"Adventures in Library Patron Borrowing" : Research for App and Talk

[Andrew Frances]

I thought I would try Montreal-Python first!!!

Intro:

During Pycon 2015, I held an Open Space talk called "Adventures in Library Borrowing." I love libraries. Not too keen on the patron experience with the information systems. I was surprised folks showed up: thanks!!!

I demoed a simple Google appEngine* prototype that I have been running for well over a year. The application takes an E-mail library loan receipt, parses the entries, places the entries in my calendar*, sends a SMS and tweets (this was to test the framework. Good thing I borrow books mostly from the QA76 and TK5105 sections :-) ). 

The folks liked the prototype, were supportive, and shared their thoughts about what they would like to see in library applications. A common theme was why no library API for patrons. 

Side Story: 

I forgot I had the app running, in part because my University's e-mail loan receipt system was so flaky. However in December 2014, I just purchased a Pebble. When I borrowed a book, it was a bit of surprise and thrill to have the Pebble vibrate (I wasn't used to the watch) and see the loan receipt. I told the clerk not to print the receipt. 

Research/Challenge:

I thought I would revisit the app, spruce it up and put it in Github. An issue is that it is tedious to write a custom parser for each library (of course, this begets the question of having a standardized format for these documents). Could natural language processing, machine learning and heuristics come to the rescue (I have just started playing with NLTK) ?

I would like to put material together for a talk I would give at Montreal-Python in Fall 2015. 

A side issue though is that I really don't have much knowledge about library systems out there....

I have five library memberships, all in the Montreal area. 

- All have web interfaces.

- Four send e-mail reminders..

- To the best of my knowledge, in lieu of paper,  two send e-mail loan receipts.

- None do SMS.

- At least three of the systems use the same OPAC,  two send email and one does not.

Although library email messages more or less contain similar information, there are variances. Enough to make writing customer parsers a wee tedious. I would like to find out a few things:

1) What are the most common elements in a particular library email?

2) What features does a particular library system support?

I would appreciate it if 

1) You can tell me about your library. For example, does it issue e-mail loan receipts? Is it straight text? Or text and HTML?

2) If you could send samples of library loan receipts, recalls, reminders and over-due notices, all personal information removed. Here is an example from two of my libraries

(McGill Loan receipt)

Title:	[title] / [author]
Location:	[call number][location]
Barcode:	[barcode]
Due:	[date]

Title:	[title] / [author]
Location:	[call number][location]
Barcode:	[barcode]
Due:	[date]

-----------

(Concordia Loan receipt)

TITLE      LABEL NO.            DUE DATE
[TITLE]    [SOME NUMBER]  [DATE]

That said, please feel either post information in your receptive mailing list. Or if polluting a mailing list is a problem, email me at andrew.francis(at)mail.mcgill.ca. Feel free to ask me questions!!!!

Again, I would like to see where the road leads me. I'll summarise all my finding in a blog post and/or a talk along with code.

Thanks,

Andrew

* I am serious thinking about dropping appEngine.