash-based) images; but, we also discovered that it was very “tricky” to fully implement/integrate this viewer in DSpace …so this is where we decided to take advantage of the power and/or usefulness of the “publication date” metadata field.

The picture below can help illustrate what we wanted to accomplish:

[View full-size image]

• Generating a Calendar-View for DSpace, initially, we had to export a tab-delimited file with two fields (publication_date and handle); then I wrote a PHP script that reads the “txt” file and through the use of a couple of arrays, it generates two drop-down menus (decade and years); these are the options that end-users can use to select/display a 12 months calendar.  I have to say that the calendar may not -necessarily- look like a regular one, but that’s in part because of the 32 bit platform issue -this project goes way back than December 12, 1901  :-(  oh well, this could be a good topic for a later post.  Anyway, in early September I plan to document/share the script I used -in case someone else finds it useful!
• Embedding OpenZoom files in DSpace, since the OpenZoom files resides out of DSpace, we thought that a good way to “relate/match” a given record in DSpace with its corresponding jp2 files in OpenZoom was to use (again) the “publication_date” metadata field.  Under the OpenZoom directory, we created a folder and named it with a publication date (e.g. YYYY-MM-DD) and the trick was to send that date from DSpace to OpenZoom as a variable in a URL (e.g. ?pub_date=1991-04-30); once we had an actual date, then a PHP script reads the (jp2) files in that folder and generates a drop down menu which allows users to navigate/view the pages of that particular issue.

Ok, …guess that’s enough for now, the next (maybe last) work for this project is to finalize a prototype for displaying DjVu files for the iPads :-)  …hasta muy pronto!

It’s been almost a week already since I first got an iPad …why now? well, with the all the attention that “mobile” services are getting in the media, our Digital Initiatives team proposed/decided to start “prototyping” a mobile version -of the The Miami Student Newspaper Archive in DSpace- for the iPads.  This “apple” device seems to a very good candidate for digital collections; its display (1024×768) makes it an ideal choice for displaying medium-large images such as newspaper pages, maps, manuscripts, etc.

Today’s test looked like this:

The actual work/test will involve a couple processes: a) decide which format to use, current choices are PDF or DjVu; b) export/create an XML file from DSpace with at least four fields: title, publication date, handle, and a URL for the actual file to display; c) modify a calendar page -for browsing- with the appropriate CSS and JavaScript files for the iPad; and hopefully that’s it!

For the regular website of this LSTA-funded project, we’re using an open source tool (OpenZoom) which does an excellent job in displaying large newspaper images with zooming and full-screen options –very cool! …but it’s based on flash, so that’s why we have to use a different format for the iPad test.  And speaking of flash issues on the web, this afternoon I came across this presentation “HTML5’s Multimedia Future” which was part of the OSCON 2010 – O’Reilly Conference in Portland, Oregon.  It seems like HTML5 will make a big difference on the use of multimedia on the web.

On a separate -but related note- I think the iPad is just what I needed to stop printing articles; as happy as I was with the double-sided option on my printer, this new device seems to be a better choice, especially if I try to go green …we’ll see how it goes!