University of Minnesota
University Relations
Our Brand: How to Convey It

Our Brand: How to Convey It.

Recently in Tabbed Search Category

Examining eCommunications

| No Comments

Join your colleagues as we explore topics specifically related to online communications.

Here's the scoop:

Who: Anyone can attend (University students/staff/faculty or the general public)
What: Examining eCommunications webinars. Each 45-minute presentation will focus specifically on online communication as it relates to the University.
Where: On the web
When: August 2010 - June 2011; 10 a.m. CDT
Why: Learn about new best practices, glean tips and tricks, and ask questions of your colleagues all from the comfort of your own chair.
How: Attendance is FREE with registration. Want to receive reminders of upcoming webinars? Send an email to with subject line "Webinar Reminder" and we'll be sure to give you ample notice.

Here's what's on tap:

August 24, 2010
Images library redesign by Jeremy Casper, University Relations

September 28, 2010
Content Store by Pete Wiringa, University Relations

October 26, 2010
Using UMContent by Tim Roman, College of Continuing Education

November 30, 2010
Video by Liz Giorgi, University Relations

January 25, 2011 **New speaker**
Usability by David Rosen, Usability Services Manager, Office of Information Technology

February 22, 2011
Content creation by Office of Information Technology

March 22. 2011
Analytics by Pete Wiringa & Jeremy Casper, University Relations

April 26, 2011
Social networking by Jennie Lijewski, University Relations

May 24, 2011
Working with web templates by Kathy Jensen, University Relations

June 28, 2011
Tabbed search and keyword by Jeremy Casper & Pete Wiringa, University Relations

July 26, 2011
Events calendar by Jeremy Casper, University Relations

P.S. The series was originally titled "The Super Karate Monkey Death Car Webinar Series: A University Relations Tribute to the Writers of Newsradio." Which title do you prefer?

Tabbed search redux

| No Comments

We've released a new version of the tabbed search application with a number of changes, including a rewritten document describing tabbed search and all its options.

We're continuing to look at tweaks and incremental changes to tabbed search and other search pages to improve the overall search experience for our visitors. Please feel free to write us at with any suggestions.

Using the Google Search Appliance to find bad links

| No Comments

The University of Minnesota offers the Google Search Appliance (GSA) to units at the University to help customize the search experience for their website users. Most groups will setup a front end and collection and never really look at anything else within the system until the next redesign of their pages. But, if you dig a little bit deeper, the GSA provides some great features that are underutilized. One of these features is the ability to find bad or old URLs that you would like to get updated. GSA can see they still exist when those old URLs show up in the Crawl Diagnostics reports.

For instance, after the UofM home page was redesigned in 2008, some of the URLs changed. A page that use to be at is now located at University Relations has put in the proper redirects so end users will get to the proper place, but if we look in the GSA, the page still comes up in the Crawl Diagnostics page. The fact that the 01_about.php page is showing in this report indicates that some site somewhere has this old link still in use.

So, how can we find the sites that have this old link?

The simple solution is to go to and do a search for link: You can see the results for the search here. From the search results, we can see the pages that have the old link. One of the first links is to the Educational Psychology site that has the link in it's navigation. We could send a note to the webmasters informing them of this wrong link.

In the case above, we were looking at a site with a known redirect. But, in another case we had a link to Now this site does not have a redirect as it was never a valid page, so in this case it is a mis-typed URL. Doing a search on it reveals a single page that has the incorrect URL. Since the site is owned by our group, we can update the link ourselves.

This trick has come in handy many times. In one instance we not only found a page with the bad link, the page itself had a lot of incorrect information that had changed over time. So instead of having a page with a lot of outdated information, we hopefully will have a page that serves our users well.

Recent changes to U search pages

| No Comments

You might have noticed that, the TC people search pages, and some of the Google Search Appliance results pages have recently been updated to use the new template headers and footers.

You might be interested in the Google Search Appliance (GSA) changes. GSA uses XSLT for the results rendering. Out of the box their XSLT (as of GSAv6.2) is still behind the times. Hint: you'll still see <br> tags, as in not resembling XHTML. This makes validation pretty ugly.

We took the XSLT from the gsa-xhtml-stylesheet project, which produces much, MUCH cleaner code. Note that gsa-xhtml-stylesheet notes compatibility with an older GSA version, so something may well be broken or additional functionality found in v6.2 not already in the sheet.

We updated it to use the standard UMN template header and footer, but generally left the results styling alone. We make no claim that the XSLT code is pretty, but it works, has browser tested OK in common browsers (including IE6), and only had eight validation errors last time we looked, as opposed to hundreds.

Want a copy? Shoot us a note at

I have had several questions recently about search engine optimization and how to bump up a website in the rankings. I came across this article and thought it presented a very good point of view. I hope you enjoy it.

6 Ways to Optimize Your SEO for Misspellings - And Why It Pays to be a Bad Speller

Tabbed search available as a service (beta)

| No Comments

The tabbed search application is now available as a beta service offering. The code isn't ready for distribution and other projects have sidelined some cleanup and documentation, but you can give it a try and we welcome all feedback. Three "front ends" are currently in use, and you can search using these front ends from the following sites:

Want to give it a try? We'll setup a front end for you. For a new front end, or if you have comments, suggestions, or other feedback, please write to A few quick notes:

  • Yes, we know about the character encoding issue.
  • The "U of M" tab must always be present for the service, but is last by default.
  • Only the standard Web templates are offered for the service; no customizations.

A front end is a collection search result sets (tabs), and breadcrumbs. Each tab utilizes the Google Search Appliance (GSA) and may be configured independently. Configuration options include:

  • Site (i.e. tc_oit_1help)
  • Collection (i.e. default_collection)
  • Additional query string, not visible by users (i.e.
  • Suppression of keymatches
  • Suppression of cache links
  • Support for additional GET parameters against the GSA (i.e. for meta tag searching)

Tabbed search is offered only as a service during the the beta, but the code will be available for download later. The code requires PHP5. Not every unit has PHP5 available, nor is it common to central, shared hosting, which is why the service is being offered. The service will also ease code updates; when bug fixes need to be deployed, template changes made, or new functionality added, this can be done centrally.

Tabbed search results

| No Comments

With priorities changing, we've finally had a chance to get back to tabbed search results and have been making excellent progress. We're currently dealing with some design issues, quashing bugs and rendering inconsistencies, looking at accessibility issues, and engaging OIT as we look for a reliable hosting solution that includes PHP5.

Needless to say, IE6 (and 7) have been problematic and are consuming far too much of our time but, we do need to support those offerings. At this point, running in Firefox 3.5.x, things are looking very good.

The quest for PHP5 may be a roadblock. We'd intended that tabbed search be flexible and extensible so that other groups could add search implementations to go against sources other than the Google Search Appliance and be able to render those results appropriately. Leveraging the better support for XML and classes/object-oriented programming in PHP5 is what makes this not only possible but, relatively easy. Backing down to PHP4 would be a detriment to the project, and require the mangling of good, working code.

Ongoing issues aside, we don't have it running on a publicly available site but, here's a screen shot of where the front-end is at the moment.
Tabbed search example

Tabbed search status

| No Comments

Since we've been asked about the status of the tabbed search results implementation, here's an update.

At this point, we have the basic code designed, but not implemented, and some mockups that need touching up. The code will be strictly PHP4, for better alignment with central, shared hosting solutions, such as OIA's shared hosting and Internet Services' web hotel. The code breaks down into several parts and is based heavily on Cornell's search results scripts (which don't seem to be available on the web anymore).

Rendering. The rendering layer renders the tabs and content items with the tabs. A standard format for output is provided but, additional code can be added by units to render differently, e.g., if the results for one tab come from a database with additional fields that would be useful to display in the results. The results will rendered in semantic HTML and progressive enhancement used to pretty up their display (jQuery).

Pagination. We are still looking at the best way to paginate results within tabs. While an AJAX solution is possible, this wouldn't degrade well and we do need to account for users with JavaScript disabled, so we're looking at options with progressive enhancement in mind.

Results. The search functionality populates results objects, which are standardized for key fields but may include other fields. Title, description and URL will be key.

Search. While we're focusing on the Google Search Appliance as the primary search engine, our intent is to allow easy integration of additional search engines and databases. Search executes the search but, result sets may come from a different data source, and code needs to be written to query those data sources. A standard set of code for use against the U's Google Search Appliance (GSA) will be included and can be configured to obtain results from other sources, contingent on those sources taking input parameters in a manner similar to the GSA, and providing output that is substantially similar (if not identical) to that provided by the GSA.

Events Calendar, Search, and UMContent update

| No Comments

So I have the privilege of being the first person to post on the eCommunication Blog. In thinking about which topic to talk about, I decided that providing quick updates on a couple of topics would be the best use of this post. So, here are some updates on some of the projects that I am working on.

Events calendar:

Thank you to everyone who responded to University Relations' request for feedback on the current events calendar and suggestions for improvement. We have taken those responses and have started to create a draft proposal to submit to the OIT so that the project can be included in their planning process. In addition to creating a draft proposal, we are working on drafting functional requirements to be used later in the process. We will be meeting with various groups to get input as the process moves along and will also be creating an advisory committee to help with feedback. The major next step will be to meet with OIT so that we can determine what an appropriate timeline will be. If you are interested in meeting with UR regarding this project, please feel free to e-mail me at


This is one project that has been delayed longer than I care for. The hope was to have the updated tabbed search results out around the same time as the templates. Due to other projects, this has been delayed. I am currently shooting for the end of March, but it will most likely be in April. We are working to get this setup so those users on OIT hosting (whether through OIA Shared or IS Group Web Hotel) will be able to just set a configuration file and be able to use it with minimal programming knowledge. To answer a question that some people have asked, yes it is possible to pull in search results other than the Google Search Appliance. Users wishing to utilize this will probably need some programming knowledge to integrate it.


Most of my time has been spent trying to get the rest of UR sites out of FileNET and into UMContent. After this project is over, more of my time will get shifted to search and events calendar. We have been working with OIT to help identify areas where performance could be improved as they work with Fishbowl Solutions. We have determined that one of the services that the system is using (SS_GET_SEARCH_RESULTS) is running full-text searching and therefore is very slow to return results. To get around this, we have switched our dynamic lists to use the GET_SEARCH_RESULTS service. This has had the effect of reducing our page load times from around 27 seconds to 2 seconds. When OIT and Fishbowl determine how to get SS_GET_SEARCH_RESULTS optimized, this will hopefully be able to speed up the contributor experience as this service is used heavily on the contribution side.

We are planning to get the new U Templates into UMContent, however we are waiting for Site Studio 10gR4 to fully implement the templates. With the 10gR4 functionality, we should be able to provide templates that users will just be able to use out of the box. This is an improvement over the previous versions where customizing the templates required groups to copy the layouts and then modify them themselves. Site Studio 10gR4 has been installed on the DEV environment, so we have had the opportunity to start to look at the architecture, but until some of my other projects are out of the way, this is a lower priority.

Other Notes:

As some people may know, my appointment is an 80/20 split with the 20% of my time reporting to OIT's Office of Communications and Advancement. This has been beneficial as there is overlap in the projects that both OCA and UR are working on. Hopefully in the next few weeks, you will see a revised OIT home page ( that uses the new U Template and also brings in graphical elements from the U home page. Hopefully this new page will provide an improved user experience.


| No Comments

Discussion about search from a site developer viewpoint, including standards, best practices, templates, and metrics.