MarcEdit 5.1 Update

By reeset / On / In LibLime, MarcEdit

Getting this last update out has taken a little more time than I would have liked, but I really wanted to think some of the issues that this update raised through so that the update process would be seamless.  Realistically, were I versioning MarcEdit in any realistic versioning process, this would likely be at the very least, a new point release.  However, I’ve already planned out my 5.5 release, and this request, while major, fell into this gray area – so I decided to keep this into the 5.1 branch.  Anyway, March 2nd is a good day to officially make this release.  March 2nd is my 32nd birthday, and this version of MarcEdit will be present to the MarcEdit user community.  Cheers.

Major Changes:

  1. In the continued work towards helping enterprise users, I’ve finally finished the installer update process that began with the last update.  In the previous update, MarcEdit’s installer was modified (as was MarcEdit itself) to make it more user aware.  What do I mean here – the program was changed so that users running MarcEdit no longer needed to be administrators when running the program, but rather, all configuration and mutable files were moved into the users Application Data directories.  This had a number of unintended benefits like supporting multiple MarcEdit users on the same machines (using custom user profiles) and making it easier to copy configuration settings from one computer to another. 

    This update takes this one step further.  As Libraries continue to move to more sophisticated application management, I’ve been running into more users that have their software managed through a central IT source.  The IT groups manage the software by automating a process to do distributed installation.  In the past, MarcEdit’s installer really didn’t do this well.  Well, about 2 months ago, I was contacted by a large IT group on the west coast wanting to know if this would be possible.  This required finishing the migration from MarcEdit’s custom installer to the Microsoft Windows Installer – while at the same time, making sure that the program cleans up the previous install while still keeping the users previous settings throughout the upgrade.  After a lot of testing (over the last month, spanning multiple institutions and users in different development environments) – I feel like this is ready to go. 

    When you install the new msi installer, what will happen. 
    1.  MarcEdit will evaluate your current installation – if you have never installed MarcEdit, it simply installs the application
    2.  If a previous version of MarcEdit is present, the program will copy, in order: config. data in the User Application directory, data in the MarcEdit Program directory – and then silently uninstall the previous version of MarcEdit.  Once the previous version has been removed, the installer when then install the new version of the Application. 
    3.  Part of the clean up process of the new install is to move the copied user data back into scope of the application.

    So what does this mean to you?  Well, if you are an individual user, who manages MarcEdit on your own machine, very little.  The one benefit that you will likely see from this migration is the eventual development of an automated updater.  The msi installer provides a number of very powerful and integrated functions that I should be able to leverage to potentially create an unmediated upgrade process for users.  For enterprise users however, this change will for the first time give your IT administrators the ability to install MarcEdit on multiple machines simply by using their enterprise software management system.  For colleges and universities that manage thousands of users, this should be a really big win.

  2. Z39.50 Changes:  This will be an incremental process, but for the first time, MarcEdit will allow users to query multiple user databases during a Z39.50 lookup.  This will allow users to query multiple Z39.50 targets to return data about a search.  This initial implementation allows multiple searching to be done in the Single Search mode.  In a planned future update, this will be extended to the batch search update, with rules regarding how to disambiguous duplicate records (for example, the ability to accept records from one target over another, etc.).  So how does this work.  Essentially, when you enter the single search, you select Select database and then select multiple items.  Up front, this is limited to 3 databases, but that limit will eventually be removed (especially as I get UI feedback).  When you select multiple database, the Single search screen changes to look like the below…

    Do you see the data in the red box?  This is how you can see what resources MarcEdit will be querying.  Also, see the green box.  You can see here that MarcEdit’s Z39.50 results list has changed slightly to let users see what institution each record is from.

  3. Help File is now local again:  Sadly, some topics are already out of date slightly (the Z39.50 info for example doesn’t represent the multiple querying functionality) – but this makes the help available both online and offline.  The online help will always be more up todate, but will be updated on each build.
  4. Youtube tutorials.  If you go to Youtube and look for marcedit, you will find a series of tutorials related to MarcEdit topics.  At present, you will find topics for:
       1.  Breaking your file
  5.    2.  Making your file
  6.    3.  Editing a MARC file
  7.    4.  Converting a files characterset
  8.    5.  Adding a new XML Function
  9.    6.  Updating a current XML Function
       7.  Using the Delimited Text Wizard
  10.    8.  Extracting a subset of records from a larger set
  11.    9.  Using the Z39.50 Client
  12. 10.  Harvesting OAI data into MARC
  13. 11.  Managing Plug-ins through the plug-in editor.
  14. OCLC Connexion Client Plugin has been updated (was needed do to updates in a few other components).  If you use the Connexion Client plugin, you will need to update this plugin once you update. 
  15. MarcEditor Editor update to better support UTF-8 data loading.  Essentially, in layman’s terms, here’s what has changed.  In previous versions of MarcEdit, loading UTF8 data into the MarcEditor would sometimes cause the process to load slowly.  The why this occurred had to do with the way that the specific windows component that I was using handled text.  I’ve updated MarcEdit so that this process has been changed (as has the component) making it so that data loaded into the MarcEditor now uses a new editing component and one that natively handles UTF8 data.  The lag time that users previously experienced should no longer be applicable.  In addition to this fix, MarcEdit’s MarcEditor’s memory footprint has been reduced.  Not drastically, but a bit.  One thing to remember when loading data into the MarcEditor.  There is roughly a 4-1 memory usage when loading bytes into a visual interface in Windows.  So, for example, load a 20 MB file, and Windows will allocate ~80 MB of memory to view the file.  Open a 120 MB file, and Windows will need to allocate ~480 MBs to render the file.  The new Editor is able to reclaim some of this memory on the high end, but this is in part, how visual interfaced work. 

    Also, note the 1/2 GB limit of data loading into the Editor, but that data of any size can continue to be edited in the editor if one makes use of the Preview mode.

  16. Yaz Update – Previous version of MarcEdit used Yaz 1.+ because it was small and fast.  I had need for some of the enhanced functionality, so I’ve updated the version of used Yaz to 3.+.
  17. ‡ Proof of concept Plug-in:  As noted in a recent post, while attending code4lib this week, the folks at Liblime demonstrated their new  ‡ platform.    For those that haven’t heard, the  ‡ platform is an attempt to create a large, shared, Open Data repository of bibliographic metadata.  What I find most interesting about LibLime’s effort has been the development of an open API to provide push/pull functionality into the database.  In theory, this allows library developers the ability to develop tools around the  ‡ platform.  The plug-in demonstrates how this works, as well as providing folks that want to work with the  ‡ platform a way to integrate their workflow with MarcEdit.  You can see the Youtube video talking about how it works, here:

Minor Changes:

  1. UI changes to the Z39.50 (to accommodate the changes in functionality)
  2. Extended regular expression support in the Replace function so that regular expressions can be run over multiple lines.
  3. Updated workflow – when converting data from UTF-8 to MARC8 using the MarcBreaker, the 9th byte in the leader isn’t set correctly.  This is partly because the previous workflow assumed moving the other direction.  Since this cause some problems with some loaders, its been corrected.
  4. Updated the Marc21XML xslt function to accommodate the following:
         a.  Up to 9 indicators (per UniMARC)
         b.  Ability for indicators to be mixed.  The current version assumes indicator order, the update allows indicators to appear in whatever order.
  5. Other minor changes

You can pick up the update at: MarcEdit_Setup.msi.

If you run into any problems, please give me a holler.


‡ MarcEdit Plug-in

By reeset / On / In code4lib09, LibLime, MarcEdit

So, one of the presentations at Code4Lib this year discussed one of the latest initiatives to come out of the LibLime company, ‡  ‡ is a repository of ~30 million MARC records released under an Open Data license. 

From my perspective, one of the things that I found most interesting about the ‡ platform is the support for developers.  ‡ provides a very simple to use push/pull data model over simple HTTP.  It’s something that I’ve been interested in taking a look at for a little while and after chatting with Joshua Ferraro after his session, decided to see how difficult it would be to actually work with this service.

So, I decided to create a plug-in.  Much in the same way MarcEdit has a Connexion batch editing plug-in, I created a ‡ helper plug-in.  I’ll make the plug-in available for download with the next version of MarcEdit through the Plugin Manager (so, likely when I get back from Code4Lib), but have posted the source code (in C#) and a short youtube video for folks wanting to see how it works.

While I’m not sure if ‡ will catch on as a service, I’m very impressed by LibLime’s effort to create a large, shared cataloging database that comes with a set of API to allow developers the ability to integrate directly with it.  Hopefully, OCLC will follow LibLime’s lead and eventually add push functionality integration for WorldCat – allowing membership and library developers an opportunity to develop their own cataloging interfaces to the resource.  Until then…

Youtube video (I think youtube has finished processing it – if note, check back):

C# Source code: