K-State Libraries: Digital Initiatives

Docs to Blogs

There's a new, simple way to get documents created by a word processing application (like Microsoft Word) posted to one of our blogs on TypePad, without the usual formatting nightmares or cut and paste headaches.  All you need is a Google account and access to Google Docs, and the steps outlined below.

1. Login to Google Docs
2. If your document is saved elsewhere, upload it using the "Upload" button near the top left of the Docs screen, just under the Google logo.   You can upload documents in .txt, .doc, .html, .rtf, .odt and .sxw formats.  It will open for editing after it is uploaded.
3. If your document is already in Docs, just open it for editing.
4. Find the "Publish" tab in the upper right corner and click it.
5. On this screen, you have two publishing options.  You are interested in the 2nd choice - publishing to a blog. You'll need to set up options the first time you do this, so click the setup link ("set your blog site settings")
6. On this page, you'll pick some non-intuitive things, so follow along (click the tiny thumbnail for a screenshot):
   1. Click the radio button for "My own server/Custom"
   2. Select "Blogger API" from the API dropdown
   3. Copy and paste this URL: http://www.typepad.com/t/api/xmlrpc.php  into the URL box
   4. Enter your TypePad userid and password in the boxes, and the name of the blog to which you're posting your document.  Make sure you enter the full name exactly as it appears on the blog, for example, "K-State Libraries: Talking in the Library", not just "Talking".
      
   5. Click OK.
7. The settings box will disappear, click Post to Blog to post your document.
8. Click OK in the confirmation box that appears.
9. Go look at your blog and see how cool it is!

If you're a multiple-blog author, just edit the settings to change to a different blog before you publish.  Another neat thing you can do is edit a document you've already posted to blog from inside Google Docs. Then, when you go to the Publish tab, one of your options will be to re-publish the previous post with the new document - handy for fixing typos or updating information!  The only flaw I've found so far is that it seems reluctant to assign a title to your post based on your document's name - you may need to login to Typepad and assign a title to your post, but that's a small price to pay for the convenience of directly publishing documents you've already created.

Who or what is big oscar, and should I be afraid?

Big oscar is a subdomain at AOL, specifically:

http://big.oscar.aol.com/

We make use of a small service offered by AOL, namely, the ability to toggle images for our chat application based on the status of our AOL login. When a page loads that uses this functionality, such as:

http://www.lib.k-state.edu/reference/vref/

or any of our catalog search pages where one finds a Meebo chat widget (or, more accurately, a picture of one), before the image of the chat widget loads, a little query goes out to big oscar, who either says, yes, the library account is logged in, in which case the "click to chat with a librarian" image loads, or no, which loads the "no one available" image. This is completely benign, and in no way poses a security risk.

Why bring this up? If a person uses Internet Explorer and cranks their security settings way up high, IE will flash a warning when this query gets sent to big oscar. Personally, I don't want to use a browser that's such a namby pamby about this kind of stuff (which is why I use Firefox, which applies a lot of common sense security fixes without being so annoying), but some people have been burned or otherwise convinced they need ultra-high security on their machine. Not much we can do there other than tell them to allow it.

Slight domain name change

A while back, we edited the settings of our Webserver to redirect all variants of our main domain to www.lib.ksu.edu. As of yesterday, we will now be using www.lib.k-state.edu as our canonical domain. As with the last change, all variants (ksu or k-state, with or without the www) will continue to work in perpetuity, so no documents need to be edited, nor does this change really impact much of anything other than the address in the browser address bar.

This was done to bring us into line with a proposal I wrote for the campus, suggesting that however one may feel about ksu contra k-state, it is time for our Webservers to declare a unified canonical domain so that our content is properly indexed and ranked by major search engines. Having access to a Google Search Appliance has opened my eyes to the chaos that serving inconsistently via two domains can create for a search engine, which can only hurt K-State and its findability. I proposed a simple change to a Webserver's configuration that will allow all of the variants to work indefinitely. The proposal was reviewed and deemed worthy by both the Faculty Senate Committee on Technology and Associate Vice Provost James Lyall. Adoption is entirely voluntary, of course, but it can only help a site's search engine rankings and findability.

Finding book reviews in Web of Science

This may be an odd place to put such documentation, but with a search appliance in the house, what does it matter.

I'm tired of having to relearn how to find book reviews in Web of Science, so I thought I'd finally write down how to do this. I originally wrote this post in July 2007, but then the nice people at Thomson ISI revamped the WoS interface, making it even harder to use their tool to find book reviews in my humble opinion. Here are the revised instructions:

• use the Advanced Search (ugh, now you have to know how to type a complex query)
• put some keywords from the title of the reviewed title in the search box; you must put ti= in front of each one and connect them with 'and' (e.g.- ti=prague and ti=territories), or put part of the title in quotes after ti, e.g.- ti="tour de france a cultural" (note: this need not be the complete title, but it must be an exact string).
• at the bottom of the search form, set the restriction in the righthand box to Book Review

That's it. If you're lucky, you'll find more reviews here than you will in such flimsy and expensive titles as Book Review Digest and Book Review Index, which both largely ignore reviews in academic journals and mainly list only those in the popular press.

This is not the only way to do this in the new Web of Science interface, but I do believe it is the most direct.

One domain: www.lib.ksu.edu

For some time, I have been mulling over the fact that our site does not respond to a www-less URL, i.e.- if you enter lib.ksu.edu in a browser, you get a site not found error. A somewhat related issue is the unfortunate situation that arose when K-State made the ill-fated choice to serve content over two domains: ksu.edu and k-state.edu. What that can lead to is a situation where a K-State site is indexed twice via both URLs, based on the entry points a bot finds. I've documented such occurrences in public Google previously, and can easily demonstrate that the problem persists. For example the following searches in Google yield wildly variant result sets:

• kansas site:k-state.edu - 48,200 hits
• kansas site:ksu.edu - 173,000

Since I have access to a Google Search Appliance here, I can see the chaos that this schizophrenic naming convention creates in an indexing scheme. I end up having to set a bunch of rules based on the preferred URL of a given subdomain (some of which, annoyingly, respond only to one or the other convention).

That was a long prelude to some documentation on changes to our production Webserver. Lawrence, our man in ETS, made changes today that will now enable our server to respond to just about any halfway valid attempt to get our domain right, including:

• www.lib.ksu.edu
• www.lib.k-state.edu
• lib.ksu.edu
• lib.k-state.edu

Moreover, all domains now redirect silently to one authoritative domain: www.lib.ksu.edu. I know that that may look like some sort of revanchistic statement, but, really, there are some good, rational reasons why I chose ksu over k-state in this instance. Our server, as noted, will dutifully respond to any form, so in the end it matters little to the user, so I've made a choice that works well for our department.

New Books, Just for You

You know that new stuff is constantly coming into our Libraries. But have you ever wanted to know exactly what new stuff has arrived? Like every week? Now you can!

Special Pages

On the left side of every wiki page is a box labeled "toolbox".  In that box, you'll find a link to Special Pages.  Special Pages are default pages created by the wiki software to hold all kinds of information about the wiki itself.  You can see statistics on various pages, check out the "most viewed" and "most linked" pages, newest pages, oldest pages, user lists, or even click a link to be sent to a random page. 

Many of these pages will become more useful as we put more information into the wiki. One of the special pages that might be immediately useful is the Orphaned Pages page.  If you've created a new page, but haven't linked it to any existing pages, it is considered an "orphan".  If you've lost track of a page, you might find it here.

Wiki preferences

Once you're logged in to the wiki, there are a variety of preferences you can set.  To access or change your preferences, click the "my preferences" link that appears near your userid in the top right corner of any wiki page when you're logged in.

The Preferences page has many tabs, but you'll probably only care about the first two.  On the first tab, you can change your password, and set your preferences for receiving email about the pages you create or edit.  The next tab, Skin, allows you to choose from several color and layout schemes for the wiki.  These schemes only affect your login, so feel free to choose whatever you like.  The rest of the tabs are probably best to just leave with the defaults, but poke around on them if you're really into wikis or just curious.

Wiki Logins and Passwords

Your wiki userid is your first initial, a period, and your last name - d.ekart, d.ireton, j.heikkilafurrey (sorry long-last-name people!).  The login link is in the top right corner of each wiki page.  If you're logged in, you will see your userid there, and a logout link. 

If you forget your password, click the login link, enter your userid, and then the "email password" button.  The wiki will email you a temporary password.  (The email will probably appear to come from Harish, but it's really just an automatic thing the wiki does, so don't call Harish and ask him about the email, because he didn't send it to you.) You can use the temporary password to login once, but you should change it immediately to something more secure.

Adding a location limit to Voyager

This is done in SysAdmin - System - Location Limit Groups. I added the Leisure Collection by clicking on new, giving it a name, a display text (taken from the display name seen in the catalog, of course), and then added the location code on the second tab. That's it. Just reload a catalog session and there it is.

GroupWise email auto-reply in easy steps

Having proposed that we should all really get in the habit of using this feature of GroupWise, I thought I'd write a handy-dandy tutorial.

1. Go to the Web version of GroupWise (http://gw.ksu.edu). Yes, Dorothy, it has one of those.
2. Log in, using the same old, same old.
3. Click on the Options button (), top middle of the screen.
4. In the resulting popup window, click on Rules.
5. Select Vacation from the dropdown list, and click on Create.
6. Give it a name, such as Out of Office, and enter a subject line and message for the auto-reply email.
7. Click Save, and you will see this rule added to your list of rules, with a check in the box indicating that it is active.

That's it. Remember to deactivate it (follow steps 1-3 above and just clear the checkbox and click Save) when you return, or you'll get all those messages from your colleagues telling you that your email account thinks you're still out.

Oddly, the installed GroupWise client doesn't offer this as a default rule type, but once you've created it in the Web client, you can then deactivate it via the GroupWise client. Frankly, I find that Rules part of the client so obtuse that I use the Web client for the whole shebang.

TypePad sidebar documentation

Made the following changes today to TypePad:

• Created a Site Search typelist, which searches all library Webservers. Currently only found on the DID blog (to test it), but I will add it to all of the public blogs as the day goes on.
• Created a Search Search Cuts typelist, which searches the Short Cuts content on both ada and TypePad. Copied the XSLT stylesheet k-state_libraries and created a new front end called short_cuts using this stylesheet. The only difference is in the displayed header--I removed the link to the general site search and renamed the search results page.
• Edited the sidebar content for Short Cuts. It belongs to the category of blogs that cannot be controlled via a central template.
• For blogs geared toward an external audience, I put the following content in the left sidebar: Library Links, K-State Links, Site Search. The blogs in this category are The Dean's Seen, New Resources, Service Status, Talking in the Library, and News & Events.

Additional RSS feeds: DVDs, CDs, etc

We had several requests for additional rss feeds from our catalog - one for DVDs, one for CDs and an another one for feeds by location. The revised version [1.5 - why not 2.0?], under test, generates seperate feeds for DVDs, CDs and one for "reference books on Hale 2nd floor" [location feed]. This version also fixes some issues that arose because of inconsistencies in the call numbers[documented in our local test wiki]. This should go into production early next week if all goes well.

The 3 new feeds are at: http://catalog.lib.ksu.edu/ksul/rss/test/newfeeds.html
Note: The links will point to the actual xml files and you need to point your rss readers to the individual locations.

Title lookup from any webpage

ISBN lookup has certain limitations in that that it works only with isbns and that too isbns that appear in the urls[amazon]. We can definitely customize it and extend it further to lookup isbns anywhere on the webpage. But after getting some input from Adam and Sara, I agree that it would be nice to have a title lookup feature rather than just an isbn lookup.

For starters, we now have a title lookup feature. This bookmarklet does a couple of things - 1) You can select text on any web page and do a title lookup on our catalog. (2) If you are on Amazon.com, then this script would extract title from the webpage automatically and do a title lookup - so you don't have to select text explicitly when on Amazon. The second feature works only for Amazon though, at the moment. To install just right click on the link below and save it to your bookmarks/favorites.
Title lookup

In version2.0, we'll try and integrate with xisbn service from OCLC. What this does is gives us an ability to go beyond the given isbn and gather information from related isbns [eg: given an isbn for a paperback, we'll get isbn for hardcover,etc]. More on this later..

A-Z List Comparison

I have completed a fairly detailed comparison of the A-Z lists supplied by Serials Solutions and SFX. The full document is linked for you to download and devour.

As of mid-September, Serials Solutions lists nearly 23,000 titles, while SFX lists slightly more than 20,000 titles. So why the big difference?
Serials Solutions is fairly free in its definition of "electronic serial." Of the nearly 3,000 items it lists that SFX does not, at least 1,000 are annual conference and symposium proceedings from IEEE and ACM. Serials Solutions is listing these items by their conference title rather than by their series title. They are difficult to locate in the list even if a user feels like browsing.
SFX directly accesses these items, either through a user's click on the K-State Get It button or through their use of the Citation Linker. Each item in these well-established series has an ISSN, so SFX can usually generate a valid OpenURL to find what is requested.

There are certainly other possible explanations for the nearly 3,000 title difference between the two A-Z lists, but none that accounts for so many titles and fits such a large pattern.

Download ComparisonAtoZlistsSerSolSFX.xls

Voyager: print and email location names

There was a bug in Unicode Voyager that made the location names in email and print output from the OPAC disappear. Normally, this appears in the emailh.cfg or saveh.cfg file as:

Location:
     1000

Where 1000 is shorthand for Voyager to pick up the location name from a table. The only fix for this in Unicode was to replace this with:

Location:
     852 XX ab

so that at least the crytpic location code would appear in the output. They fixed this in v5.

OPAC changes

I made a host of changes to the catalog today, and thought I should record them here.

CSS to make printable version

The following CSS removes the sidebars from the print output, widens the center column, and creates a small buffer to allow for weird spacing in certain blogs:

@media print {.module {display:none}}
@media print {.layout-three-column #beta {margin-left:5px; margin-right:5px}}

The first bit uses the parent class of all sidebar content (module) to hide them, then takes the center column, beta, and gives it a 5px buffer. TP uses greek letters to signify columns, from right to left: alpha (left sidebar), beta (main content), gamma (right sidebar), delta (far-right sidebar in four-column designs).

Incidentally, never knew one could use a parent class like module to declare display:none and hide a bunch of stuff, but was very glad to make this serendipitous discovery.

One last note about printing: this is a rare instance where I would recommend using IE rather than FF. FF renders posts almost as images (treats the div space as an unsplittable whole), so will actually dump the rest of a post that goes longer than a page. Also creates horrid spacing between longish posts. IE renders them as continuous HTML, ignoring the div. I'm sure it was just an oversight on Microsoft's part, not intentionally smart design!