neblogska's blog

I've decided to start blogging here again.  I'll be honest, I'm a little vain; I like to see what I've written get wide exposure.  Hopefully I'll write something worth a wider audience than I have on my current work blog. 

I used Creighton's local instance of WordPress to keep notes on creating a Fedora Commons 3.0 test server.  I will keep those notes here, and write about some other projects, too.

In simple, here are my professional growth projects from now (August 30 2008) to the end of the calendar year (December 31 2008):

   1. Get a copy of Fedora Commons 3.0 and the Fez front-end working.

   2. Attempt to move content from EMC's Documentum (ApplicationXtender 5.4 to be specific) to Fedora Commons 3.0.

   3. See what makes better sense:

      - Using an enterprise search tool like the Google Enterprise Search Appliance for federated search, too.

      - Using a federated search tool like DeepWeb for enterprise search, too.

...is amateur software development.  No sooner to I promise my family NOT to bring my laptop home from work than I do what I've done many times: set up an account on my wife's eMac.  I've tricked out the machine with a tool chain, but I've discovered a few problems:

   1) I installed the DarwinPorts and MacPorts port management tools, only to discover they work with Xcode 3.0.  So...

   2) I downloaded Xcode 3.0 (1.1 GB) only to find it doesn't run on OS X 10.4.11, only 10.5.  Remedy?

   3) Download Xcode 2.5 (997 MB), remove Xcode 3.0, install Xcode 2.5, and discover that a) DarwinPorts and MacPorts don't appear to run with an earlier version of Xcode.  What's more installing the smallest piece of ported software wants a command-line version of Subversion installed.  So...

   4) I find and download a Universal binary for Subversion, only to find it won't install.  I have to find an earlier version.

The nightly reports are cleaned out and a missing icon is now in the catalog, on both the test and production servers.  I also checked free space on both servers (the font makes spacing funny):

Production:   Filesystem   kbytes   used   avail   capacity  Mounted on
/dev/md/dsk/d0   60514909   15348091   44561669   26%   /
/dev/md/dsk/d1   5953134   4946839   946764   84%   /var
/dev/md/dsk/d80   23051496   8571729   14249253   38%   /u01
/dev/md/dsk/d83   82624470   45383537   36414689   56%   /s01
/dev/md/dsk/d84   25827239   4699396   20869571   19%   /d01

I'm still seeing occasional core dumps, which I've reported to SDCC.  The diagnosis is "yeah, that happens sometimes...the fix is to install the latest patch cluster."  I've noticed no performance impact because of this error, and there have been no service interruptions observed or reported by the library staff.  We (the library staff and I) have jointly decided to put off installing PC5d or PC5e (whichever is latest); we will use PC5c until some other error comes along that requires the patch to fix, or we are ready to install Symphony 3.2. 

We hope to install Symphony 3.2 on our test server in July 2008.

I just completed regular maintenance on the production and test servers.  This consists of making sure regularly scheduled reports ran without error last night, and erasing them after inspection.

I am continuing manual user record loads, too.  There have been no complaints about latency, so I'll wait another day or two before loading outstanding user records.  I am still looking for unexpected, large (3 MB) user load files transferred from other campus systems.  If this error occurs, it means something has gone wrong with the nightly export process, an error my colleagues are very interested in.  I don't want to load a file this large, as loading wipes out manual changes to records in the library system.

I still need to fix a shell script that automatically backs up the user files in the library system before new records are loaded.  I call this taking a "snapshot" of the user records before new records are loaded.

The latest changes came up on the online catalog, some of them, anyway.  The bolded text near the search tips looks fine, but the "Help" link does not appear in the toolbar on the production server.  It looks okay on the test server, though.  So I have to find the difference between the two.

I've go to do this quickly.  Friday night I'll take the production server down and tar up the working software, and also do an Oracle dump of the database contents.  This is a read-only operation, so there is (theoretically) no danger of changes to the production system while its down or after I restart it.  Once the export is done, I'll restart the production server, and move the Oracle dump file and tar archives to the test server at my leisure.

I posted the latest, custom page files to the production server.  I sent a note to the internal, iLink discussion list that the changes will appear Tuesday morning.  I backed up the current "custom" folder so I can put it back if anything goes wrong.

I continued work on iLink this morning.  I started by comparing the delivered pages to the customized pages I copied from the test server.  I discovered that the delivered pages on the test and production boxes are slightly different.  I installed the same patch on both machines, but the patch files were downloaded and installed at different times.  So I can't be sure that a given patch is stable!

I deleted "custom" pages where there were few or no differences with the "delivered" pages.  Then I backed up the existing custom pages on the test server, and tried my new batch of customized files.  Two missing features returned straight away:  the link to our local help pages, and the "revise search filters" option.

I added some bold text around the search hints.

I found the "X-Ref" icon was missing.  I asked a colleage to make a new icon for me, to match the others on our customized catalog.

Rather than maintain two blogs, I am moving the first and only post from this site to https://people.creighton.edu/wordpress/mja30807/.

My blog at EDUCAUSE is now closed.  Please see my "work" site for new content.

My topic is my work on our online catalog, and my attempts to make it more useful and interesting for the library public at

Creighton

University . 

Creighton

University uses the Hyperion & Unicorn products from SirsiDynix, Inc.  We are on version 2003.1.4.3.

Like my "official" blog, hosted by my employer, I want my writing here to be helpful to and useful for others.  I also want to bring credit to myself and the University.  To this end I publicly commit myself to

Creighton

University 's acceptable use policy and the EDUCAUSE acceptable use policy.  I value my personal and professional relationship with our vendor, SirsiDynix, and the many libraries using SirsiDynix products; I won't knowingly post anything here that would harm or hinder those relationships.  Public information is welcome here, as is informed, respectful speculation based on public information.  Conversely, the appearance of potentially proprietary information and/or inflammatory or offensive comments is not welcome or permitted here.

The impetus for this web log is my hope for (not yet the certainty of) an upgrade to a newer version of our public access catalog of library resources - in this matter I defer to my managers.  If the upgrade comes to pass that is great; if not, I will work with the resources available to me.

To begin, it is helpful to take an inventory of our own catalog, and the associated electronic and web-based resources of Creighton's libraries.  Additionally, it is also helpful to have some sense of what is possible, based on the work other libraries are doing.  I will use this site to catalog and document various WebCat, iLink, iBistro and EPS sites across the Unicorn customer base.  My main source of information is the Internet itself, through Google.  If a library would like to be added to any list I create, I'd be glad to do so.  However, if a library is on this list and wants to be removed, I am also glad to do that.  Either request must come from an authorized representative of a given library.  You don't have to give a reason, either - it is enough to say "Put us on the list" and I'll add you or "Take us off the list" and I'll remove your entry, it’s that simple.

Two final points of information:  First, I am going to focus on how the catalog is customized, not necessarily the preceding or surrounding web pages on a library's web site.  Second, I am not quite sure how to organize this list of catalogs.  It is not desirable, necessary or possible to catalog each and ever instance of WebCat, iLink, iBistro or EPS.  My particular interest is in academic libraries, so I'll focus there, but that doesn't mean the work of corporate, government, laboratory, military, national, public, special, school and other libraries (have I missed any?) is not also applicable and interesting.  I will include links to the catalogs of the other-than-academic-libraries, too.