Blogs

I have long been concerned about technical problems with our chat software. Now and then I run a report showing, by week, how many chat requests we had and how many were joined by librarians.

The data looks something like this:

What I didn't know was why the chats aren't being picked up. I looked at week 8 (last week) and found of the 158 requests we missed:

137 patrons waited too long
  9 sessions are duplicates that no one picked up
  6 patrons disconnected before we got to them
  5 show librarians but not patrons for no apparant reason - why?
  1 shows a patron but no librarian - why?

So less than 1% had no connection between patron and librarian for unexplained reasons. It is possible these are technical problems - the server thinks the patron is there, but they are not. There still could have been other technical problems along the way, but I am pleased that our software works so well at getting people connected.

But the 137 patrons waiting too long is still a concern. Many of them waited over 10 minutes before being automatically disconnected. I dug a little deeper and found that 92 of the chats went to L-net and 45 of them to L-net local. L-net local makes up about 20% of the total traffic, but was responsible for almost one third of patrons timing out last week.

If L-net patrons are waiting too long, we are understaffed, but if L-net local patrons are waiting too long, something else is going wrong - L-net local is not usually subject to the same group visit pressures as L-net. I'll send out a reminder to L-net local libraries - the main point is to remember to log out before you go home.

One of the great things about chat 'widgets' on library websites is that people don't have to enter a lot of information about themselves to get help from a librarian. The down side is that libraries don't get to collect as much information about patrons. I think that's a fair trade for having more patrons getting service.

L-net has adopted this philosophy even in our non-widgetized forms, and the data we used to collect about what "level"of questions people were asking ("General Interest", "College / Research", "High School", "Elementary/Middle School" or "Professional" is no longer available. Since we used this data to infer who was using our service, we have also lost demographic information.

What we do instead is ask a patron to identify the library they usually use. If they come to L-net via a library website, we infer they use that library. If not, we ask them for the name of their library and a zip code, one of which we eventually map to at least one academic, public, school or tribal library, or an area of Oregon unserved by public libraries.

Our statistics by patron's library page reflects this information. This grant year so far, the numbers show up like this:

This doesn't add up to 100% of our patrons because no category is exclusive of the others. If a patron enters "Knight Library" for their library and "97405" for their zip code, we identify them as a patron of both the University of Oregon and the Eugene Public Library. There are also patrons we can't identify at all, and I don't the exact number (I haven't created that report), but I would guess it would be at least 12%, but not much more.

If you're paying attention, and if you staff L-net, you'll also notice that saying that 2% of our patrons are school library patrons is a big, fat, lie. In fact, most of our patrons are kids. Kids who say they use the public library.

To get a better sense of just how many they were, I took a sample by looking at the 158 questions asked on February 23, 2010 and categorized them by what kind of question they were. It's fairly unscientific for categorizing the service as a whole, it is just an analysis of what happened on L-net last Tuesday:

The other day I had a chat with a teen patron and ending up recommending a book, The Basketball Diaries by Jim Carroll. Knowing about this book that covered a subject that the patron was very interested in - sex and drugs - helped turn around an "inappropriate" question into a fun and successful chat.

Afterwards I asked Sara Ryan, the teen services librarian at Multnomah County Library if there was such thing as a Sex and Drugs booklist. After all, I've got to have more than one ace up my sleeve the next time this comes up.

Multnomah County Library has a set of booklists for teens and a teen section of their website. We also considered if lists of banned books or lists from watchdog groups like Parents Against Bad Books in Schools might help.

These are all good places to start, and I do think it helps if you can say you have read the books you recommend (how impossible is that?), but what do you do with a patron who is expressing an explicit interest in a topic such as sex and drugs?

We brainstormed and came up with these books and lists of books. Sara gets most of the credit. Like all of it:

Sex

• YALSA's '08 Popular Paperbacks list "Sex is..."
• Multnomah County Library's If you Like Zane... is also called "African-American Erotic Fiction" - I'm not sure where the general list of erotic fiction is, though you can always use the subject heading.

Drugs

• The Guardian UK has a recent article on books with fictional drugs
• Biographies of famous musicians are bound to intersect with drugs, but also to cover much broader topics.

More

• San Francisco Public Library's "Gang-Related: Nonfiction" and "Gang-Related: Fiction" lists
• Multnomah County Library's Urban Fiction and Beyond booklist
• Little Willow (prominent YA blogger)'s master list of "Tough Issues for Teens" books (first created in '06 and updated consistently ever since)

These lists will never be complete - please brainstorm away.

Sarah Ludwig on the YALSA blog recently expressed the idea that libraries shouldn't try too hard to be cool. Instead, we should show that we are passionate about something. Anything. Books? That sounds like a good place to start.

I've been getting asked about this a bunch lately, so here is an answer for everyone. You may have noticed that telephone numbers are displaying in completed transcripts as hash marks:

(###) ###-####

What is going on? Did the patron get the correct number in the answer?

Questions that are 'closed' are displayed as having been scrubbed, though answers sent to patrons are not. Questions that are 'open' are displayed without scrubbing. When you answer a question, it is automatically set to 'closed', so your answer will immediately display as scrubbed.

The idea right now is to make sure the scrubbing works well before we make it permanent.

Names (if we know it), e-mail addresses, phone numbers, groups of 5 numbers or more (ie zip codes) and groups of 11 numbers-and-spaces or more (sometimes people enter barcodes and credit card numbers) are removed from transcripts if they are not part of a URL.

We scrub phone numbers because the computer system doesn't know who it belongs to - if it is a patron's phone number, we want it removed, but who can tell? So we are removing anything that looks like it might be a phone number.

And all of this begs the question, why keep any of it at all? I used to be more firm in my conviction that the transcripts were valuable for statistics, for education, to be re-used, and to be mined for the connection between the questions people ask, the words we use in conversation, and the resources librarians use to answer them. I still believe this, but since we are quite slow to do anything about it lately, I am leaning more and more towards deleting transcripts altogether.

To help make things a little clearer, I added a message to closed transcripts:

Thanks to the diligence of many librarians recently, and especially Michael Bowman at Portland State University and Laura Orr at the Washington County Law Library, I was finally able to track down and squash a bug that had been causing patrons problems for at least three months.

In one patron's words,

[the librarian] never answered my question, he just left the chat.. great survice this place is....

Another, who replied to a transcript after her chat session said,

The chat never worked.

Something was happening where the patron would seem to be connected but couldn't see messages on the patron's screen.

I tried tracking down this problem by comparing the users' browsers, but it was a different one every time. I tried communicating with patrons who it was happening to, but didn't get very far. I tried ensuring that JavaScript was enabled before connecting patrons, but it still happened.

As it turned out, the one place I didn't look was in the question itself. Before our software upgrade on November 16th of last year, I had tested the chat by submitting a question with every character on the keyboard:

`~1!2@3#4$5%6^7&8*9(0)-_=+QqWwEeRrTtYyUuIiOoPp{[}]|\AaSsDdFfGgHhJjKkLl:;"'ZzXxCcVvBbNnMm,,>.?/

Only I forgot to test the enter key. Patrons who wrote long questions and broke up their thoughts into multiple lines, and patrons who just happened to hit enter somewhere in their question, even at the end, had the problem described above. All they saw was a blank chat screen, even when the librarian sent messages.

When we upgraded the software, we added a feature for the patron's question to display on their screen and ours. When the question had a line break, this feature broke and stopped the whole chat.

Oops, grr, d'oh, and phew.

The problem is fixed now, for both the regular L-net chat and L-net Local. Here is a look at some of the damage that may have been done:

Chatssince 11/16/2009: 7,599
That include a line break in the question: 1,648
Probable technical problems: 22%

I'm going to keep paying attention to this issue, but for now am pleased that more patrons will get a good experience from L-net.