Terry’s Worklog

I’ve just updated the MarcEdit Beta Build. The issues corrected in this build were specific to MarcEdit’s Unicode functionality. Changes:
1) First, the MarcEditor and the MARCEngine has support added for entities…i.e., characters can be represented as W3C notation: 齊…Of course, this leads to a problem in terms of tracking the input stream since .NET requires the byte stream to be idential to the io stream. The program tracks this internally now and after doing some testing, it appears that its working now. If anyone notices something different, please let me know.
2) CJK corrections — the UTF-8 to MARC conversion needed a small correction. The translator was incorrectly inserting extra characters on certain occasions.
3) Mnemonic replacements — when creating UTF-8 to MARC-8 records, the mnemonics were not being replaced correctly every time. This primary affected the $ sign – which is represented as a mnemonic. In CJK cases, it would be represented occasionally as {lcur}dollar{rcur}. This has been corrected.
4) MARCXMLtoMARC — the checkbox allowing direct translation to MARC-8 has been enabled and tested to ensure that it is working.
5) MarcEditor — Edit subfield fuction. I’ve added new functionality into the Edit subfield utility. There is a new option called Move Subfield data. This allows users to do smart subfield swapping. It allows you to use 5 arguments. An example of its use:
Field: 300
Subfield: b
Field Data: [leave blank]
Replace with: 994|b|.||$t

The primary arguments are found in the Replace with options. The five arguments are:
1) Field: This is the field to swap data to
2) subfield: This is the subfield to swap the data to.
3) normalization: This is data that should be normalized out of the data to be swapped.
4) substitution: This allows you to provide additional data. An example would be: {data} this is added. {data} represents the data to be swapped — all data surrounding the {data} mnemonic will be inserted with the swapped data.
5) Swap data: If this field is specified, the program will look for this string in the specified field, and if found, will put the swapped data at that location.

The new download can be found at: http://oregonstate.edu/~reeset/marcedit/software/development/marcedit50_2005_07_21.zip.

FYI — I’ve had some folks start testing the UTF-8 to MARC-8 conversion and there are a few problems when dealing with CJK items. I’ve made the corrections and will be updating the build later today.

Updated:
I refreshed the build to include a small correction to the UTF-8 conversion routine. The problem occurred when you had dangling combining characters in a record. The program will now detect these danglers and include a space for diacritic to combine on. This way, the diacritic isn’t lost and the combining character doesn’t attach itself to the wrong character.

–Terry

I’ve refreshed the beta build. As with previous versions, you will need to uninstall any current beta builds before installing this version. What’s changed:
1) MARCEngine fixes dealing with mnemonic replacement, MARCXML=>MARC function expanded to provide optional MARC-8 conversion
2) MARC Tools Windows cleaning up
3) Delimited Text Client has been included in this setup

Coming this week — the Z39.50 client. I have a working version that includes just the Z39.50 functionality. If I finish, I’ll include SRU, OpenSearch and OAI functionality in this first test release — but more likely, it will be in the following week. Also, I’ll likely start looking at setting up the scripting objects for those wanting to script in PERL or Vbscript sometime in the next couple of weeks.

Thanks for those testing — keep sending in your questions and comments. All are appreciated.

–Terry

Couple of bugs worth noting in the MarcEdit 5.0 beta (updated July 14th).
1) When a left bracket ({) is encountered in its mnemonic form — the mnemonic reader doesn’t reposition itself correctly to read the rest of the line. This will cause the engine to stop unexpectedly.
2) Delimited Text Translator appears to have been left out of the installation.

I’ll be refreshing the build tonight — I’ll correct the two noted errors above as well as finishing the MARC Tools window by connecting the UTF-8 to MARC-8 conversion to the MARCXML=>MARC translation, in addition to a few other changes. Also, the Z39.50 client should be making an appearance this week or next. I have a working version right now an am starting to work on adding support for OpenSearch, SRU and OAI harvesting (though when it is first included in the beta build, it may just include the Z39.50 functionality).

Kyle Banerjee and I took a trip to Mary’s Peak this weekend to get a bit of a hill workout. This is the largest mountain in the Coast Range in Oregon (about 4000+ feet) — not terribly large, but pretty constant climb. Reaching the summit is very worth it. On a clear day, you can see the Ocean, Mt. Hood, Mt. Adams, Mt. Rainer, etc. I really haven’t had a climbing workout this year, so I thought this would be a good first run. I’ve mostly been riding in the valley and getting pretty good speed on my 25 mile commute (I average about 20-23 mph) but I found that I’m definitely not in climbing shape. Sunday was hot (96 degrees) and clear and after an inital 4 or so miles up the base, Kyle flat out left me up the hill. I think I burned myself on the bottom — but I definitely wasn’t expecting the climb to be as difficult as it ended up being. Looks like I’ll have to get more climbing in my life.

Well, I’ve just installed Wordpress and we’ll see how this whole blogging thing goes. I’ll be using this mainly to chat about things I’m interested in and to put out information on projects I’m working on.